METHOD FOR COUNTING AND CHARACTERIZING PARTICLES IN A FLUID IN MOTION

专利摘要:
The invention is a method for tracking particles in a moving fluid by an optical method. The particles are moving in a fluidic chamber. An image of the fluidic chamber is acquired, so as to obtain three-dimensional positions of particles in the fluid chamber at a first instant. Three-dimensional positions of particles are also obtained at a second instant, the second instant being subsequent to the second instant. From the three-dimensional positions obtained, potential particle displacements between said instants are established. Based on a model of particle displacement, potential displacements are validated. The validated displacements make it possible to count the particles in the fluid. In addition, if the particles are of a different nature, the displacement pattern may comprise a component of movement of the particles with respect to the fluid that is characteristic of this difference. The determination of this component then makes it possible to characterize the particles.
公开号:FR3061297A1
申请号:FR1663475
申请日:2016-12-28
公开日:2018-06-29
发明作者:Pierre Joly；Rodrigue ROUSIER；David ELVIRA
申请人:Commissariat a lEnergie Atomique CEA；Commissariat a lEnergie Atomique et aux Energies Alternatives CEA；
IPC主号:

专利说明:

@ Holder (s): COMMISSION FOR ATOMIC ENERGY AND ALTERNATIVE ENERGIES Public establishment.
® Agent (s): INNOVATION COMPETENCE GROUP.
® METHOD FOR COUNTING AND CHARACTERIZING PARTICLES IN A MOVING FLUID.
FR 3,061,297 - A1
©) The invention is a method for monitoring particles in a moving fluid, by an optical method. The particles are moving in a fluid chamber. An image of the fluid chamber is acquired, so as to obtain three-dimensional positions of particles in the fluid chamber at a first instant. Three-dimensional positions of particles are also obtained at a second instant, the second instant being after the second instant. From the three-dimensional positions obtained, potential displacements of particles, between said instants, are established. Based on a particle displacement model, potential displacements are validated. The validated displacements make it possible to count the particles in the fluid. In addition, if the particles are different in nature, the displacement model may include a component of movement of the particles relative to the fluid which is characteristic of this difference. The determination of this component then makes it possible to characterize the particles.
Ml; 5U
4C o
QS>
O ° o,
I2
i
Method for counting and characterizing particles in a moving fluid Description
TECHNICAL AREA
The technical field of the invention is the counting of particles circulating in a fluid chamber using an optical method.
PRIOR ART
Several optical methods have already been implemented for counting particles circulating in a fluid, the fluid being a gas or a liquid. A very widespread principle is based on illumination, using a light beam of a fluid in which particles circulate. When a particle crosses the beam, part of the light is scattered and can be detected by a photodetector. This principle has been implemented for the detection of particles in the air, or for the detection of cells in liquids, for example biological liquids.
Other methods are based on the analysis of images, for example using a microscope, but the images provide only two-dimensional information as to the position of the particles.
The document WO2008090330 describes a device allowing the observation of samples comprising cells by lensless imaging. The sample is disposed between a light source and an image sensor, without having an optical imaging system between the sample and the image sensor. Thus, the image sensor collects an image, also called a hologram, formed from interference patterns between the light wave emitted by the light source and transmitted by the sample, and diffraction waves, resulting from diffraction by the sample of the light wave emitted by the light source. These interference figures are generally formed by a succession of concentric rings. They are sometimes called diffraction figures, or designated by the English term "diffraction pattern". We thus acquire images, whose field of observation is much larger than that of a microscope. When the concentration of cells in the sample is sufficiently low, each cell can be associated with an interference pattern; their enumeration allows the counting of the cells present in the sample. However, the hologram does not allow reliable counting of cells when the concentration increases, and / or when the particles are in motion. The method also shows limits when the hologram has a low noise ratio, for example when the size of a particle is small or when a particle has a refractive index close to that of the medium forming the sample.
Patent application US2012 / 148141 describes a process repeating the principles of WO2008090330, by implementing a holographic reconstruction algorithm with a succession of acquired images to reconstruct complex images of spermatozoa. The objective is to characterize their motility. It is a process based on individual monitoring of trajectories of moving particles, in an immobile fluid, on the basis of three-dimensional coordinates of the particles, obtained by holographic reconstruction. Indeed, such a reconstruction makes it possible to provide an estimate of a distance between an image sensor and a particle, allowing access to so-called depth information, supplementing the two-dimensional information obtained by conventional image sensors. . Furthermore, the method makes it possible to determine a displacement model of each particle, the displacement model being a result obtained by the implementation of the method.
In addition, numerous flow imaging methods, designated by the acronym Particle Imaging Velocimetry, use optical particle detection methods in order to characterize their movement, which is representative of the movement of the fluid studied.
The inventors of the present invention have proposed a method for locating and counting particles circulating in a fluid chamber, which can be automated and implemented in a moving fluid. The method can be implemented automatically, and address high speeds or quantities of particles. In addition, when particles of different types are present in the fluid chamber, the method can allow discrimination between different types of particles, so as to count the number of particles of different types, based on their respective displacements.
STATEMENT OF THE INVENTION
An object of the invention is a method for counting particles in motion in a fluid circulating in a fluid chamber, the method comprising the following steps:
a) arrangement of the fluid chamber between a light source and an image sensor, the image sensor extending along a detection plane;
b) illumination of the fluid chamber by the light source, the light source emitting an incident light wave propagating along a propagation axis and acquisition, by the image sensor, of an image, called the first image, representative of a light wave, called exposure wave, to which the image sensor is exposed, the image sensor comprising different pixels, with each pixel being associated with a radial coordinate in the detection plane;
c) from the acquired image, obtaining coordinates, in particular three-dimensional, of particles, in the fluid chamber, at a first instant,
d) obtaining coordinates, in particular three-dimensional, of particles in the fluid chamber at a second time, after the first time;
e) from the coordinates of the particles obtained at the first instant and at the second instant, determination of potential displacements of the particles between said instants.
f) taking into account a model of displacement of the fluid in the fluid chamber;
g) from the fluid displacement model considered during step f), validation of displacements among the potential displacements calculated during step e);
h) from the displacements validated during step g), determination of a number of particles and / or of the coordinates of the particles at the first instant and / or at the second instant.
Steps f) and g) are optional. When they are not implemented, the method comprises a step of determining a number of particles and / or the coordinates of the particles at the first instant and / or at the second instant from the potential displacements determined during the step e).
Step c) can include:
obtaining a first image of interest from the first image acquired during step b), and applying a digital propagation operator to the first image of interest, according to at least a distance of reconstruction, along the axis of propagation, so as to obtain at least one complex reconstructed image;
from each reconstructed complex image, obtaining radial coordinates of particles in the fluid chamber at the first instant.
Step c) can include the following sub-steps:
ci) obtaining a first image of interest from the first image acquired during step b), and applying a digital propagation operator to the first image of interest, according to a plurality of reconstruction distances , along the propagation axis, so as to obtain a first stack of reconstructed complex images, called the first stack of images, comprising as many complex reconstructed images as reconstruction distances, each reconstructed complex image being representative of a exposure light wave to which the image sensor is exposed;
cii) for at least one radial coordinate defined by the first image of interest, determination of a reconstruction distance maximizing the evolution of a component of each complex image forming the first stack of images, along an axis parallel to the axis of propagation and passing through said radial coordinate, said determined reconstruction distance forming a transverse coordinate associated with said radial coordinate, the value of the component calculated at said reconstruction distance being a so-called maximum value associated with said radial coordinate , the sub-step cii) being carried out for all or part of the radial coordinates associated with the pixels of the first image of interest;
ciii) establishment of a list of three-dimensional positions, each three-dimensional position comprising a radial coordinate and the associated transverse coordinate, determined during the sub-step cii), with each three-dimensional position being associated the maximum value obtained during the sub-step cii);
civ) selection of three-dimensional positions according to the maximum value associated with them.
The first image of interest can be:
the first image acquired during step b);
or the first image acquired during step b), from which an image of the fluidic chamber, subtracted by the image sensor, is subtracted before or after the acquisition of the first image, the subtraction being weighted by a weighting term, the weighting term possibly being a real between 0 and 1;
or the first image acquired during step b), from which an average of images acquired respectively before and after the acquisition of the first image is subtracted.
The component considered during the sub-step cii) can comprise the real part, or the imaginary part, or the module, or the phase of each complex image forming the stack of images.
The civ) sub-step can include:
an image formation, known as the first maxima image, each pixel of which is associated with a three-dimensional position determined during sub-step ciii), and is assigned the maximum value determined, during sub-step ciii), for said three-dimensional position;
a selection, in the first image of the maxima, of pixels whose value is maximum in a neighborhood area defined around each pixel;
a calculation, for each selected pixel, of a signal-to-noise ratio as a function of said maximum value and of the pixel value of the first image of the maxima situated in a calculation area extending around said pixel;
so that each three-dimensional position is selected according to the signal to noise ratio calculated for the pixel of the first image of the maxima associated with it. The pixel selection step is optional, but preferable, and the signal-to-noise ratio can be calculated on all the pixels of the maxima image. A position whose associated pixel, on the first image of maxima, is not selected, is invalidated.
