• 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:
    The invention relates to a method for monitoring and managing security at pedestrian crossings and to the device necessary for its practical realization. The device consists of at least one complete monitoring system (1), a system for transmitting information to drivers (2) and an interaction system with traffic light controllers (3) that manage pedestrian crossings, while the monitoring system is in turn (1) composed of a system of stereoscopic cameras, an artificial lighting system, a synchronization device, a processor and a storage system. The procedure uses the images provided by the stereoscopic cameras of the different monitoring systems (1), said images being analyzed by the processor, which is also connected to a storage system. The processor is adapted to extract three-dimensional information about the most relevant objects present in the scene observed by the monitoring system (1), which corresponds to waiting areas and pedestrian crossing areas in pedestrian crossings. The processor is adapted to analyze the relevant three-dimensional objects of the analyzed areas and classify them into pedestrians, vehicles and other elements of the road, such as poles, signaling and bollards, using machine learning systems previously confronted with real and synthetic images. The artificial lighting system is activated when the lighting conditions are below a predetermined value, so that the analysis areas are always illuminated optimally for the operation of the analysis procedure. The activation of the lighting system is accompanied by the acquisition of stereoscopic images through the synchronism system. The processor is adapted to store the result of the analysis and the stereoscopic images in the storage system. (Machine-translation by Google Translate, not legally binding)
    公开号:ES2561912A1
    申请号:ES201431262
    申请日:2014-08-28
    公开日:2016-03-01
    发明作者:David Fernández Llorca；Miguel Ángel Sotelo Vázquez；Ignacio Parra Alonso；Iván GARCÍA DAZA；Raúl QUINTERO MÍNGUEZ；Pablo HYAM BORRACHERO
    申请人:Fundacion Centro De Innovacion De Infraestructuras Inteligentes；Universidad de Alcala de Henares UAH；
    IPC主号:
    专利说明:

    5
    10
    fifteen
    twenty
    25
    30
    35
    40
    Four. Five
    fifty
    or warning vehicle drivers in the vicinity of the intersection, or interacting with the traffic light controller that regulates the intersection. There is a commercial system called SafeWalk [Favoreel 2011, IEEE International Conference on Advanced Video and Signal-based Surveillance] capable of monitoring the presence of pedestrians using a platform of stereoscopic cameras. The main disadvantage of this system is its narrow field of vision, which makes it applicable only in very short range, so that it can only be used to monitor a single waiting area of a pedestrian crossing located on the sidewalk. Monitoring a complete crosswalk would require the combination of SafeWalk with other systems of another nature. Likewise, SafeWalk does not interact with traffic light controllers, so it does not have the ability to change the green time for the crosswalk. This means that SafeWalk cannot increase road safety nor can it optimize the flow of road traffic near the crosswalk.
    There is therefore, to date, no publication or patent in the framework of the monitoring and safety management systems in pedestrian crossings that performs the analysis of the scenes using geometric disparity techniques based on SemiGlobal Block Matching, which extract three-dimensional objects of interest from the images of disparity using geometric analysis techniques that perform the classification of pedestrians, vehicles and other objects from machine learning and geometric and speed characteristics using real and synthetic samples for training, and that interact with information systems for drivers of vehicles and with traffic control systems to optimize the safety conditions at crosswalks.
    EXPLANATION OF THE INVENTION
    The invention relates to a procedure for monitoring and management of safety in crosswalks and to the device for practical implementation. Pedestrian crossings are indicated by road markings on the asphalt that may consist of zebra crossings or crossing lines (Pelikan Crossing) in those cases where the pedestrian crossing is regulated by traffic lights. The main objective is to increase the safety of crosswalks while optimizing the fluidity of road traffic. For this, the invention carried out analyzes the state of the pedestrian crossing, obtaining the number, speed and direction of the existing pedestrians in the waiting and crossing areas of the pedestrian crossing, as well as the number, speed and direction of the approaching vehicles imminently to said crosswalk. Depending on the conditions of the pedestrian crossing relative to the presence of pedestrians in it, visual caution signals are sent to the drivers of vehicles approaching the pedestrian crossing. Likewise, if the crossing is regulated by traffic lights, it interacts with the control system (3) of the same so that the green crosswalk time is modulated within a preset minimum and maximum value, in order to improve the comfort and safety of pedestrians during their transit through the crosswalk, on the one hand, and to optimize the fluidity of road traffic in the vicinity of the crosswalk, on the other.
