• 专利附录
  • 专利说明
  • 权利要求
  • 类似技术
  • 同族专利
  • 引用文献
  • 法律状态
  • 优先权
    • 专利摘要:
    defect categorization. it is a method of categorizing defects in a media item. the method comprises the steps of: receiving a binary image of the media item, where the binary image comprises a plurality of pixels, with each pixel having a potential defect intensity or a non-defect intensity; and identifying one or more blobs that comprise contiguous pixels having a potential defect intensity. for each identified blob, the method involves comparing a blob size with a damage limit; ignore the blob if the blob size is less than the damage limit; and for each identified blob having a size exceeding or equaling the damage limit, categorize the identified blob.
    公开号:BR102012023807B1
    申请号:R102012023807-1
    申请日:2012-09-20
    公开日:2020-09-29
    发明作者:Ping Chen；Chao He；Gary Ross
    申请人:Ncr Corporation；
    IPC主号:
    专利说明:

    FIELD OF THE INVENTION
    The present invention relates to the categorization of defects. In particular, although not exclusively, the invention refers to the automatic categorization of defects in media items, such as those deposited in self-service terminals (SSTs). The invention also relates to the evaluation of a media item based on categorized defects. BACKGROUND OF THE INVENTION
    Some SSTs, such as ATMs, can receive media items in the form of bank notes (or checks) deposited by a customer.
    Some currency issuing authorities (such as the European Central Bank) have determined that banks should capture and return to that authority any currency that is considered unsuitable for continued circulation. This is relatively easy to implement when the currency is deposited at a bank teller, because the teller can physically inspect bank notes, however, it is more difficult to implement when bank notes are deposited at ATMs because no human teller is involved. SUMMARY OF THE INVENTION
    Consequently, the invention generally provides methods, systems, devices, and software to automatically categorize defects in media items.
    In addition to the Summary of the Invention provided above and the subject matter described below in the Detailed Description, the following paragraphs of this section are intended to provide an additional basis for alternative claim language for possible use during the process of this application, if necessary. If this order is granted, some aspects may refer to claims added during the process of this application, other aspects may refer to claims excluded during the process, other aspects may refer to the subject matter never claimed. In addition, the various aspects detailed below are independent of each other, except where stated otherwise. Any claim corresponding to one aspect should not be built incorporating any element or feature of the other aspects, except where explicitly stated in that claim.
    According to a first aspect, a method of categorizing defects in a media item is provided, the method comprising the steps of: receiving a binary image of the media item, where the binary image comprises a plurality of pixels, if that each pixel has a potential defect intensity or a non-defect intensity; identify one or more blobs that comprise contiguous pixels having a potential defect intensity; for each identified blob, compare a blob size with a damage limit; ignore the blob if the blob size is less than the damage limit; and for each identified blob having a size exceeding or equaling the data limit, categorize the identified blob.
    Depending on the usage in question, a blob comprises contiguous pixels in an image that have a similar property (for example, intensity) that is different from the corresponding property (in this example, intensity) of the surrounding pixels.
    The two intensity values can represent high intensity and low intensity, respectively. The potential defect intensity can comprise a high intensity if the transmission is used to create the image; alternatively, the potential defect intensity may comprise a low intensity if reflectance is used to create the image.
    The step of identifying one or more blobs may include using a region growth algorithm (an example of which is described in IEEE Region Growing: A New Approach SA Hojjatoleslami and J. Kittier TRANSACTIONS ON IMAGE PROCESSING, VOL. 7, NO. 7, JULY 1998 pages 1079 - 1084), a splitting and joining algorithm (an example of which is described in SL Horowitz and T. Pavlidis, Picture Segmentation by a Directed Split and Merge Procedure, Proc. ICPR, 1974, Denmark, pp.424 -433) or any other convenient algorithm. These two algorithms are computationally intensive.
    Advantageously, the step of identifying one or more may comprise a modified disjoint set structure algorithm.
    The modified disjunct set structure algorithm can comprise: identify in each row of the binary image, each group of pixels that is contiguous and that has a potential defect intensity; identify for each row limit, (i) each group in an upper row that at least partially overlaps with a group in the lower row, and join these two groups into a new single group as a growing blob, (ii) each group in a top row that does not overlap with a group in the bottom row, and characterizes each group in the top row as a complete blob, and (iii) each group in the bottom row that does not overlap a top row, and characterizes each such group as a growing blob. When the modified disjoint set structure algorithm is complete, then one or more complete blobs can be presented. Preferably, for each complete blob, the blob location, size and dimensions are stored.
    The step of categorizing the identified blob can include the categories of: tear, missing portion (including a missing corner), corner fold, and hole.
    The step of categorizing the identified blob can comprise the additional step of: (a) categorizing the blob identified as a hole if the blob does not touch one of the edges of the media item. If an identified blob has no connection to an edge of the media item, it can be categorized as a hole.
    The step of categorizing the identified blob can comprise the additional step of: (b) if the identified blob does not touch one of the edges of the media item, however, if it is not in one of the corners of the media item, then it is categorized the blob identified as a tear / missing portion. Whether the identified blob is a tear or a missing portion depends on the size of the identified blob, so the step of categorizing the identified blob can comprise the additional step of comparing the size of the identified blob with a maximum tear size to categorize the identified blob as a tear or as a missing portion.
    The step of categorizing the identified blob can comprise the additional step of: (c) if the identified blob touches one of the corners of the media item, then it is verified whether the identified blob is a corner fold.
    The step of verifying that the identified blob is a corner fold may include comparing the pixel intensities from a potential corner overlap region with the pixel intensities from a neighboring non-overlap region.