According to one embodiment, step d) may comprise an acquisition, by the image sensor, of a second image, each pixel of which is associated with a radial coordinate in the detection plane (P _o ). According to this embodiment, step d) comprises the following sub-steps:
di) obtaining a second image of interest from the second acquired image and applying a digital propagation operator to the second image of interest, according to a plurality of reconstruction distances, along the propagation axis, so as to obtain a second stack of complex reconstructed images, called the second stack of images, comprising as many complex reconstructed images as reconstruction distances, each complex reconstructed image being representative of a light wave of exposure to which is exposed the image sensor at the second instant;
dii) for at least one radial coordinate defined by the second image of interest, determination of a reconstruction distance maximizing the evolution of a component of each complex image forming the second stack of images, along an axis parallel to the propagation axis and passing through said radial coordinate, said reconstruction distance forming a transverse coordinate associated with said radial coordinate, the value of the component calculated at said reconstruction distance being a so-called maximum value associated with said radial coordinate, sub-step dii) being carried out for all or part of the radial coordinates associated with the pixels of the second image of interest;
diii) establishment of a list of three-dimensional positions, each three-dimensional position comprising a radial coordinate and the associated transverse coordinate, determined during sub-step dii), with each three-dimensional position being associated the maximum value obtained during the sub-step dii);
div) selection of three-dimensional positions according to the maximum value associated with them.
During sub-step di), the second image of interest can be:
the second acquired image;
or the second acquired image, from which an image of the fluidic chamber is subtracted, acquired by the image sensor, before or after the acquisition of the second image, the subtraction being weighted by a weighting term;
or the second acquired image, from which an average of images acquired respectively before and after the acquisition of the second image is subtracted.
During sub-step dii), the component can include the real part, or the imaginary part, or the module, or the phase of each complex image forming the stack of images.
The sub-step div) can include:
forming an image, called the second maxima image, each pixel of which is associated with a three-dimensional position determined during sub-step diii), and is assigned the maximum value determined, during sub-step diii), for said three-dimensional position;
a selection, in the second image of the maxima, of pixels whose value is maximum in a neighborhood area defined around each pixel;
a calculation, for each selected pixel, of a signal to noise ratio as a function of said maximum value and of the value of pixels of the second image of the maxima situated in a calculation zone extending around said pixel;
so that each three-dimensional position is selected as a function of the signal-to-noise ratio calculated for the pixel of the second image of the maxima associated with it. The selection step is optional, but preferable, and the signal to noise ratio can be calculated on all the pixels of the maxima image. A position whose associated pixel, on the second image of maxima, is not selected, is invalidated.
According to one embodiment:
step b) comprises two successive illuminations of the fluid chamber by the light source, at the first instant and at the second instant, so that the first image (/) represents the wave of exposure (14) to each of the moments;
steps c) and d) are combined in the same step of obtaining the coordinates of particles at the first instant and at the second instant.
The process can include one of the characteristics taken in isolation or according to the technically feasible combinations:
step e) may include a comparison of the coordinates of the particles in the fluid chamber determined at the first instant and at the second instant, so as to establish a list of potential displacements of the particles between said times, step g) includes a determination of a range of displacements using the displacement model taken into account during step f), the potential displacements being validated when they are included in said displacement range.
step g) comprises a subtraction, at each potential displacement, of a displacement according to the displacement model.
According to one embodiment, the fluid comprises particles, each particle having and moving, relative to the fluid, according to a displacement model, called the particle displacement model, depending on said property. According to this embodiment, the method comprises, on the basis of displacements validated during step g), a step i) of taking into account at least one model of particulate displacement, so as to count the particles as a function of 'a value of said property. According to a variant, the method comprises, on the basis of potential displacements determined during step e), a step e ') of taking into account at least one particle displacement model, so as to count the particles as a function of 'a value of said property. The property is a mass or an electric charge, or an ability to move in the fluid.
According to this embodiment, the method can include:
taking into account a particular displacement model for a predetermined value of the property;
a calculation of deviations of the displacements of each particle with respect to the particle displacement model;
so that the property of each particle is determined according to said deviations as well as said predetermined value of the property.
The fluid can flow in one direction of flow, and the particulate movement can take place in another direction not parallel to said direction of flow.
The method may be such that no imaging optics are disposed between the image sensor and the fluid chamber. It can also be such that the image sensor includes an image-forming optic between the image sensor and the fluid chamber, the image formed during step b) being a defocused image.
Another object of the invention is a device for counting particles circulating in a fluid chamber, the device comprising:
a light source configured to illuminate the fluid chamber;
an image sensor extending in a detection plane, the fluid chamber being interposed between the image sensor and the light source, the image sensor being configured to acquire a plurality of images of the illuminated fluid chamber by the light source, at an instant or at successive instants;
the device comprising a processor capable of implementing steps c) to e) or c) to h) of a method as described in this application, from at least one image acquired by the image sensor.
Another object of the invention is a device for the observation of particles circulating in a fluid chamber comprising the following steps:
arrangement of the fluid chamber between a light source and an image sensor, the image sensor extending along a detection plane;
illumination of the fluid chamber by the light source, the light source emitting an incident light wave propagating along a propagation axis and acquisition, by the image sensor, of an image representative of a light wave, called exposure wave, to which the image sensor is exposed, the image sensor comprising different pixels, with each pixel being associated a radial coordinate in the detection plane;
selection of an image, acquired at an acquisition instant, and formation of an image of interest by subtraction, from the selected image, of an image acquired at a later instant and / or of an image acquired at an instant prior to the acquisition instant, the subtraction being able to be weighted by a weighting term, for example between 0 and 1;
application of a digital propagation operator to the image of interest, according to a plurality of reconstruction distances, so as to obtain a stack of reconstructed complex images, called a stack of images, comprising as many complex reconstructed images as reconstruction distances, each reconstructed complex image being representative of an exposure light wave to which the image sensor is exposed;
for at least one radial coordinate defined by the image of interest, determination of a reconstruction distance maximizing the evolution of a component of each complex image forming the stack of images, along an axis parallel to l propagation axis and passing through said radial coordinate, the component comprising the real part of each complex image, or its opposite, said determined reconstruction distance forming a transverse coordinate associated with said radial coordinate, the value of the component calculated at said distance reconstruction being a so-called maximum value associated with said radial coordinate, this step being carried out for all or part of the radial coordinates associated with the pixels of the image of interest;
establishment of a list of three-dimensional positions, each three-dimensional position comprising a radial coordinate and the associated transverse coordinate, determined during the previous step, with each three-dimensional position being associated with the maximum value obtained during the previous step;
selection of three-dimensional positions according to the maximum value associated with them.
The selection of three-dimensional positions can be carried out by forming an image of the maxima and by determining a signal-to-noise ratio at each pixel of the image of the maxima, as previously described.
Other advantages and characteristics will emerge more clearly from the description which follows of particular embodiments of the invention, given by way of nonlimiting examples, and represented in the figures listed below.
FIGURES
FIG. 1A represents an example of a device allowing the implementation of the invention. Figure IB is a detail of Figure IA.
FIG. 2A illustrates a succession of steps allowing an implementation of a first embodiment of the method. FIGS. 2B, 2C and 2D show profiles of a part of an acquired image extending around a hologram corresponding to a particle, at three successive instants. Figure 2E shows a profile corresponding to a fixed component, obtained by combining the profiles of Figures 2B to 2D. FIG. 2F shows a profile corresponding to a mobile component, this profile being obtained by subtracting the fixed component, represented in FIG. 2D, from the profile of FIG. 2C. FIG. 2G represents a profile modeling the evolution of the module of a complex image reconstructed from an image corresponding to the profile represented in FIG. 2F. FIG. 2H represents a profile modeling the evolution of the opposite of the real part of a complex image reconstructed from an image corresponding to the profile represented in FIG. 2F.
FIG. 3A shows a fluid chamber implemented during a first experimental test. FIG. 3B is a plot of results obtained during the first experimental test. Figure 3C shows the impact of a threshold value on metering performance.
FIG. 4A illustrates a succession of steps allowing an implementation of a second embodiment of the method. Figure 4B shows results obtained from a second experimental trial.
FIG. 5 shows the results obtained from a third experimental test, implementing a variant of the invention.
EXPLANATION OF PARTICULAR EMBODIMENTS
Figure IA shows an example of a device for implementing the invention. A light source 11 is capable of emitting a light wave 12, called the incident light wave, propagating towards a sample 10, along a propagation axis Z. The light wave is emitted according to a spectral band Δλ.
The sample 10 is a sample comprising particles 10a which it is desired to count, the particles being placed in a transparent or translucent carrier fluid medium 10b. The particles are small elements, and are inscribed in a diameter between 0.1 pm and 100 pm; or between 1 pm and 100 pm. The particles are solid or liquid. It can be dust, or cells or microorganisms or microbeads, usually used in biological applications, or even microalgae. They may also be droplets insoluble in the fluid 10b, for example oil droplets dispersed in an aqueous phase. The carrier medium 10b is a fluid, for example air or a liquid, for example water or a biological liquid. The sample may for example be an aerosol, comprising particles in suspension in a gas, the latter possibly in particular being air.