    The invention consists of a method and a device for its practical implementation. The device consists of a monitoring system (1) for each analysis area, with waiting areas and transit zones for crosswalks being understood as such. Depending on the geometric characteristics of the crossing, a single monitoring system (1) covering the transit zone and the two waiting areas of the crosswalk may be used. The device is completed with an information system (2) for vehicle drivers and an interaction system with traffic light regulator (3) whose objective is respectively to inform the drivers of vehicles approaching the pedestrian crossing about the presence of pedestrians in the waiting and / or transit zones thereof, as well as regulating the green time of the traffic light that controls the pedestrian transit time
    5
    10
    fifteen
    twenty
    25
    30
    35
    40
    Four. Five
    fifty
    at the crosswalk. Each monitoring system (1) consists of two digital cameras (4), an artificial lighting system (5), a synchronization system (6), a processor (7) and a storage system (8). The cameras (4) acquire stereoscopic images (9) of the analyzed area, being understood as such a waiting area, a transit zone or the conjunction of two waiting areas and a transit zone. The stereoscopic images acquired constitute the entrance to the analysis procedure performed by the detection of pedestrians, vehicles and other objects (12) of the road, such as posts, vertical signage and bollards. For each object its position, speed and direction are determined. As part of the procedure, the optimal green time needed to improve the comfort and safety of pedestrians during their transit through the crosswalk is calculated, while avoiding traffic light opening situations in the event that there is no Pedestrian waiting to cross or transit through the crossing area of the pedestrian crossing. All this results in an optimal management of the pedestrian crossing that not only increases the road safety of vulnerable road users, but also optimizes the flow of road traffic in the vicinity of the crosswalk.
    Digital cameras
    These are at least two digital cameras (4) for each monitoring system (1), located in an isolated industrial housing, specially designed for outdoor use on a pole, or on a pole or wall bracket. The cameras must be positioned so that stereoscopic images (9) are captured synchronized from the waiting areas and / or the crosswalks where the pedestrians waiting or passing through the crosswalk are visible. The digital camera can have IEEE1394, IEEE1394-B, USB or Ethernet format output. The camera provides black and white or color images with a minimum resolution of 640x480 pixels, so that pedestrians can be seen with sufficient resolution.
    Image processor
    This system is based on an industrial computer, or an electronic card with DSP or FPGA type processors. In the image processor (7) all the necessary functions are performed to control the process of image capture (9) and memory storage (8), pre-processing, extraction of characteristics of the areas of interest, class classification between pedestrians and vehicles (12) and finally the storage of information in a storage unit (8). The whole process is regulated by the software implemented in the image processor (7). In general, the following functions are performed:
    - Sending orders for capture and storage of stereoscopic images (9).
    - Execution of algorithms for preprocessing images.
    - Execution of any method for obtaining geometric disparity images from three-dimensional images (10).
    - Obtaining the three-dimensional regions of interest (11) where pedestrians are located within the monitoring scenes from the disparity images.
    - Analysis of the three-dimensional regions of interest using techniques of a geometric nature (11) and machine learning for classification (12) between pedestrians, vehicles and other road objects, such as posts, vertical signage and bollards.
    - Monitoring of the objects detected using Bayesian filtering techniques (13).
    5
    10
    fifteen
    twenty
    25
    30
    35
    40
    Four. Five
    fifty
    - Calculation of the optimal time in green for the crosswalk from the conditions measured in it based on a pre-established minimum and maximum time.
    - Sending information to vehicle drivers (15) about the presence of pedestrians waiting to cross or passing through the crosswalk.
    - Interaction with the traffic light controller management system (16) that regulates the crosswalk.
    - Storage of information related to pedestrians and vehicles detected
    (8).