    The step of categorizing the identified blob can comprise the additional step of: (d) if the identified blob touches one of the corners of the media item, however, if it is not a corner fold, then the identified blob is categorized as a faulty portion, or, more precisely, as a faulty corner, if necessary.
    The step of categorizing the identified blob can comprise the additional step of: (e) if the identified blob touches one of the corners of the media item and is a corner fold, then the identified blob is categorized as a corner fold.
    The method can comprise the additional step of applying adaptation rules to categorized blobs to see if the media item should be classified as inappropriate.
    Adaptation rules can comprise a different set of rules for each defect category. For example, a media item can be rejected as inappropriate if it has a hole greater than or equal to ten square millimeters (10 mm2); or if a corner fold has a shorter edge greater than or equal to ten millimeters (10 mm) and its area is greater than or equal to 130 mm2.
    The method can comprise the additional steps of capturing an image of the media item and then creating a binary image of the captured image before the step of receiving a binary image of the media item.
    The step of capturing an image of the media item may further comprise capturing a broadcast image of the media item. A transmission image can be captured using an electromagnetic radiation transmitter on one side of the media item and an electromagnetic radiation detector on the opposite side of the media item. In one embodiment, the electromagnetic radiation used is infrared radiation.
    The step of capturing an image of the media item may include using eight bits to record the intensity value for each pixel (providing a range of intensity values from 0 to 255). Alternatively, any convenient number of bits can be used, such as 16 bits, which would provide a range of intensity values between 0 and 65535.
    The method can comprise the additional step of adjusting spatial dimensions of the received image in such a way that the received image is compatible with the spatial dimensions of a reference for such media item. This compensates for any media items that have portions added (such as tape) or have shrunk or expanded, or something like that. Techniques for automatically aligning a captured image with a reference image, and then cropping or adding to the captured image to satisfy the spatial dimensions of the reference image are well known in the art.
    According to a second aspect, an operable media validator is provided to categorize defects on a presented media item, the media validator comprising: a transport of a media item that serves to transport a media item; an image capture device aligned with the transport of the media item and capture a two-dimensional array of pixels corresponding to the media item, each pixel having a pixel intensity referring to a property of the media item in a spatial location on the item of media corresponding to that pixel; and a processor programmed to control the media transport and the image capture device, and also programmed to: (a) identify one or more blobs that comprise contiguous pixels having a potential defect intensity; (b) for each identified blob, compare a blob size with a damage limit; (c) ignore the blob if the blob size is less than the damage limit; and (d) for each identified blob having a size not less than the damage limit, categorize the identified blob.
    The processor can also implement the additional steps mentioned in relation to the first aspect.
    The media item transport may comprise one or more between infinite tracks, sliding plates, rollers, or the like.
    The image capture device may comprise a two-dimensional sensor, such as a CCD contact image sensor (CIS), which has a sensor area at least as large as the media item area. This allows a complete two-dimensional image to be captured at one point in time. Alternatively, the image capture device may comprise a linear sensor (covering one dimension of the media item, but not both dimensions) that captures a strip of the media item as the media item passes through the linear sensor, in order to such that, once the entire media item has passed through the linear sensor, then a complete two-dimensional image of the media item can be constructed from the sequence of images captured by the linear sensor. This would allow a lower cost sensor to be used, because a smaller catchment area (just as large as a dimension of the media item) would be sufficient.
    The image capture device may further comprise a light source. The light source may comprise an infrared radiation source.
    The image capture device can be located on the opposite side of the media item (the opposite side of the media item's path when none in the media item is present) to the light source in such a way that a broadcast image is captured. Alternatively, however, less advantageously, the image capture device may be located on the same side as the media item of the light source such that a reflectance image is captured.
    The media validator can comprise a banknote validator. The banknote validator can be incorporated into a media deposit, which can be incorporated into a self-service terminal, such as an ATM.
    According to a third aspect, a computer programmed to implement the steps of the first aspect is provided. The computer program can be executed by a media validator.
    According to a fourth aspect, a defect profile configuration file is provided, the file comprising: a defect type parameter; a defect size field; and a logical parameter.
    The type of defect can comprise a type of no defect in a media item. A type of absence of defect occurs when there is no substrate in a portion of the media item where there should be a substrate.
    The defect configuration file can comprise a plurality of defect size parameters within the defect size field.
    Each defect size parameter can comprise: a defect length parameter, a defect width parameter, or a defect area parameter.
    The logical parameter can be used to indicate how the defect size parameters are associated. For example, if a defect type parameter is a “missing portion”, and if a defect length is 6 mm and a defect width is 5 mm, and the logical parameter is “OR”, then a media item it will be identified as inappropriate if it has an identified blob categorized as a missing portion having a width exceeding or equaling 5 mm or a length exceeding or equaling 6 mm.
    According to a fifth aspect, a method is provided to characterize a media item as inappropriate, the method comprising: implementing the method of the first aspect to categorize defects in the media item; for each categorized defect, access a defect profile configuration file to restore (i) a defect type compatible with the categorized defect, and (ii) a defect size associated with that defect type; and characterize the media item as inappropriate if a categorized defect includes a dimension greater than or equal to a defect size for that type of defect.
    According to a sixth aspect, a programmed media validator is provided to implement the fifth aspect.
    Preferably, the media validator implements additional media item processing functions, such as media item recognition, validation, spot detection, wear detection, foreign matter detection, and the like.
    The media item can comprise a bank note.
    For reasons of clarity and simplicity of description, not all combinations of elements provided in the aspects mentioned above were expressly presented. Despite this, individuals versed in the technique will recognize directly and unambiguously that, unless it is technically impossible, or explicitly stated to the contrary, the consistency clauses referring to an aspect are destined to apply, with the necessary changes, as optional features of each other aspect to which these consistency clauses could possibly relate.