The sample 10 is contained in a fluid chamber 15. The thickness e of the sample 10, along the propagation axis typically varies between 10 pm and 2 cm or 3 cm, and is preferably between 20 pm and 1 cm. The sample extends along a plane, called the plane of the sample, preferably perpendicular to the axis of propagation Z. The fluid chamber 15 is held on a support 10s facing the image sensor 20.
The particles 10a are movable in the fluid chamber 15, being carried by the fluid 10b, the latter being movable in the fluid chamber 15. In this example, the fluid flows, in the fluid chamber 15, along an axis of longitudinal flow X. The particles 10a are then entrained by the fluidic movement of the medium 10b, the latter acting as a carrier medium, and forming a fluidic current inside the fluidic chamber 15. The displacement of the medium can be modeled. The particles 10a can also be mobile relative to the medium 10b, the movement of the particles relative to the fluid which carries them being designated by the term particulate movement. Thus, the movement of the particles 10a in the fluid chamber 15 is not random and obeys a predetermined displacement model, combining the fluid movement of the medium 10b and, optionally, the particle movement of the particles relative to the fluid.
The distance D between the light source 11 and the sample 10 is preferably greater than 1 cm. It is preferably between 2 and 30 cm. Advantageously, the light source, seen by the sample, is considered as a point. This means that its diameter (or its diagonal) is preferably less than a tenth, better a hundredth of the distance between the sample and the light source. In the example shown, the light source 11 is a laser diode. Alternatively, the light source 11 is a white light source or a light emitting diode. In this case, a spatial filter is advantageously placed between the light source and the sample, so that the light source appears as a point. The spatial filter can be a pinhole camera or an optical fiber. A wavelength filter is also preferably placed between the light source and the sample, to adjust the emission spectral band A de of the incident light wave 12. Preferably, the emission spectral band Δλ of the incident light wave 12 has a width less than 100 nm. By spectral bandwidth is meant a width at half height of said spectral band.
The fluid chamber 15 is disposed between the light source 11 and the previously mentioned image sensor 20. The latter preferably extends parallel, or substantially parallel to the plane along which the sample extends. The term substantially parallel means that the two elements may not be strictly parallel, an angular tolerance of a few degrees, less than 20 ° or 10 ° being allowed. The image sensor 20 is able to form an image I according to a detection plane P _o . As shown in FIG. 1B, the image sensor comprises a matrix of pixels, with each pixel being associated with coordinates (x, y), called radial coordinates, in the detection plane P _o . The image sensor can in particular be a CCD type sensor or a CMOS. The detection plane P _o preferably extends perpendicular to the axis of propagation Z of the incident light wave 12. The distance between the sample 10 and the pixel matrix of the image sensor 20 is between a distance minimum d _min and maximum distance d _max . The thickness e of the fluid chamber corresponds to the difference between d _max and d _min . d _min can be between 50 pm and 2 cm, preferably between 100 pm and 2 mm. The thickness of the fluid chamber is generally between 100 μm and 5 cm.
Note the absence of an optical imaging system, in particular a magnification optic between the image sensor 20 and the sample 10. This does not prevent the possible presence of focusing microlenses at each pixel of the image sensor 20, the latter having no function of enlarging the image acquired by the image sensor. The image sensor 20 is thus placed in a configuration known as lensless imaging. Such a configuration makes it possible to obtain a high field of observation. Other configurations can nevertheless be envisaged, in particular a configuration according to which a focusing optic is interposed between the image sensor 20 and the fluid chamber 15. In such a configuration, the image sensor acquires a defocused image of the sample 10.
Under the effect of the incident light wave 12, the particles present in the fluid chamber can generate a diffracted wave 13, capable of producing, at the level of the detection plane P _o , interference with part of the incident light wave 12 transmitted by the sample. Furthermore, the sample can absorb part of the incident light wave.
12. Thus, the light wave 14, called the exposure light wave, transmitted by the sample 10 and to which the image sensor 20 is exposed, can comprise:
a component 13 resulting from the diffraction of the incident light wave 12 by each particle of the sample;
a component 12 ′ resulting from the absorption of the incident light wave 12 by the sample.
These components form interference in the detection plane. Also, the image I acquired by the image sensor 20 includes interference patterns (or diffraction patterns), each interference pattern being generated by a particle 10a of the sample 10.
A processor 30, for example a microprocessor, is configured to process each image I acquired by the image sensor 20. In particular, the processor is a microprocessor connected to a programmable memory 32 in which is stored a sequence of instructions for performing the image processing and calculation operations described in this description. The processor can be coupled to a screen 34 allowing the display of images acquired by the image sensor 20 or calculated by the processor 30.
The fluid chamber 15 is fixed relative to the image sensor 20. Thus, the fluid medium 10b and the particles 10a circulating in the fluid chamber are in motion relative to the image sensor 20.
As indicated in relation to the prior art, a propagation operator h can be applied to each image acquired by the image sensor, so as to calculate a complex quantity representative of the exposure light wave 14. It it is then possible to calculate a complex expression A of the light wave 14 at any point of coordinates (x, y, z) of space, and in particular according to a reconstruction surface extending opposite the image sensor 20 The reconstruction surface is usually a plane P _z , called the reconstruction plane, extending parallel to the image sensor 20, at a coordinate z of the detection plane P _o . The reconstruction plane P _z is then parallel to the detection plane P _o . One then obtains an image, called complex image A _z , representative of the light wave of exposure 14 in the reconstruction plane P _z . The complex image A _z is obtained by a convolution of the image I acquired by the image sensor 20 by the propagation operator h, according to the expression: A _z = I * h.
The propagation operator h describes the propagation of light between the detection plane P _o and the reconstruction plane P _z . In this example, the detection plane P _o has the equation z = 0.
The propagation operator is for example a so-called Fresnel operator, defined according to
-ί ^ίπ ( _χ 2 i, .2-) the following expression: h _z (x, y) = - e ^ '' (D
Λ.Ζ
A feature of the invention is that the particles 10a move, being entrained by the fluid 10b. The fluid moves between an inlet and an outlet of the fluid chamber 15, along a flow axis X. In order to count them, it is necessary to obtain three-dimensional positions of the particles at a first instant t _± and at a second instant t ₂ , after the first instant, the time difference Δί = t ₂ - t _± between the two instants depending on a maximum speed V _max of the fluid in the fluid chamber 15 as well as on the dimension of the part of the fluid chamber seen by the sensor. If L denotes a dimension of the fluid chamber 15, seen by the image sensor 20, along the axis of propagation X of the fluid, it is preferable that: Δί <^ V _max .
Different embodiments are possible. According to a first embodiment, the image sensor acquires two successive images / (G) and i (f ₂ ), respectively at the first instant and at the second instant t ₂ . From each image, three-dimensional coordinates of the particles are obtained at each instant. According to a second embodiment, the same image of the fluid chamber is acquired at the two instants, the acquisition of this image being carried out at the first instant and at the second instant.
The main steps of the first embodiment of the method are described below, in connection with FIG. 2A.
Step 100: acquisition. It is a question of acquiring an image / (fj at different times ί _έ , according to an acquisition frequency. During a first iteration, the instant L is a first instant and one acquires an image called first image / ( G). During a second iteration, the instant L is a second instant t ₂ , the second instant being after the first instant.The image acquired at the instant t ₂ is a second image i (f ₂ ).
Step 110: extraction of an image of interest from the acquired image, the image of interest representing a mobile component / _m (ij) of the acquired image. The acquired image Z (tj) comprises a component called the fixed component, representing the elements considered as not dependent on time, and a component / _m (ij) called the moving component, representing the elements considered to be in motion in the image. The particles moving in the sample are in motion and form the motion component. The first filtering aims to remove the fixed component from the acquired image. The fixed component can be obtained by means of one or more images acquired at different times different from the instant of acquisition of the filtered image. The fixed component If (ti) can be estimated by an initial image I (t ₀ ), acquired while no particle is circulating in the fluid chamber 15. This makes it possible to obtain an image of the fixed elements, for example dust, not representative of the moving particles to be counted. Preferably, the fixed component If (ti) is estimated by an average between an image acquired at an anterior instant and an image acquired at an instant posterior to the acquisition instant tj of the acquired image. It can be for example the instant before t _i _ ₁ and the following instant t _{i + i} the acquisition instant in which case the fixed component is such that / fe + i) + (2)
The estimate of the fixed component is thus renewed with each new acquisition of an image. It corresponds to an average of two images respectively acquired before and after the image acquired considered, the average being weighted by a weighting factor of L. This allows a regular updating of the fixed component.
The fixed component is subtracted from each acquired image, so as to obtain a mobile component I _v , representative of the mobile elements in the image, and in particular of the mobile particles. / _V (L) = / (tj) - Iy (i) (3).
The mobile component forms an image of interest on the basis of which the following steps are carried out. At the first instant t _1; the image of interest is noted / ^ (^). At the second instant t ₂ , the image of interest is denoted I _v (t ₂ ).
FIGS. 2B to 2D represent modeled examples of intensity profiles of a hologram, corresponding to a particle, on an image acquired by the image sensor 20, respectively at times f and t _{i + 1} . The particle moves along the axis of flow of the fluid X, which results in a translation of the hologram, the latter being represented by a brace in each of these figures. The undulations observed on both sides of the hologram correspond to the effect of mounting imperfections. These imperfections are in particular the non-uniformities of illumination and the interference between reflections taking place at the interfaces of the chamber. This is reflected in the fact that in FIGS. 2B, 2C and 2D, the profiles at the hologram level are asymmetrical and different. Figure 2E shows the fixed component, If (tt) as determined by expression (2). The fixed component If (tt) includes the effect of the imperfections of the assembly as well as the holograms corresponding to the instants tj ^ and t _{i + 1} , the latter being weighted by a weighting factor equal to ¹ Λ. FIG. 2F represents the mobile component / _ν (ί,) obtained according to expression (3). We observe that the effect of imperfections has disappeared. The central hologram, corresponding to the position of the particle at time t ,, is symmetrical. The holograms corresponding to the instants tj ^ and t _{i + 1} are also symmetrical and are weighted with a weighting factor equal to -1/2. These residual holograms are called echoes.
Thus, this step makes it possible to estimate a mobile component of the acquired image, this mobile component being representative of the mobile elements, relative to the image sensor, at the acquisition time tt. This mobile component makes it possible to better show the mobile particles that we are trying to count.
Step 120: frequency filtering. The image of interest I _v (tt), resulting from step 110 is subject to a bandpass frequency filtering: such a filtering makes it possible to eliminate low spatial frequencies, linked to heterogeneities of the illumination of the sample, as well as high spatial frequencies, the latter being considered as noise. The passband of the frequency filter is preferably between a low cutoff frequency and a high cutoff frequency. The low cutoff frequency can be 0.02 f. The high cutoff frequency can be equal to 0.5 f. / is a frequency corresponding to half the spatial frequency defined by the size of the pixels: f = -. £ representing a dimension of a pixel (length or width).
Step 130: propagation of the filtered image. The image resulting from step 120 is propagated according to different reconstruction distances z _; , along the propagation axis Z. The reconstruction distances are determined so that the reconstruction planes P _z respectively associated with each reconstruction distance are included in the sample. Thus, from each acquired image / (£,), after the steps of extracting the image of interest 120 and filtering 130, a stack of complex images A _z (tf) reconstructed at different distances from reconstruction z _; . If d _min and d _max respectively denote the minimum distance and the maximum distance between the sample and the image sensor, the reconstruction is carried out so as to obtain different reconstruction planes between d _min and d _max . The number of reconstruction distances considered conditions the spatial resolution with which the coordinates of the particles are determined, as described below. The interval between two different reconstruction distances may for example be 100 µm. At the end of this step, there is a stack of complex images, each complex image extending parallel to the detection plane, at a z coordinate _; , called transverse coordinate.
Step 140: Extraction of a component from each complex image. It is a question of associating, with each pixel of the complex image, a real number. Thus, the stack of complex images A _z . (Ti) is replaced by a stack of images A ' _z . (Ti) of real numbers, each pixel A' _z . (Ti, x, y) of each real image . comp being a component (a _'z (ti, x, yf) of a complex image a _z (ti), to the transverse coordinate _z;., at the same radial coordinate (x, y) (that is by the same pixel). By component of a complex image, we mean a quantity obtained from the complex image at the radial coordinate (x, y). The component can be or include the real part, the part imaginary, or the module, or the phase, of the complex amplitude A _z . (x, y) of the complex image A _z . (ti) at the radial position (x, y). The component can combine quantities . listed in the preceding sentence here is sought to obtain a transverse coordinate _z;, denoted z _xy, maximizing the component, and for each radial coordinate (x, y) This maximization is subject to the step 145..
FIGS. 2G and 2H respectively represent a profile of a component of a reconstructed complex image, at a radial coordinate z _; , the complex image being obtained by holographic reconstruction of the image I _v (ti) whose profile is shown in FIG. 2F. In FIG. 2G, the profile of the module of the reconstructed complex image is shown. In FIG. 2H, the profile of the opposite of the real part of the reconstructed complex image has been shown. In other words, FIGS. 2G and 2H represent the profile of an image of real numbers obtained after extraction of a component from the reconstructed complex image, the component being respectively the module, a _Zj . (X, y) | , or the opposite of the real part, —Re ^ A _z . (x, y) ^. It is observed that when the component is the opposite of the real part, the central hologram, visible on the profile of FIG. 2H, is represented by a high and positive value. The holograms located on either side of the central hologram correspond to residues, or echoes, resulting from the extraction of the image of the mobile component. Their amplitude is lower than that of the central histogram, and is negative. We understand that it will be easier to discriminate the central hologram, corresponding to the particle at time ί _έ , holograms corresponding to a residue (or echo) of the particle at the instants anterior t _i _ ₁ and posterior t _{i + i} , these holograms not being representative of the particle at time tt. On the profile of FIG. 2G, the hologram of the particle at time tj results in a peak of positive and high amplitude, while the holograms of the particle at anterior times tj ^ and posterior t _{i + 1} are result in peaks of equally positive, but lower amplitude. We understand that it is more difficult to discriminate, on the basis of the module, the useful holograms, that is to say corresponding to a position of the particle, to the holograms of echoes, corresponding to a position of a particle at a anterior or posterior instant.
The comparison of FIGS. 2G and 2H shows that considering the real part, or its opposite, has an advantage, in particular compared to conventional approaches to holographic reconstruction, according to which the module of a complex amplitude is considered. In fact, unlike a module, a real part of an image is signed, in the sense that it can be positive or negative. It allows discrimination based on a sign, which is not possible when considering modules. The taking into account of the real part appears particularly relevant when it follows arithmetic combinations of images comprising a subtraction of images.
Step 145: Digital focus. During this step, for each pixel of the acquired image, that is to say for each radial position (x, y), a transverse coordinate z, along the propagation axis Z for which the comp component (A _z (ti)) of a complex image A _z . of the image stack, at the pixel of coordinates (x, y), is maximum. This involves applying a principle known as digital focusing known to those skilled in the art. A particle is present at an unknown distance from the image sensor. The closer the reconstruction distance is to this distance, the more the particle forms, on the reconstructed complex image, a narrow and intense spot. The complex image comprising a real part and an imaginary part, the search for the distance separating the particle from the detector is carried out by analyzing the spatial evolution of a component of each complex image along the axis of propagation.
we determine z _xy , such that z _xy = argmax (comp (A _Zj . (x, y) ^ = argmax Ç — Re (A _z . (x, y)) ^ (4).
This step is repeated for all or part of the radial positions (x, y) of the image sensor so that with each radial coordinate (x, y) is associated a transverse coordinate z _xy as defined in expression (4 ).
Step 150: formation of the image of the maxima.
Following step 145, a so-called maxima image is constituted, such that:
A _max (x, y) = comp (A _Zxy (x, y)) = -Re (A _Zxy (x, y)) (5).
This image comprises, at each pixel (x, y), the maximum value of the component, in the stack of complex images A _z ., Along the axis of propagation Z, determined during step 145. A each pixel (x, y) of the image of the maxima A _max is associated with the transverse coordinate z _xy identified during step 145.
During the first iteration (/ y = tf), we obtain a first image of the maxima. During the second iteration (/ y = t ₂ ), we obtain a second image of the maxima.
Step 160: search for local maxima in the image of the maxima.
During this step, a search for local maximum values is carried out by groups of adjacent pixels. For example, each group of pixels has 51 * 51 adjacent pixels. A pixel of the image of maxima A _max is considered to be local maximum if it is the pixel with the highest value in a group of 51 * 51 pixels centered on said pixel. The image of maxima A _max can be smoothed before the search for local maxima. It can be a smoothing by applying a Gaussian filter or a low-pass filter.
We can thus obtain a list of the coordinates of each local maximum pixel (x _max , y _max ) as well as the value A _max (x _max , y _max ) of the image of the maxima A _max at this pixel, as well as the coordinate transverse z _xv associated with this pixel. Other known digital focusing algorithms can be applied to the stack of images resulting from step 140, making it possible to define such a list. Such algorithms can for example be based on a criterion of sharpness on each image of the stack of images. These algorithms also make it possible to define local maxima as well as a transverse coordinate associated with these local maxima.
Step 170: taking into account the signal to noise ratio.
The search for local maxima in the image of maxima A _max can be affected by a non-homogeneous background. This non-homogeneous background is notably caused by fluctuations in interference fringes produced by the multiple interfaces between the light source 11 and the image sensor 20. Therefore, the inventors considered that it is preferable to take into account counts a signal-to-noise ratio at each radial coordinate determined during step 160. Thus, at each radial position x _max , y _m a _X defined in step 160, a signal-to-noise ratio SNR (x _max) is calculated , y _max ), this ratio being calculated using the information contained in the image of the maxima A _max . A local noise level is calculated, in the maxima image, around each radial position (x _max , y _max ), for example in a noise calculation zone centered on the position (x _max , y _max ) and of diameter equal to 200 pixels. The pixels considered for the calculation of the local noise can be all the pixels of the noise calculation zone, or certain pixels of this zone. The inventors have for example taken into account 100 pixels regularly distributed over the circle delimiting the noise calculation zone, the noise level being estimated by a calculation of the median of the value of these 100 pixels.
This step makes it possible to establish a list of radial coordinates (x _max , y _max ) corresponding to a local maximum in the image of the maxima, each pair of radial coordinates being associated with a transverse coordinate z _xv . list of three-dimensional positions (x _max , y _m ax> ^z x _max ymax ^ in the sample likely to contain a particle. Each three-dimensional position (% _max , y _max , is associated with an estimate of the signal-to-noise ratio SNR (x _max , y _max ) of the image A _max at the position ( ^X max> ymax) · Each of these three-dimensional positions is likely to be occupied by a particle 10a at the instant L considered.
Step 180: thresholding. During this step, the three-dimensional positions are subject to a thresholding of the signal to noise ratio which is respectively assigned to them. The thresholding is carried out according to a threshold value S which can be predetermined. Only three-dimensional positions whose associated signal-to-noise ratio is greater than the threshold value are retained, the others being considered as not representative of particles. The threshold can be predetermined, for example on the basis of calibrations, or optimized as described later, in connection with step 250.