    These and other aspects will become apparent from the specific description below, given by way of example, with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS
    Figure 1 is a schematic diagram of a defect categorization system that comprises a media validator that serves to implement a method of categorizing defects in a media item inserted in it according to one embodiment of the present invention;
    Figure 2 (divided into two sheets in the drawings) is a flowchart that illustrates the steps taken by the media validator in Figure 1 to capture and process an image of a media item as part of the categorization of defects and adaptation assessment of media;
    Figure 3 (divided into two sheets in the drawings) is a flow chart that illustrates the steps implemented by the media validator in Figure 1 (a Digital Signal Processor) to detect potential defects in the media item;
    Figure 4 is a pictorial drawing that illustrates part of the media item, with a corner edge folded over itself, along with the threshold lines calculated by the DSP of Figure 3;
    Figure 5 is a pictorial drawing that illustrates part of the media item, with a corner edge folded over itself, next to the optional additional threshold lines calculated by the DSP of Figure 3; and
    Figure 6 illustrates an example of entries in a defect profile configuration file that is used to classify the defects categorized by the system in Figure 1. DETAILED DESCRIPTION
    First, reference is made to Figure 1, which consists of a simplified schematic diagram of a defect categorization system that comprises a media item validator 12 (in the form of a banknote validator) that serves to implement a method of categorizing defects on a media item (and also to assess the fit of the media item) according to one embodiment of the present invention.
    The banknote validator 12 comprises a compartment 13 that supports a transport mechanism 15 in the form of a train of traction rollers comprising upper traction rollers 15a aligned with lower traction rollers 15b, extending from an input port 16 to a capture port 18.
    Entry and capture ports 16,18 are in the form of openings defined by compartment 13. In use, capture port 18 would typically be aligned with parts of a deposit module.
    In use, the drive rollers 15a, b first guide a short edge of the media item (in this embodiment a bank note) 20 through an examination area 22 defined by a gap between the adjacent drive roller parts. Although bank note 20 is being transported through examination area 22, bank note 20 is selectively illuminated by light sources, which include a lower linear array of infrared LEDs 24 arranged to illuminate through the short edge of bank note 20. Infrared LEDs 24 are used for transmitting measurements. Additional light sources are provided for other functions of the bank note validator 12 (for example, bank note identification, counterfeit detection, and the like), however, these are not relevant to this invention, so they will not be described in this document.
    When the infrared LEDs 24 are illuminated, the emitted infrared radiation is incident on a bottom part of bank note 20, and an optical lens 26 focuses the light transmitted through bank note 20 to optical image maker 28 (in this mode, a sensor CCD contact image (CIS)). This provides a transmitted infrared channel emitted from the optical image former 28. In this embodiment, the optical image former 28 comprises an array of elements, each element providing an eight-bit value of detected intensity. The CIS 28 in this modality is a sensor of 200 points per inch, however, the outputs are averaged, in this modality, in such a way that 25 points per inch are provided.
    The light source 24, the lens 26, and the image trainer 28 comprise an image collection component 30.
    The banknote validator 12 includes a data and power interface 32 that serves to allow the banknote validator 12 to transfer data to an external unit, such as an ATM (not shown), a media deposit (not shown), or a PC (not shown), and receive data, commands, and power from it. The banknote validator 12 would typically be incorporated into a media deposit, which, in turn, would typically be incorporated into an ATM.
    The banknote validator 12 also has a controller 34 that includes a digital signal processor (DSP) 36 and an associated memory 38. Controller 34 controls the pull rollers 15 and the image collection component 30 (including energize and de-energize the lighting source 24). Controller 34 also combines and processes the data captured by the image collection component 30, and communicates this data and / or results of any analysis of this data to the external unit via the data and energy interface 32. Controller 34 receives the data from infrared transmission from the optical image former 28.
    Reference is now made to Figure 2, which consists of a flowchart 100 that illustrates the steps taken by the bank note validator 12 (as controlled by DSP 36) to categorize defects in a bank note.
    Initially, a bank note 20 is inserted into the validator 12, which the bank note validator 12 receives (step 102).
    Then, controller 34 transports bank note 20 to examination area 22 (step 104) and causes the image collection component 30 to capture an image of bank note 20 (transmitted by IR) (step 106).
    It should be considered that the image capture process can be used for several different purposes. For example, a captured IR image can be used for bank note recognition, spot detection, and other purposes. Additionally, additional images (such as red and green channel images) can be captured at the same time for use in validating the banknote. In other words, the banknote validator 12 may include other light sources (for example, a green light source), not shown in Figure 1 for clarity. However, these other features and purposes are not essential to understanding this invention, therefore, they will not be described in detail in this document. This is sufficient for a knowledgeable individual to realize that the same banknote validator can be used to perform multiple functions regarding media evaluation.
    Once an image of bank note 20 has been captured, the image is normalized by DSP 36 in such a way that the captured image is compatible with the size of a reference image for that type of bank note (step 108). This is to compensate for any media items that have portions added (such as tape) or have shrunk or expanded, or something like that. Techniques for automatically aligning a captured image with a reference image, and then cropping or adding to the captured image to satisfy the spatial dimensions of the reference image are well known in the art.
    DSP 36 stores the normalized image in memory 38 as a raw image file 39 (which includes the complete pixel intensity information) for further processing (step 110), as described in greater detail below.
    Then, the normalized image is binarized (step 112) such that the binary image comprises a plurality of pixels, each pixel having a potential defect intensity or a non-defect intensity. In this modality, a transmission is used to create the initial image, therefore, any high intensity area indicates a potential defect intensity, and any low intensity area indicates a non-defect intensity. Depending on the usage in question, "potential defect" and "non-defect" refer to defects based only on absence, such as tears, perforations, folds, missing portions, and the like. A feature of absence-based defects is that they cause extraordinarily high light transmission in an area that would normally have much less light transmission. A bank note can have a stained portion that transmits very little infrared; this would be a defect in terms of bank note adaptation, but not a “potential defect” in the context of defects based on absence (tears, perforations, folds, and the like).