Step 190: reiteration. Steps 110 to 180 are repeated on the basis of an image / (t ₂ ) acquired at the second instant t ₂ . This makes it possible to obtain a list of three-dimensional positions ( ^x maxNmax> ^z xmaxymax) ^at time t2, as well as a signal-to-noise ratio associated with each position. Thus, at the end of step 190, there is a first list of three-dimensional positions (x _ma xNmax> ^z xmaxymax) (ti) ^{at the} first instant t ± and a second list of three-dimensional positions (x _ma xNmax> ^z xmaxymax) (t2) ^{at the} second instant t2, as well as a signal-to-noise ratio associated with each position.
Step 200: Calculation of potential displacements. During this step, potential displacements Δ are determined resulting from the comparison between each three-dimensional position at the first instant (x _m ax>ymax> ^z x _max y _max ) (. Ti) and at the second instant ( ^x maxNmax ' ^z x _max y _max XhP H θ results a list of vectors of potential displacements, the coordinates of which represent potential displacements Figure 3B represents vectors of displacements whose coordinates are indicated according to the X axis (abscissa axis) and the axis Z (ordinate axis). Each displacement vector corresponds to a couple comprising a three-dimensional position of a particle (x _max , y _max , z _Xmax y _max ) (ti) ' ^{at the} first instant, chosen from the first list and a other position of a three-dimensional particle (jcmax>ymax> ^z xmaxy _max ) (t2) ' ^{at the} second instant, chosen from the second list. A first sort is made, based on a minimum displacement and maximum displacement according to each he axis, as well as on the basis of a criterion relating to the signal-to-noise ratio assigned to each three-dimensional position: the signal-to-noise ratio SNR (x _max , y _max ) of the position at the first instant must correspond to the signal-to-noise ratio assigned to the position at the second instant to the nearest uncertainty.
Step 210: Taking into account a displacement model mod. This is based on a knowledge of the kinematic parameters of the displacement of the particles 10a in the fluid chamber 15. For example, the medium 10b in which the particles 10a evolve is in motion in the fluid chamber 15, the medium 10b carrying the particles. The movement of the medium 10b can be modeled, the particles being considered as following the movement of the medium, at least in one plane. For example, when the fluid chamber 15 is horizontal, the particles are assumed to follow the pattern of displacement in the horizontal plane, to the nearest fluctuation corresponding to a movement of the particles in a vertical plane, the latter being due to gravity and dependent on the mass of the particles.
Taking into account the displacement model mod makes it possible to define a displacement range, extending between a first terminal and a second terminal. The displacement range defines the coordinates of the possible displacement vectors taking into account the displacement model adopted. Potential movements outside the travel range are invalidated.
The displacement model can be a parametric model, the parameters of which are adjusted experimentally, based on a statistical processing of the displacements detected on a series of image acquisitions. FIG. 3B represents for example in the form of a point cloud all the displacements obtained following an analysis of a series of 500 image acquisitions. Each displacement is represented by a circle whose abscissa is the component Δχ of the displacement along the longitudinal axis X and whose ordinate is the coordinate of the reconstruction plane Zj corresponding to the position of the particle at the start of the displacement. The points having the same ordinate correspond to displacements whose starting position is situated in a plane parallel to the image sensor situated at the distance Zj from the latter. Taking into account multiple image acquisitions makes it possible to constitute displacement-related data, the statistics of which are sufficient to determine or adjust the parameters of the model. The point cloud clearly presents a zone of high density which has the shape of a boomerang.
At the center of the fluid chamber 15 (zj close to 35), the displacements have a maximum amplitude. At the edges of the fluid chamber 15 (zj close to 0 or zj close to 60), the displacements are smaller, due to the presence of the walls of the fluid chamber. Thus, preferably, the displacement model is three-dimensional, so as to take into account a distribution of flow speed of the fluid in a transverse plane YZ perpendicular to the flow axis X of the fluid, in particular because edge effects resulting from the walls of the fluid chamber 15.
In this example, the boomerang shape is modeled by a polynomial of degree 3. The coefficients of this polynomial can be determined by a quadratic fit with respect to the measured data. It is thus possible to determine or refine the parameters of the model, on the basis of the images acquired. Thus, one is based on a parametric displacement model, the parameters of the model being able to be determined or updated by experimental measurements.
In FIG. 3B, a range has also been represented corresponding to an accepted tolerance with respect to the model, which is the zone comprised between the curves M1 and M2.
Step 220: validation of trips.
During step 220, the potential displacements Δ determined in step 200 are compared with the displacement range defined in step 210. The displacements not included in the displacement range are considered to be invalid and are eliminated . The displacements Δ _ν included in the range are validated. In the example of FIG. 3B, the displacement range is defined according to a plane (Ax, zj), in which case the validation is carried out on the basis of a projection of each potential displacement vector along this plane.
Step 230: definition of the positions and / or the number of particles corresponding to valid displacements.
Each displacement Δ _ν validated during step 220 makes it possible to define a position of a particle at the first instant and a position of a particle at the second instant. One then determines a list of positions (x, y, z) (tf) validated of particles at the first instant and a list positions (x, y, z) (t ₂ ) of particles validated at the second instant. This list is made by considering that at the first instant and at the second instant, a particle is associated with only one displacement. Each list thus obtained makes it possible to estimate a position of the particles at the first instant, as well as a position of the particles at the second instant, as well as the number N of particles 10a circulating in the fluid chamber 15.
Preferably, to validate the position of a particle at an instant L we consider 3 different instants, for example three successive instants t ^, f and t _{i + 1} . The instant L represents a so-called current instant, the instants and t _{i + i} being instants respectively before and after the current instant. From the displacements Av ^ t ^, tj) validated between tj ^ and iy we establish a first list of pairs of positions between the instants tj ^ and f. From displacements Δ _ν (ί ;, t _{i + 1} ) validated between and t _{i + 1} , we establish a second list of pairs of positions between the instants and t _{i + 1} . The list of particles at the current time is obtained by performing the union of the first list and the second list, the duplicates being eliminated.
Step 250: optimization of the threshold
A parameter which may be important for the implementation of the method is the threshold S used during step 180, to select or not particles positions. This threshold conditions the number of particles considered to establish the potential displacements. FIG. 3C represents an evolution of the number of particles counted, by implementing the method described above, the signal to noise ratio of which is greater than the value of the abscissa. Such a representation allows a posteriori modification of the threshold, by fixing for example the value of the threshold at an optimal value corresponding to a flattest part of the curve. The inventors consider that an optimal threshold corresponds to a flattest part of the curve, that is to say to a weak derivative, the derivative being calculated with respect to the value of the threshold. In the example shown in FIG. 3C, and described below, the optimal value of the threshold, by implementing the method, is 2.2 or 2.3. It is therefore possible to remove the posteriori the particles of which having a signal to noise ratio below the threshold.
By way of comparison, the figure also shows an evolution of the number of particles N 'counted without considering a displacement, that is to say based only on an image acquired at a given instant. It is observed that taking account of displacements makes it possible to reduce the number of enumerated particles, in particular when the threshold is low.
During a first experimental test, a vertical chamber, as shown in FIG. 3A, was used. The test was carried out in a configuration as described in FIG. 1, the axis X, along which the particles propagate, being vertical and oriented downwards.
The sample consists of polystyrene particles with a diameter of 1 μm circulating in an air flow. The experimental parameters are as follows:
Fluid chamber: Starna type 45-F: internal dimensions of 5 x 10 x 45 mm.
Light source: CiviLaser laser diode - 405 nm - pulse duration 100 ps.
- Image sensor: CMOS MIGHTEX BTN-B050-U - 2592 x 1944 pixels of size 2.2 pm x 2.2 pm.
Acquisition frequency: 10 Hz.
64 reconstruction planes were used, corresponding to distances from the image sensor, regularly spaced between 1.5 mm and 7.8 mm.
At each instant we determine a list of particles, of coordinates (x, y, z). Being in the case of weak signals, the detection is made by favoring the detection of a large fraction of the particles with the disadvantage of having many false detections.
From the positions of the particles at two successive instants, the potential displacements Δ are determined, the latter being represented in the form of circles, having a Δχ coordinate along the X axis, a Δζ coordinate along the Z axis and a Δγ coordinate along the Y axis. The potential displacements were obtained by taking into account the following sorting criteria: 0 <Δχ <2.2 mm; 0 <Δγ <66 pm; 0 <Δζ <200 pm.
FIG. 3B illustrates these potential displacements in the form of a cloud of points in the plane (Δχ, Z). A displacement model has been taken into account, forming bounds represented by the curves M1 and M2 plotted in FIG. 3B. The displacements located between these curves have been validated.
From the validated displacements Δν, the number N of particles has been counted, as a function of the signal to noise ratio threshold considered during step 180, the evolution of the number N of particles counted as a function of the signal ratio threshold S on noise being shown in Figure 3C. This figure also represents a number of particles N ′ as a function of the signal-to-noise ratio threshold, without taking displacement into account. It can be seen that beyond a certain threshold, the estimate without taking the displacement into account is reliable.
According to a second embodiment, the sample is illuminated by two pulses respectively at a first instant t _± and at a second instant t ₂ , and an image I is acquired whose acquisition time includes the first instant and the second instant . Thus, on the same image, a signal representative of the positions of the particles is obtained at the first and at the second instant. The steps of this embodiment are shown in FIG. 4A, and described below:
Step 300: successive illumination of the sample at the first instant and at the second instant, and acquisition of an image I, called the first image, during the first instant and during the second instant. The time interval between the two instants can be very short, for example 5 ms.
Step 320: frequency filtering, analogously to step 120.
Step 330: propagation of the filtered image, analogously to step 130, to obtain a stack of complex images
Step 340: extraction of a component from each complex image from the complex image stack.
Step 345: digital focusing, analogously to step 145.
Step 350: formation of an image of the maxima from the acquired image, in a similar manner to step 150.
Step 360: search for local maxima in the image of the maxima, similar to step 160.
Step 370: taking into account the signal-to-noise ratio, analogously to step 170. This step makes it possible to establish a list of radial coordinates (x _max , y _max ) corresponding to a local maximum in the image of the maxima , each pair of radial coordinates being associated with a transverse coordinate z _rv . We then have a list of three-dimensional positions (x _max , y _m ax> ^z xmaxymax ^> ) in the sample likely to contain a particle. Each three-dimensional position (% max, y _max , is associated with an estimate of the signal-to-noise ratio SNR (x _max , y _max ) of the image I _max at the position ( ^X max> ymax) · Unlike the first mode embodiment, each of these three-dimensional positions is likely to be occupied by a particle 10a at the first instant ^ or at the second instant t ₂ .
Step 380: thresholding as a function of a signal-to-noise ratio threshold, in a similar manner to step 180. Only the three-dimensional positions whose associated signal-to-noise ratio is greater than the threshold value are retained, the others being considered as not representative of particles.
Step 400: Calculation of potential displacements. During this step, potential displacements Δ are determined resulting from the comparison between each three-dimensional position obtained following step 380. This results in a list of vectors of potential displacements, the coordinates of which represent potential displacements. FIG. 4B represents potential displacements obtained following a second experimental test described below. This step can take into account a sorting criterion, based on a minimum displacement, and therefore a minimum spacing between two positions. Furthermore, the sorting criterion can also take into account the fact that the particles move, in one direction, in a predetermined direction. For example, along the X axis, we consider that Δχπιίη <Δχ <Δχτηαχ, with Δχτηίη> 0. This takes into account the fact that the three-dimensional positions of the acquired image are likely to correspond to positions of particles at the first instant. or at the second instant.
Step 410: taking into account a displacement model, similar to step 210.
Step 420: validation of displacements, on the basis of a displacement model, as described in connection with step 220. In FIG. 4B, a displacement model (curve M3) has been represented.
Step 430: definition of the positions and / or the number of particles corresponding to displacements validated during step 420.
According to this second embodiment, the method may include a step 450 of adjusting the signal-to-noise ratio threshold used, in a similar manner to step 250 previously described.
An advantage of this embodiment is to avoid the use of image sensors having too high an acquisition frequency. For example, when the time interval between the first instant and the second instant is 5 ms, the first embodiment, based on an acquisition of two successive images, would impose an acquisition rate of 200 images per second, which is not within the reach of usual image sensors. This embodiment is therefore suitable for particles exhibiting high speeds.
This embodiment was the subject of a second experimental test, the particles being polystyrene balls with a diameter of 2 μιτι moving in the air. FIG. 4B represents the displacements as well as a modeled border.
A limitation of this embodiment is that it only takes into account the particles present in the observation field of the image sensor at the two instants considered. The inventors have estimated that by applying a weighting factor to each displacement detected, the number of particles counted is more reliable. The weighting factor for each displacement A _fc , is determined according to a probabilistic approach. The probability of detection p _k of coordinates AX _k , AY _k is such that: p _k = ^ ^LX _gt ^ γ denote | _he dimensions of the field observed by the detector 20 in the fluid chamber 15, respectively along the X axis and the Y axis.
If K denotes the number of validated displacements A _k , each displacement having the coordinates AX _k and AY _k , the number of particles in the fluid chamber can be estimated _Dar Λ7 ₌ yk = K _J_ _ yk = K ______ LXxLY______ ^p 'Pk (Ax- lAXfci) x (ir- | âr _fc |)
This only remains valid if | <LX or if | <LY
We now describe a variant that can be applied to each embodiment, from the list of potential displacements A. This list is obtained at the end of step 200 of the first embodiment or of step 400 of second embodiment. According to this variant, the particles circulating in the fluid chamber are of different types, for example of different masses. Therefore, each type of particle can have a displacement, called particle displacement, relative to the fluid, which is specific to it. The particle displacement can be induced by a property of the particle, conditioning the displacement of the latter with respect to the fluid. The particle then moves in the fluid under the effect of a force depending on said property, for example under the effect of a field to which the particle is subjected. It may for example be an electric or magnetic field, in which case a particle is subjected to a force depending on its charge. It can also act from a gravitational field, in which case the particle moves relative to the fluid as a function of its mass. Thus, one can define a particle displacement model of particles with respect to the fluid, one parameter of which is said property of the particle. The particle displacement of each particle is preferably oriented in an orientation not parallel to the direction of flow of the fluid, but this condition is not necessary. It is optimal that the particle displacement is perpendicular to the direction of flow of the fluid. By applying the particle displacement model to the previously validated three-dimensional displacements Δ, one can determine the property of the particle forming a parameter of the particle displacement model. It is then possible to classify the particles according to their property and to count the particles according to a value of said property.
One can for example take into account a model of particle displacement corresponding to a predetermined value of the property. Then we determine, for each particle, a deviation ε from this model. One can then classify the particles according to the deviation ε, compared to the model of particle displacement, which was allotted to them. The particles are then classified according to their particle displacement. The particles for which the deviation is zero have a property corresponding to the predetermined value. The property of the other particles depends on the difference calculated for each of them.
A third experimental trial was carried out to implement this variant, using polystyrene beads of diameter lpm and diameter 2pm. The fluid chamber was kept arranged so that the particles were entrained by air circulating horizontally, the flow axis X being horizontal. The experimental device is shown in Figure IA, the XZ plane being a horizontal plane. The particles were entrained horizontally by the carrier fluid, in this case air, along the horizontal axis X. They were subjected to the effect of gravity, along a vertical axis Y, perpendicular to the flow axis . The dimensions of the fluid chamber 15 were 10 mm and 20 mm respectively along the axes Z and Y. The fluid chamber had, in a plane YZ, a rectangular section of dimensions 9.6 mm x 20 mm.
We can show that if Δί = t ₂ - t _± , a variation of the displacement Δ7 along the Y axis, is such that: Δ7 = K (p _b dl - Padfobt, with:
p _b : density of the second type of particle;
d _b : diameter of the second type of particle;
p _a : density of the first type of particles;
d _a : diameter of the first type of particles;
K is a constant, equal to 34.7.
In this example, the property of each particle considered is the aerodynamic diameter, corresponding to the product of the diameter of a particle by the square root of its density.
For an acquisition frequency of 10 Hz or 4 Hz, ΔΚ is respectively equal to 12.4 and 31 pm, or 5.6 and 14.1 pixels, for the first type and the second type of particles.
On each reconstructed image, the particles of diameter 2pm appear more clearly than the particles of diameter lpm: thus, the signal to noise ratio corresponding to the particles of large diameter is greater than the signal to noise ratio corresponding to the particles of small diameter.
When establishing the potential displacements Δ (step 200), a displacement is considered to be potential when the signal-to-noise ratio associated with the two positions, defining the displacement, is close. It is then possible to assign a signal to noise ratio Sa to each displacement Δ, this ratio being obtained by an average of the signal to noise ratios respectively associated with each position forming the displacement. The signal-to-noise ratio Sa of the displacements of the first type of particle (particles of diameter lpm) is lower than the signal-to-noise ratio of the displacements of the second type of particles (particles of diameter 2 pm). Furthermore, the displacement, along the vertical axis Y, of the first type of particles is less than the displacement, along the same axis, of the second type of particles.
We subtracted, at each displacement ΔΥ determined along the Y axis, the particle displacement modeled for the second type of particles.
FIG. 5 represents the displacements ΔΥ of each particle after the subtraction of the particle displacement model, as a function of the signal to noise ratio assigned to each displacement. The acquisition frequency is 10 Hz. Each point in Figure 5 represents a validated displacement. We observe a segmentation of displacements:
the displacements corresponding to the second type of particles are grouped according to a group Gb, centered on the coordinate ΔΥ = 0.
the displacements corresponding to the first type of particles are grouped according to a group Ga, extending around the coordinate ΔΥ = 12.5 pm.
We also observe that the displacements associated with the first type of particle have a signal to noise ratio Sa lower than the displacements associated with the second type of particles.
This variant makes it possible to count particles according to a property, of mass, charge, aerodynamic diameter type. It can also be used to discriminate bacteria according to their motility. We can then discriminate bacteria of the Staphylococcus type (non-motile which follow the fluid) from bacteria of the Escherichia coli type (motile, which move relative to the fluid).
In the embodiments described above, the images are acquired by an image sensor 20 placed in a lensless imaging configuration, no image forming optics being disposed between the image sensor and the fluid chamber. Indeed, such a device allows a determination of three-dimensional positions of particles using a two-dimensional image sensor, by implementing inexpensive instrumentation. Such a device is therefore particularly suitable for implementing the invention. However, the invention applies to other imaging configurations making it possible to obtain positions of particles at two successive instants, and in particular three-dimensional positions. The embodiments described above apply to a defocused image sensor, forming a defocused image of the sample, according to the known principle of digital holographic microscopes. The advantage is to be able to observe smaller particles, to the detriment of a reduced field of observation. It is also possible to obtain the three-dimensional positions of particles by other imaging methods, using several image sensors. These sensors can for example extend parallel to each other, the three-dimensional position of the particles being obtained by stereovision. Two sensors extending along different planes, for example perpendicular to one another, are also conceivable.
The invention may be applied to the detection of solid particles in the air, for example pollutants or dust, but also to the detection of particles, in particular biological particles, in a liquid. It will find applications in applications related to fluid control, for industry, the environment, health or the food industry.