    The techniques for creating a binary image are well known. For example, a limit can be applied to the captured pixel intensities, or to a stretched version in contrast to the captured pixel intensities.
    Then, DSP 36 detects all blobs in the binary image (step 114) using a modified union-find structure algorithm. A blob (sometimes referred to as a large binary object) comprises contiguous pixels in an image that have a similar property that is different from the corresponding property of the surrounding pixels. In this mode, a blob comprises pixels of contiguous potential defect (high intensity).
    With reference to Figure 3 (divided into two sheets of drawings as Figures 3A and 3B for reasons of clarity), it consists of a flowchart that illustrates the steps implemented by DSP 36 in the realization of the modified disjoint set structure algorithm. In other words, Figure 3 illustrates the sub-steps of the blob detection step 114.
    Initially, DSP 36 detects all high-intensity pixels (the potential defect pixels) in the first row of the image (step 202). Then, DSP 36 exclusively labels each contiguous group of high-intensity pixels (step 204).
    Then, DSP 36 increases the row (step 206). If this new row exists (step 208) (that is, if the current row is not the last row), then step 202 is repeated for this next row. This continues until all rows have been analyzed, and all groups of high-intensity pixels have been identified.
    Once all rows have been analyzed, then DSP 36 starts in the second row (step 210), and compares each group in that row (initially, the second row) with the groups in the previous row (initially, the first row ) (step 212).
    For each group in the current row (initially, the second row), DSP 36 checks whether that group overlaps any groups in the previous row (step 214).
    If a group in the current row does not overlap any group in the previous row, then that group is assigned a unique blob number (step 216). This is referred to as a growing blob because it can increase in size, depending on whether it overlaps any groups in the next row, as described in greater detail below.
    DSP 36 checks if there are any other groups left in the current row (step 218). If there are more groups, then steps 212 and 214 are repeated for these groups. If there are no more groups in the current row, then processing will continue as described below.
    Going back to step 214, if a group in the current row does overlap some group in the previous row, then DSP 36 checks whether the group overlaps some other group or several other groups (step 220) (see Figure 3B).
    If a group in the current row overlaps just one group in the previous row, then DSP 36 joins the current group with such another group and assigns a unique blob number to the new joined group (if it is in the second row) or uses the number exclusive blob assigned to such another group (if it is in any row after the second row) (step 222). The effect of this is to unite groups that overlap and retain only a single blob identification number for the joined group.
    If a group in the current row overlaps several groups in the previous row, then DSP 36 joins the current group with all groups that overlap the current group in order to create a single joined group (step 224). DSP 36 uses unique blob identification for the group with the highest level in the joined group. This means that the identification is taken from the group that appeared in the first (first, second, third, etc.) rows containing the groups.
    Regardless of whether one or more groups overlap the current group, the next step is for DSP 36 to return to step 218 (Figure 3A) (step 226).
    As mentioned earlier, step 218 is used to ensure that all groups in a row are processed before the DSP increments the row (step 230).
    DSP 36 checks whether this new row exists (step 232). If this row exists, then steps 212 onwards are repeated for this new row.
    If this new row does not exist (in other words, if the current row is the last row in the image), then the blob detection is complete, and the position, identification, and dimensions (length, width, and area) of the blobs detected are all saved in memory 38 by DSP 36 in a blob identification file 40 (Figure 1) (step 234).
    In practical software implementations, the loop of steps 202 to 208 and the loop of steps 212 to 232 can be joined, and the above mentioned tasks of detecting adjacent high-intensity pixel group and the subsequent analysis and joining of group connectivity can be joined. be implemented in a scan of the image pixels.
    Turning to Figure 2, once all the blobs have been detected in step 114, the next phase consists of categorizing the detected blobs.
    This starts at step 116, where each detected blob is compared to a blob area limit (in this mode, it is 8 mm2).
    If a detected blob is less than the blob area limit, then the blob is ignored as insignificant (step 118) and the process moves to step 136 (described in more detail below).
    If a detected blob is greater than (or equal in size to) the blob area limit, then DSP 36 checks whether the detected blob is connected to an edge of bank note 20 (step 120). In other words, is the blob an island in the central portion of the bank note, or does it slope to an edge of the bank note
    If the detected blob is not connected to a bank note edge, then it is categorized by DSP 36 as a hole defect (step 122) and the process moves to step 136 (described in more detail below).
    If the detected blob is connected to a bank note edge, then DSP 36 checks whether the detected blob is in one of the four corners of bank note 20 (step 124).
    If the detected blob is not located in a corner of bank note 20, then it is categorized by DSP 36 as a torn / defective portion defect (step 126) and the process moves to step 136 (described in more detail below) ).
    If the detected blob is located in a corner of bank note 20, then DSP 36 performs calculations to check whether the corner is missing, or if it is folded over itself (referred to as a corner fold) (step 128).
    The calculations performed in this step (step 128) will not be described in more detail with reference to Figure 4, which illustrates part of bank note 20, with a corner edge folded over itself.
    In Figure 4, bank note 20 has a vertical (short) border 60 and a horizontal (long) border 61. Part of the corner (marked “I”) is folded over the top side of bank note 20 to create a region of corner fold 62, as shown by the dark gray portion in Figure 4, leaving an empty area 63. The empty area 63 is shown most clearly in Figure 4 by the portion of the vertical dotted line 64 and the portion of the horizontal dotted line 65. bank note 20 has its image formed and binarized, there will be a blob corresponding in size and shape to the empty area 63. The length of the dotted vertical line 64 is illustrated by the arrow 66; and the length of the dotted horizontal line 65 is illustrated by the arrow 67.