权利要求:
Claims (22)
[1" id="c-fr-0001]
1. Method for counting particles (10a) in motion in a fluid (10b), circulating in a fluid chamber (15), the method comprising the following steps:
a) arrangement of the fluid chamber (15) between a light source (11) and an image sensor (20), the image sensor extending along a detection plane (P _o );
b) illumination of the fluid chamber by the light source, the light source emitting an incident light wave (12) propagating along a propagation axis (Z), and acquisition, by the image sensor (20), d a first image (/, / (^)), representative of an exposure wave (14) to which the image sensor (20) is exposed, the image sensor comprising different pixels, with each pixel being associated a radial coordinate (%, y) in the detection plane (P _o );
c) from the first acquired image, obtaining three-dimensional coordinates ((x, y, z) (ti)) of particles, in the fluid chamber, at a first instant (tffi
d) obtaining three-dimensional coordinates ((%, y, z) (t ₂ )) of particles in the fluid chamber at a second time (t ₂ ), after the first time;
e) from the coordinates of the particles obtained at the first instant and at the second instant, determination of potential displacements (Δ) of the particles between said instants;
f) taking into account a displacement model (mod) of the fluid in the fluid chamber;
g) from the fluid displacement model considered during step f), validation of displacements among the potential displacements (Δ) calculated during step e);
h) from validated displacements (Δ _ν ) during step g), determination of a number (N) of particles and / or of the coordinates of the particles at the first instant and / or at the second instant.
[2" id="c-fr-0002]
2. Method according to claim 1, in which step c) comprises:
obtaining a first image of interest (/ _v (ti)) from the first acquired image (1.1 (tf)) during step b), and applying an operator of digital propagation (h) to the first image of interest, along at least one reconstruction distance (z _; ), along the propagation axis, so as to obtain at least one complex reconstructed image (Azj);
from each reconstructed complex image (Azj), obtaining radial coordinates (x, y) of particles in the fluid chamber at the first instant.
[3" id="c-fr-0003]
3. Method according to claim 1 or claim 2 wherein step c) comprises the following substeps:
ci) obtaining a first image of interest (/ _v (ti)) from the first image acquired (I, I (ti)) during step b), and applying a digital propagation operator (h) at the first image of interest along a plurality of reconstruction distances (z _; ), along the propagation axis (Z), so as to obtain a first stack of complex reconstructed images, called the first stack images, comprising as many reconstructed complex images (Azj) as reconstruction distances, each reconstructed complex image (Azj) being representative of an exposure light wave (14) to which the image sensor is exposed ( 20);
cii) for at least one radial coordinate (x, y) defined by the first image of interest, determination of a reconstruction distance (z _xy ) maximizing the evolution of a component (comp ^ A _Zj . (x, y) ^, —Re ^ A _Zj . (x, y) ^ of each complex image (Azj) forming the first stack of images, along an axis parallel to the axis of propagation and passing through said radial coordinate , said reconstruction distance thus determined (z _xy ) forming a transverse coordinate associated with said radial coordinate, the value of the component calculated at said reconstruction distance being a so-called maximum value A _max (x, y) associated with said radial coordinate ( x, y), the sub-step cii) being carried out for all or part of the radial coordinates associated with the pixels of the first image of interest;
ciii) establishment of a list of three-dimensional positions, each three-dimensional position comprising a radial coordinate (x, y) and the associated transverse coordinate (z _xy ), determined during the sub-step cii), with each three-dimensional position being associated the maximum value (A _max (x, y)) obtained during the sub-step cii);
civ) selection of three-dimensional positions (x, y, z _xy ) according to the maximum value (^ max (^ y)) associated with them.
[4" id="c-fr-0004]
4. Method according to claim 2 or claim 3, in which the first image of interest (ZvCti)) is:
the first image (/ (tj) acquired during step b);
or the first image (Z (t _x )) acquired during step b), from which an image of the fluid chamber is subtracted, acquired by the image sensor, before or after the acquisition of the first image , the subtraction being weighted by a weighting term;
or the first image acquired (/ (^)) during step b), from which an average of images acquired respectively before and after the acquisition of the first image is subtracted.
[5" id="c-fr-0005]
5. Method according to any one of claims 3 or 4, in which during the sub-step cii), the component (comp ^ A _z . (X, y) ^ considered comprises the real part, or the imaginary part, or the module, or the phase of each complex image (Azj) forming the image stack.
[6" id="c-fr-0006]
6. Method according to any one of claims 3 to 5, in which the civ sub-step comprises:
an image formation (A _max ), called the first image of the maxima, each pixel of which is associated with a three-dimensional position (x, y, z _xy ) determined during the sub-step ciii), and is assigned the maximum value ( A _max (x, y)) determined, during the sub-step ciii), for said three-dimensional position;
a selection, in the first image of the maxima, of pixels (% _max , y _max ) whose value (A _max (x, y)) is maximum in a neighborhood area defined around each pixel;
a calculation, for each pixel selected, of a signal to noise ratio (SNR (x _max , y _max )) as a function of said maximum value (A _max (x, y ')) and of the pixel value of the first image of the maxima (A _max ) located in a calculation area extending around said pixel;
so that each three-dimensional position is selected according to the signal to noise ratio calculated for the pixel of the first image of the maxima associated with it.
[7" id="c-fr-0007]
7. Method according to any one of the preceding claims, in which step d) comprises an acquisition, by the image sensor (20), of a second image (I (t ₂ )), each pixel of which is associated with a radial coordinate (x, y) in the detection plane (P _o ).
[8" id="c-fr-0008]
8. Method according to claim 7, in which step d) comprises the following sub-steps:
di) obtaining a second image of interest (/ _v (t ₂ )) from the second acquired image (/ (t ₂ )) and application of a digital propagation operator (h) to the second image d interest, according to a plurality of reconstruction distances, along the propagation axis (Z), so as to obtain a second stack of reconstructed complex images (rice ₇ ), called the second stack of images, comprising as many complex images reconstructed only for reconstruction distances (z _; ), each reconstructed complex image being representative of an exposure light wave (14) to which the image sensor (20) is exposed at the second instant;
dii) for at least one radial coordinate (x, y) defined the second image of interest, determination of a reconstruction distance (z _xy ) maximizing the evolution of a component of each complex image forming the second stack of images, along an axis parallel to the propagation axis and passing through said radial coordinate, said reconstruction distance forming a transverse coordinate associated with said radial coordinate, the value of the component calculated at said reconstruction distance being a value said maximum associated with said radial coordinate, the sub-step dii) being carried out for all or part of the radial coordinates associated with the pixels of the second image of interest;
diii) establishment of a list of three-dimensional positions, each three-dimensional position comprising a radial coordinate (x, y) as well as the associated transverse coordinate, determined during sub-step dii), with each three-dimensional position being associated the maximum value obtained during sub-step dii);
div) selection of three-dimensional positions according to the maximum value associated with them.
[9" id="c-fr-0009]
9. Method according to claim 8, in which during the sub-step step di) the second image of interest (/ _v (t ₂ )) is ^: the second acquired image (/ (t ₂ ));
or the second acquired image (/ (t ₂ )), from which an image of the fluidic chamber is subtracted, acquired by the image sensor, before or after the acquisition of the second image, the subtraction being weighted by a weighting term, which can be between 0 and 1;
or the second acquired image (/ (t ₂ )), from which an average of images acquired respectively before and after the acquisition of the second image is subtracted.
[10" id="c-fr-0010]
10. Method according to any one of claims 8 or 9, in which during the sub-step dii), the component comprises the real part, or the imaginary part, or the module, or the phase of each complex image forming the stack of images.
[11" id="c-fr-0011]
11. Method according to any one of claims 8 to 10, in which the sub-step div) comprises:
an image formation (A _max ), called the second maxima image, each pixel of which is associated with a three-dimensional position (x, y, z _xy ) determined during the substep diii), and is assigned the maximum value ( A _max (x, y)) determined, during sub-step diii), for said three-dimensional position;
a selection, in the second image of the maxima, of pixels (x _max , y _max ) whose value (A _max (x, y)) is maximum in a neighborhood area defined around each pixel;
a calculation, for each selected pixel, of a signal to noise ratio (SNR (x _max , y _max )) as a function of said maximum value (ri _max (x, y)) and of the pixel value of the second image maxima (A _max ) located in a calculation area extending around said pixel;
so that each three-dimensional position is selected as a function of the signal-to-noise ratio calculated for the pixel of the second image of the maxima associated with it.
[12" id="c-fr-0012]
12. Method according to any one of claims 1 to 5, in which:
step b) comprises two successive illuminations of the fluid chamber by the light source, at the first instant and at the second instant, so that the first image (/) represents the wave of exposure (14) to each of the moments;
steps c) and d) are combined in the same step of obtaining the coordinates of particles at the first instant and at the second instant.
[13" id="c-fr-0013]
13. Method according to any one of the preceding claims, in which step e) comprises a comparison of the coordinates of the particles in the fluid chamber determined at the first instant and at the second instant, so as to establish a list of potential displacements (Δ ) particles between said instants.
[14" id="c-fr-0014]
14. Method according to any one of the preceding claims, in which step g) comprises a determination of a range of displacements using the displacement model taken into account during step f), the potential displacements. being validated when they are included in said displacement range.
[15" id="c-fr-0015]
15. Method according to any one of the preceding claims, in which step g) comprises a subtraction, at each potential displacement, of a displacement according to the displacement model (mod).
[16" id="c-fr-0016]
16. Method according to any one of the preceding claims, in which:
the fluid comprises particles, each particle having a property and moving, with respect to the fluid, according to a displacement model, called the particle displacement model;
the particle displacement model depends on said property of the particles;
the method comprising, from displacements validated during step g), a step i) of taking into account at least one model of particulate displacement, so as to count the particles as a function of a value of said property .
[17" id="c-fr-0017]
17. The method of claim 15, wherein the property is a mass or an electric charge, or an ability to move in the fluid.
[18" id="c-fr-0018]
18. Method according to any one of claims 16 or 17, comprising:
taking into account a particular displacement model for a predetermined value of the property;
a calculation of deviations of the displacements of each particle with respect to said particle displacement model;
so that the property of each particle is determined according to said deviations as well as said predetermined value of the property.
[19" id="c-fr-0019]
19. Method according to any one of claims 16 to 18, in which the fluid flows in a direction of flow, and in which the particulate movement takes place in another direction not parallel to said direction of flow.
[20" id="c-fr-0020]
20. Method according to any one of the preceding claims, in which no image-forming optics are arranged between the image sensor and the fluid chamber.
[21" id="c-fr-0021]
21. Method according to any one of claims 1 to 19, in which the image sensor comprises an image-forming optic between the image sensor and the fluid chamber, the image formed during step b ) being a defocused image.
[22" id="c-fr-0022]
22. Device for counting particles circulating in a fluid chamber, the device comprising:
a light source (11) configured to illuminate the fluid chamber (15);
an image sensor (20) extending in a detection plane (P _o ), the fluid chamber being interposed between the image sensor and the light source, the image sensor being configured to acquire at least one image of the fluid chamber illuminated by the light source;
5 - the device comprising a processor (30) capable of implementing steps c) to h) of a method according to any one of claims 1 to 21 from at least one image acquired by the sensor picture.
1/6