    As can be seen from Figure 4, the corner fold region 62 has an upper angled edge 68 extending from an upper fold point 69 (marked "A") at which the horizontal line portion 67 starts, up to an original corner point 70 (marked “C”). The length of the upper angled edge 68 is equal to the length of the arrow 67.
    The corner fold region 62 also has a lower angled edge 71 extending from the original corner point 70 (marked “C”) to a lower fold point 72 (marked “B”) at which the vertical line portion 64 starts. The length of the lower angled edge 71 is equal to the length of the arrow 66.
    The corner fold region 62 also has a fold edge 74 that extends from the upper fold point 69 (marked "A") to the lower fold point 72 (marked "B").
    The original corner position of bank note 75 (ie, if the corner has not been folded back) is shown at the junction of the two 64.65 line portions and is marked “C0”.
    A bisector 76 is shown extending from the original corner position 75 (marked "C0") to the original corner point 70 (marked "C") and passing through the fold edge 74 at right angles to it .
    An overlap boundary line 77 is shown extending parallel to the fold edge 74 and passing through the original corner point 70 (marked "C"). The non-overlapping area 78 (marked "O"), which lies between the overlap boundary line 77, the corner fold region 62, the vertical border 60, and the horizontal border 61, is shown with lines vertical.
    The detected blob (equivalent to the empty area 63) that corresponds to this corner fold region 62 will directly provide the coordinates for the upper fold point 69 (marked "A") and for the lower fold point 72 (marked as “B”) because the detected blob will correspond to the shape of the corner fold region 62. Since these points are connected by straight lines, these lines can be estimated using the following equation (equation 1): y = mx + c equation 1 In equation 1, m is the gradient of the line and c is the point at which it passes through the geometric axis y (the intersection point).
    Using the coordinates for the upper bend point 69 (marked “A”) and the lower bend point 72 (marked “B”), DSP 36 can estimate the gradient (k) and the intersection point (b) for the bend edge 74.
    DSP 36 accesses the normalized image stored in the raw image file 39 (which was stored in memory 38 in step 110).
    DSP 36 can extrapolate lines 60 and 61 from the standard image to calculate the original corner position 75 (marked “C0”). Once the original corner position 75 is known, it can be symmetrically projected around the fold edge 74 (since the equation for the fold edge is now known) to locate the original corner point 70 (marked as "Ç").
    Once the original corner point 70 (marked "C") is known, DSP 36 uses the coordinates for the upper bend point 69 (marked "A") and the original corner point 70 (marked as “C”) to estimate the gradient (I) and the intersection point (g) for the upper angled edge 68.
    DSP 36 also uses the coordinates for the original corner point 70 (marked "C") and for the lower bend point 72 (marked "B") to estimate the gradient (r) and the intersection point (h ) to the bottom angled edge 71.
    Once DSP 36 has calculated the position of the original corner point 70 (marked "C"), it can then calculate the equation for the overlap limit line 77. This is because the original corner point 70 (marked “C”) is over the overlap line 77, and the gradient of the overlap line (k) is equal to the gradient of the bend edge 74.
    At this stage, DSP 36 calculated the location of the potential corner fold.
    Turning to the flowchart in Figure 2, the next step is for DSP 36 to check if there is a corner fold or a missing portion (step 130). In this sense, DSP 36 accesses the normalized image from the raw image file 39 (which includes the full pixel intensity information that was captured by the bank note validator 12) and uses the calculated points and lines to compare (i) the pixel intensities in the normalized image for the corner fold region 62, with (ii) the pixel intensities in the normalized image for the region defined by the non-overlap area 78.
    If the pixel intensities in the corner fold region 62 are substantially less than the pixel intensities in the non-overlap area 78, then this indicates that there is a corner fold because two layers of substrate are likely to be present in this region. However, if the pixel intensities in the corner fold region 62 are not substantially less than the pixel intensities in the non-overlap area 78, then this indicates that there is a missing corner instead of a corner fold.
    In this modality, DSP 36 compares the lower quartile (that is, the lower 25%) of pixel intensities in the corner fold region 62 of the normalized image, with the lower quartile of pixel intensities in the non-overlapping area 78 of the normalized image.
    In an alternative modality, DSP 36 can calculate two extra threshold lines, as shown in Figure 5. An upper threshold line 82 is parallel to bisector 76 (that is, it has the same gradient as bisector 76) and passes through the bend point higher than 69 (“A”). A lower threshold line 84 is also parallel to bisector 76 and passes through the lower fold point 72 ("B"). The new non-overlap area 88 (marked “Q”) has the same area as the corner fold region 62.
    Similar to Figure 4, DSP 36 compares the lower quartile (that is, the lower 25%) of pixel intensities in the corner fold region 62 of the normalized image, with the lower quartile of pixel intensities in the non-area. -overlay 88 of the normalized image to check for a corner fold (indicated by the pixel intensities in the corner fold region 62 being substantially lower in intensity d than the pixel intensities in the non-overlap area 88) or a missing portion (indicated by the pixel intensities in the corner fold region 62 being similar in intensity to the pixel intensities in the non-overlap area 88).
    If the pixel intensity ratio indicates that a defective portion exists, then DSP 36 categorizes the defect as a defective portion defect (step 132) and the process moves to step 136 (described in more detail below).
    If the pixel intensity ratio indicates that there is a corner fold, then DSP 36 categorizes the defect as a corner fold defect (step 134).
    Then, DSP 36 checks if there is any other detected blob that has not been categorized (step 136). If there are still non-categorized blobs, then DSP 36 returns to step 114 as needed until all detected blobs have been categorized.