类似技术:

---

公开号 | 公开日 | 专利标题

---

EP3274689B1|2022-02-16|Method and apparatsu for analysing particles

---

EP3274694B1|2019-12-18|Method for determining the state of a cell

---

EP3234550B1|2022-01-05|Method for identifying biological particles using stacks of defocused holographic images

---

EP3343201A1|2018-07-04|Method for counting and characterising particles in a moving fluid

---

EP2420822A2|2012-02-22|Method and device intended for obtaining three-dimensional images of a sample by microscopy

---

EP2556164B1|2018-05-30|Method for detecting clusters of biological particles

---

FR3049348A1|2017-09-29|METHOD OF CHARACTERIZATION OF A PARTICLE IN A SAMPLE

---

EP2909671B1|2016-12-07|Method for designing a single-path imager able to estimate the depth of field

---

EP2356429B1|2019-06-05|Device for analyzing a polyphase mixture via a light beam backscattered by said mixture

---

WO2018060589A1|2018-04-05|Method for counting white blood cells in a sample

---

EP2912475B1|2016-09-07|Method for detecting turbulence conditions, using the interaction of a laser beam with a thin photochromic film, and device for implementing said method

---

EP3640743A1|2020-04-22|Method for observing a sample

---

WO2017207184A1|2017-12-07|Device and method for acquiring a particle present in a sample

---

EP3230713B1|2020-07-29|Method for obtaining an image of a sample, and associated lens-free imaging system

---

Barbu2014|Tridimensional estimation of turbulent fluid velocity

---

FR3059113A1|2018-05-25|METHOD FOR FORMING A HIGH RESOLUTION IMAGING IMAGE WITHOUT A LENS

---

FR3076618A1|2019-07-12|METHOD AND SYSTEM FOR OPTICALLY INSPECTING A SUBSTRATE

---

EP3584560A1|2019-12-25|Method for observing a sample by lens-free imaging, with acknowledgement of a spatial dispersion in a sample

---

FR3111998A1|2021-12-31|Focusing process for holographic imaging system

---

WO2019115901A1|2019-06-20|Method for calibrating an analysis device, and associated device

---

FR3090107A1|2020-06-19|Method for characterizing a particle from a hologram.

---

FR3085474A1|2020-03-06|METHOD FOR MEASURING THE LINE ROUGHNESS OF A LINE ARRAY BY X-RAY DIFFRACTION

---

WO2021037935A1|2021-03-04|Method and device for the optical characterisation of particles

---

FR3066816A1|2018-11-30|OPTICAL DEVICE FOR MEASURING THE BENDING OF A REFLECTIVE SURFACE

同族专利:

---

公开号 | 公开日

---

US10467764B2|2019-11-05|

---

FR3061297B1|2019-05-24|

---

EP3343201A1|2018-07-04|

---

US20180189963A1|2018-07-05|

引用文献:

---

公开号 | 申请日 | 公开日 | 申请人 | 专利标题

---

WO2008090330A1|2007-01-22|2008-07-31|Cancer Research Technology Ltd|Detecting objects|

---

FR3031180A1|2014-12-30|2016-07-01|Commissariat Energie Atomique|METHOD AND DEVICE FOR CONTROLLING THE POSITION OF A TRANSPARENT PARTICLE BY A PILE OF HOLOGRAPHIC IMAGES|EP3584560A1|2018-06-20|2019-12-25|Commissariat à l'Énergie Atomique et aux Énergies Alternatives|Method for observing a sample by lens-free imaging, with acknowledgement of a spatial dispersion in a sample|WO2008132995A1|2007-04-12|2008-11-06|The University Of Electro-Communications|Particle measuring device and particle size measuring device|

---

JP5478814B2|2007-06-05|2014-04-23|株式会社東芝|Ultrasonic diagnostic apparatus and ultrasonic speed measurement method|

---

US8994945B2|2011-10-27|2015-03-31|Fluid Imaging Technologies, Inc.|Method of treatment analysis with particle imaging|JP6971842B2|2017-12-28|2021-11-24|キオクシア株式会社|Measuring device and measuring method|

---

EP3839636A1|2019-12-20|2021-06-23|Imec VZW|A device for detecting particles in air|

法律状态:
2018-01-02| PLFP| Fee payment|Year of fee payment: 2 |

---

2018-06-29| PLSC| Publication of the preliminary search report|Effective date: 20180629 |

---

2019-12-31| PLFP| Fee payment|Year of fee payment: 4 |

---

2020-12-28| PLFP| Fee payment|Year of fee payment: 5 |

---

2021-12-31| PLFP| Fee payment|Year of fee payment: 6 |

优先权:

---

申请号 | 申请日 | 专利标题

---

FR1663475A|FR3061297B1|2016-12-28|2016-12-28|METHOD FOR COUNTING AND CHARACTERIZING PARTICLES IN A FLUID IN MOTION|

---

FR1663475|2016-12-28|FR1663475A| FR3061297B1|2016-12-28|2016-12-28|METHOD FOR COUNTING AND CHARACTERIZING PARTICLES IN A FLUID IN MOTION|

---

EP17210133.9A| EP3343201A1|2016-12-28|2017-12-22|Method for counting and characterising particles in a moving fluid|

---

US15/856,679| US10467764B2|2016-12-28|2017-12-28|Method for counting and characterization of particles in a fluid in movement|