    If there are no more non-categorized blobs, then DSP 36 accesses a defect profile configuration file 42 (Figure 1) stored in memory 38 (step 138). The defect profile configuration file 42 includes an entry for each type of defect. Figure 6 illustrates an example of entries in the defect profile 42 configuration file. Each entry includes: a defect type parameter; a defect size field; and a logical parameter.
    As shown in Figure 6, there is an entry 300 for a type of hole defect. Each line includes the defect type parameter (“hole”) as part of the line. The first line 302 includes the defect size field (which comprises a single defect size parameter, that is, an area equal to ten square millimeters); and the second line 304 includes the logic parameter (in this example, no logic connector is required). This entry means that if a type of defect hole is detected that has an area greater than or equal to 10 mm2, then the bank note with such defect must be characterized as inadequate.
    Inlet 310 serves for a corner fold defect. The defect size field comprises two defect size parameters. The first line 312 indicates an area of 130 mm2 (which is the first defect size parameter); the second line 314 indicates a length of 10 mm (which is the second defect size parameter); and the third line 316 indicates that there is a logical AND connection between the two different size parameters. In other words, if a type of defect corner fold is detected that has an area greater than or equal to 130 mm2 and a length greater than or equal to 10 mm, then the bank note having such defect must be characterized as inappropriate (seen that no size parameter is satisfied). If only one of these size parameters is not satisfied, then the bank note must be characterized as adequate (unless other defects are present that would lead to an opposite result).
    Inlet 320 serves for a defect of the missing portion. The first line 322 indicates a length of 6 mm, the second line 324 indicates a width of 5 mm; and the third line 326 indicates that there is an OR logic connection between the two parameters of different size. In other words, if a type of defective portion of defect is detected that has a length greater than or equal to 6 mm or a width greater than or equal to 5 mm, then the bank note having such defect must be characterized as inadequate. If none of these size parameters are met, then the bank note must be characterized as adequate (unless other defects are present that would lead to an opposite result).
    Turning to the flowchart of Figure 2, in step 140, DSP 36 compares each categorized defect with the corresponding entry in the defect profile configuration file 42. This serves to verify whether that defect satisfies the defect profile (step 142) .
    The position, identification, and dimensions (length, width, and area) of the detected blobs were saved in the blob identification file 40, which DSP 36 accesses as part of this step.
    If a categorized defect does not satisfy the defect profile (for example, because a type of defect hole has an area exceeding 10 mm2), then bank note 20 is characterized as unsuitable for continued circulation (step 144).
    If the categorized defect does not satisfy the defect profile, then the next step is for DSP 36 to check if there are any remaining categorized blob that need to be compared with the defect profile configuration file 42 (step 146).
    If there are still some categorized blobs remaining that need to be compared to the defect profile configuration file 42, then DSP 36 returns to step 140 as needed until all categorized blobs have been compared to the configuration file defect profile 42.
    Once all categorized blobs have been evaluated, then bank note 20 is characterized as suitable for continued circulation if all categorized blobs satisfy the defect profile configuration file 42. However, even if a categorized blob does not satisfy defect profile configuration file 42, then bank note 20 is characterized as inappropriate and must be removed from circulation. This can be implemented by segregating bank note 20 in a different storage compartment within the media deposit on which the bank note validator 12 is mounted.
    The categorization and characterization of bank note adaptation is then stopped (step 148).
    This modality has the advantage of being able to be implemented quickly while the bank note is being validated. The defect profile configuration file can be easily updated to accommodate changes in defect size parameters that are considered unacceptable. This mode requires only a single transmission channel, and can use relatively low resolution images (such as 25 dpi).
    Various modifications can be made to the modality described above within the scope of the invention, for example, media items other than bank notes may have been categorized for damage using this technique.
    In other modalities, a different blob detection algorithm can be used than the one previously described.
    In other modalities, a different defect categorization process can be used. For example, a type of tear defect / missing portion can be further categorized as a tear or a missing portion by comparing the size of the blob with a threshold size for a missing portion. If the detected blob is less than the threshold size, then the detected blob is a tear; if it is equal to or greater than the threshold size, then the detected blob is a missing portion.
    In other embodiments, different numerical methods can be used to estimate the location of points in a corner fold. For example, the original corner point 70 (“C”) can be calculated using the lengths of the vertical discontinuous line 64 and the horizontal line 65 (66 and 67, respectively). The original corner point 70 being the intersection of an arc centered at the upper bend point 69 (“A”) having a radius equal to length 67 and an arc centered at the lower bend point 72 (“B”) having an equal radius to length 66.
    In other embodiments, a different data structure can be used for the defect profile configuration file.
    In other modalities, the defect size parameter may be different from those previously described.
    The steps of the methods described here can be performed in any suitable order, or simultaneously when appropriate. The methods described here can be performed by software in machine-readable form on a tangible storage medium or as a propagation signal.
    The terms "that understands", "that includes", "that incorporates", and "that has" are used in this document to quote an open list of one or more elements or steps, not a closed list. When such terms are used, these elements or steps mentioned in the list are not exclusive to other elements or steps that can be added to the list.
    Except where otherwise indicated by context, the terms "one" and "one" are used in this document to denote at least one of the elements, integers, steps, resources, operations, or components mentioned, however, they do not exclude elements, integers, additional steps, resources, operations or components.
    The presence of comprehensive words and phrases, such as “one or more,” “at least,” “but not limited to” or other similar phrases in some cases does not mean, and should not be construed as meaning, that the most limited is intended or required in cases where such comprehensive phrases are not used.
    The reader's attention is directed to all reports and documents that are filed simultaneously, or before this specification according to this request and that are open to public inspection with this specification, and the contents of all these reports and documents are here incorporated for reference.
    权利要求:
    Claims (9)
    [0001]
    1. Method of characterizing defects in a media item, comprising the steps of: receiving a binary image of the media item (step 112), where the binary image comprises a plurality of pixels, with each pixel having a potential defect intensity or an intensity of non-defect; identifying one or more blobs that comprise contiguous pixels each having a potential defect intensity (step 114); for each identified blob, compare a blob size with a damage limit (step 116); ignore the blob if the blob size is less than the damage limit (step 118); and for each identified blob having a size exceeding or equaling the damage limit, categorize the identified blob as a type of defect; for each categorized defect, access a defect profile configuration file to restore (i) a defect type compatible with the categorized defect, and (ii) a defect size associated with that defect type (step 138); and characterize the media item as inappropriate if a categorized defect includes a dimension that meets the defect size requirements for that type of defect (step 144), CHARACTERIZED by the fact that the step of categorizing the identified blob includes one or more of the following categories: tear, missing portion, corner fold and hole.
    [0002]
    2. Method, according to claim 1, CHARACTERIZED by the fact that the step of identifying one or more blobs comprises using a modified disjoint set structure algorithm.
    [0003]
    3. Method, according to claim 2, CHARACTERIZED by the fact that the modified disjoint set structure algorithm comprises: identifying in each row of the binary image, each group of pixels that is contiguous and that has a potential defect intensity; identify for each row boundary, (i) each group in an upper row that overlaps at least partially with a group in the lower row, and join these two groups into a new single group as a growing blob, (ii) each group in a top row that does not overlap with a group in the bottom row, and characterize each group in the top row as a complete blob, and (iii) each group in the bottom row that does not overlap a group in the top row, and characterize each group like a growing blob.
    [0004]
    4. Method according to any one of claims 1 to 3, CHARACTERIZED by the fact that the step of categorizing the identified blob comprises the additional step of: (a) categorizing the blob identified as a hole (step 122) if the blob do not touch one of the edges of the media item.
    [0005]
    5. Method, according to claim 4, CHARACTERIZED by the fact that the step of categorizing the identified blob comprises the additional step of: (b) if the identified blob touches one of the edges of the media item, but is not in a from the corners of the media item, then categorize the blob identified as a tear / missing portion (step 126).
    [0006]
    6. Method, according to claim 5, CHARACTERIZED by the fact that the step of categorizing the identified blob comprises the additional step of: (c) if the identified blob touches one of the corners of the media item, then verify that the blob identified is a corner fold (step 130).
    [0007]
    7. Method, according to claim 6, CHARACTERIZED by the fact that the step of categorizing the identified blob comprises the additional step of: (d) if the identified blob touches one corner of the media item, but it is not a fold corner, then categorize the blob identified as a missing portion (step 132).
    [0008]
    8. Media validator (12) operable to categorize defects in a media item (20) presented to it, CHARACTERIZED by the fact that the media validator (12) comprises: a media item transport (15) to transport a media item (20); an image capture device (30) aligned with the media item transport (15) to capture a two-dimensional array of pixels corresponding to the media item, each pixel having a pixel intensity referring to a property of the media item (20) in a spatial location on the media item (20) corresponding to that pixel; and a processor (34) operable to characterize the image capture device (30) and the media item (20) presented thereon as inadequate by implementing the steps of the method as defined in any of claims 1 to 7.
    [0009]
    9. Media validator, according to claim 8, CHARACTERIZED by the fact that the media validator implements additional media item processing functions on bank notes.
    类似技术:
    公开号 | 公开日 | 专利标题
    BR102012023807B1|2020-09-29|METHOD OF CATEGORIZING DEFECTS IN A MEDIA ITEM AND MEDIA VALIDATOR
    US8805025B2|2014-08-12|Stain detection
    JP5825278B2|2015-12-02|Defect inspection apparatus and defect inspection method
    KR20190004807A|2019-01-14|Banknote management method and system
    JP4113393B2|2008-07-09|Coin discrimination method and apparatus
    RU2562758C2|2015-09-10|Method and apparatus for determining reference data set of class for classification of valuable documents
    KR20050118209A|2005-12-15|Sheet paper identification device and method
    JP2009293999A|2009-12-17|Wood defect detector
    CN106447908B|2020-06-09|Paper money counterfeit distinguishing method and device
    US9047723B2|2015-06-02|Defect categorization
    JP2015180981A|2015-10-15|Device for detecting loading state and method for detecting loading state of bundled paper
    US9852568B2|2017-12-26|Apparatus and method for checking value documents, in particular bank notes, and value-document processing system
    JP6534278B2|2019-06-26|Paper sheet discrimination apparatus and paper sheet discrimination method
    US20130272598A1|2013-10-17|Method and device for examining the optical state of value documents
    WO2013127185A1|2013-09-06|Papery medium authentication device and authentication method thereof
    KR20130125150A|2013-11-18|A media sensing apparatus and financial device
    US10347069B2|2019-07-09|Apparatus, method and assembly for checking value documents, in particular bank notes, and value-document processing system
    JP5743819B2|2015-07-01|Paper sheet processing apparatus and paper sheet processing method
    JP5976477B2|2016-08-23|Character reading device and paper sheet processing device
    AU2016407875B2|2019-10-10|Sheet information collection device and sheet information collection method
    JP2019124519A|2019-07-25|Tablet inspection method and tablet inspection device
    WO2013127186A1|2013-09-06|Paper medium identifying device and identifying method
    JP2010186410A|2010-08-26|Passage detection device
    JP2015001759A|2015-01-05|Paper sheet processing device and paper sheet processing method
    JP2005122587A|2005-05-12|Method and device for discriminating coin
    同族专利:
    公开号 | 公开日
    US20130287264A1|2013-10-31|
    EP2660787A1|2013-11-06|
    CN103377509B|2016-12-21|
    EP2660787B1|2015-10-14|
    CN103377509A|2013-10-30|
    US8983168B2|2015-03-17|
    BR102012023807A2|2014-12-02|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    US5544256A|1993-10-22|1996-08-06|International Business Machines Corporation|Automated defect classification system|
    GB9703191D0|1997-02-15|1997-04-02|Ncr Int Inc|Method and apparatus for screening documents|
    US6233364B1|1998-09-18|2001-05-15|Dainippon Screen Engineering Of America Incorporated|Method and system for detecting and tagging dust and scratches in a digital image|
    JP3693993B2|2002-01-28|2005-09-14|日立オムロンターミナルソリューションズ株式会社|Paper sheet confirmation device|
    DE10335147A1|2003-07-31|2005-03-03|Giesecke & Devrient Gmbh|Method and apparatus for determining the status of banknotes|
    DE602004015615D1|2003-08-01|2008-09-18|Cummins Allison Corp|DEVICE AND METHOD FOR PROCESSING BANKNOTES|
    US7027934B2|2003-09-24|2006-04-11|3M Innovative Properties Company|Apparatus and method for automated web inspection|
    GB0427484D0|2004-12-15|2005-01-19|Money Controls Ltd|Acceptor device for sheet objects|
    JP4709596B2|2005-07-06|2011-06-22|日立オムロンターミナルソリューションズ株式会社|Handling of banknotes that are partially broken|
    US7978899B2|2005-10-05|2011-07-12|Cummins-Allison Corp.|Currency processing system with fitness detection|
    CN101331526B|2005-12-16|2010-10-13|Ncr公司|Banknote validation|
    JP4065893B1|2006-12-04|2008-03-26|東京エレクトロン株式会社|Defect detection device, defect detection method, information processing device, information processing method, and program thereof|
    JP5210013B2|2008-03-19|2013-06-12|株式会社ユニバーサルエンターテインメント|Paper sheet processing equipment|
    JP2011028512A|2009-07-24|2011-02-10|Toshiba Corp|Method for creating dictionary for fitness determination of paper sheet, paper sheet processing apparatus, and paper sheet processing method|
    CN102792154B|2010-03-10|2015-09-23|3M创新有限公司|Repeated defects special in web preparation technology detects|
    US8305650B2|2010-11-16|2012-11-06|Xerox Corporation|Image to target comparator software|US10082454B2|2013-11-04|2018-09-25|Spectra Systems Corporation|Device and method for gasochromic porosity sensing|
    CN103593907B|2013-11-08|2016-09-28|广州广电运通金融电子股份有限公司|A kind of detection foreign body method|
    DE102014002273A1|2014-02-19|2015-08-20|Giesecke & Devrient Gmbh|Method for examining a value document and means for carrying out the method|
    US10073044B2|2014-05-16|2018-09-11|Ncr Corporation|Scanner automatic dirty/clean window detection|
    JP6603718B2|2014-12-27|2019-11-06|ヒルズ・ペット・ニュートリシャン・インコーポレーテッド|Food processing method and system|
    CN104574650A|2015-01-26|2015-04-29|上海古鳌电子科技股份有限公司|Sorting method|
    CN104504727B|2015-01-26|2018-01-12|广州广电运通金融电子股份有限公司|banknote crack detection method and device|
    AU2016265876C1|2015-05-15|2021-02-25|Motion Metrics International Corp|Method and apparatus for locating a wear part in an image of an operating implement|
    WO2016183831A1|2015-05-20|2016-11-24|深圳怡化电脑股份有限公司|Method and device for identifying defects of paper currency|
    US10354471B2|2015-08-19|2019-07-16|Spectra Systems Corporation|Device and method for using gasochromic sensing for evaluating fitness|
    CN105894461A|2015-12-25|2016-08-24|乐视云计算有限公司|Gray morphological image processing method and device|
    US10275971B2|2016-04-22|2019-04-30|Ncr Corporation|Image correction|
    US10055669B2|2016-08-12|2018-08-21|Qualcomm Incorporated|Methods and systems of determining a minimum blob size in video analytics|
    CN109463969A|2018-09-10|2019-03-15|深圳蓝胖子机器人有限公司|Recognition methods, robot and the computer storage medium of express delivery cabinet and its quick despatch|
    US11263388B2|2020-02-17|2022-03-01|Wipro Limited|Method and system for dynamically generating summarised content for visual and contextual text data|
    法律状态:
    2014-12-02| B03A| Publication of a patent application or of a certificate of addition of invention [chapter 3.1 patent gazette]|
    2018-12-11| B06F| Objections, documents and/or translations needed after an examination request according [chapter 6.6 patent gazette]|
    2019-10-15| B06U| Preliminary requirement: requests with searches performed by other patent offices: procedure suspended [chapter 6.21 patent gazette]|
    2020-03-03| B07A| Application suspended after technical examination (opinion) [chapter 7.1 patent gazette]|
    2020-07-28| B09A| Decision: intention to grant [chapter 9.1 patent gazette]|
    2020-09-29| B16A| Patent or certificate of addition of invention granted [chapter 16.1 patent gazette]|Free format text: PRAZO DE VALIDADE: 20 (VINTE) ANOS CONTADOS A PARTIR DE 20/09/2012, OBSERVADAS AS CONDICOES LEGAIS. |
    优先权:
    申请号 | 申请日 | 专利标题
    US13/460,172|US8983168B2|2012-04-30|2012-04-30|System and method of categorising defects in a media item|
    US13/460,172|2012-04-30|
    [返回顶部]