    Artificial lighting system
    It is a system that emits visible light or infrared light, there must be an artificial lighting system (5) for each monitoring system (1) that analyzes either a waiting area or a transit zone or the conjunction of two zones Waiting and a transit zone. The artificial lighting system (5) will be activated only if the ambient lighting conditions of the pedestrian crossing do not reach acceptable minimum values for the correct operation of the image analysis procedure provided by the cameras of the crossing monitoring system. The power of the lighting system can be variable depending on the dimensions of the area to be analyzed by the monitoring system. The activation of the artificial lighting system is synchronized with the stereoscopic imaging of the monitoring system from a synchronization system based on a microprocessor or microcontroller external to the processor.
    Vehicle driver information system
    It is a system that emits a visual warning or caution message directed towards the drivers of the vehicles (15) that approach the pedestrian crossing that is being monitored by the system. A driver information system will be placed for each direction of progress towards the monitored crossing. Thus, two-way driver information systems will be located on two-way roads, one for each direction of traffic, while on the one-way roads a single driver information system will be located, except in those cases in which the The route has more than two lanes, in which case two driver information systems will be located to optimize their visualization, one on each side of the road. The vehicle driver information system is installed on a pole or staff and is fully visible from the position of a driver of a vehicle at a minimum distance of 50 meters.
    Interaction system with traffic light controller
    It is a system that interacts between the processor and the traffic light control system that regulates transit times at the crosswalk. The interaction system with traffic light controller (16) adapts the signals provided by the processor to the levels, duration and protocol of the signals handled by the traffic light controller. This system is based on a microcontroller or microprocessor system.
    BRIEF DESCRIPTION OF THE DRAWINGS
    Figure 1 shows a perspective view of a pedestrian crossing where a pedestrian is crossing. In said pedestrian crossing there is a device composed of a pedestrian crossing monitoring system (1) that is connected to an information system (2) to vehicle drivers approaching the pedestrian crossing and to a system of interaction (3) with the Traffic light controller that regulates the crosswalk.
    5
    10
    fifteen
    twenty
    25
    30
    35
    40
    Four. Five
    fifty
    Figure 2 shows the composition of the pedestrian crossing monitoring device
    (1) which includes two digital cameras (4) responsible for acquiring a stereoscopic image of the analysis area, an artificial lighting system (5) that emits visible or infrared light to ensure that the lighting level of the scene a synchronization system (6) that guarantees the synchronized acquisition of images from digital cameras (4), an image processor (7) in which the general procedure for monitoring and safety management in pedestrian crossings is processed is suitable and a storage system (8) connected to the digital processor.
    Figure 3 shows the general procedure for monitoring and managing pedestrian crossing safety. The procedure begins with the acquisition of two images (9) from the digital cameras (4). The geometric disparity image is then created using the Semiglobal Block Matching technique or any other geometric disparity calculation technique (10). In the next step we proceed to obtain the three-dimensional objects relevant in the scene, using geometric and velocity techniques of the objects (11). The categorization or classification of the various objects in the scene is then carried out using machine learning techniques based on a training process that uses real and synthetic images (12). The objects are classified into pedestrians, vehicles and other road objects, such as posts, vertical signage and bollards. Based on the classified objects, the detection and monitoring of the pedestrians (13) found in the waiting areas and / or in the transit area of the pedestrian crossing is carried out. Likewise, the detection and monitoring of vehicles located in the vicinity of the crosswalk is carried out. In the last steps of the procedure two processes are carried out in parallel: on the one hand, it interacts with an information system
    (2) that sends visual information to the drivers of vehicles approaching the crosswalk (15). This information warns them of the presence of pedestrians waiting or passing at the crosswalk. At the same time, the traffic light control system (3) that regulates the crosswalk is interacted with the aim of optimizing the green crosswalk time (16).
    MODE OF REALIZATION
    The device for the monitoring and management of safety in pedestrian crossings consists of a monitoring system (1) for each area of analysis of the pedestrian crossing, understanding as such the crossing areas and waiting areas, being able to manage two waiting areas and a crossing area together in a single monitoring system (1) depending on the geometric characteristics of the crosswalk. Each monitoring system (1) consists of two digital cameras (4) installed in an outdoor industrial housing on a pole, pole or wall bracket. The cameras (4) can be in color or gray levels, with a minimum resolution of 320x240 pixels, and a recommended resolution of at least 640x480 pixels. The cameras (4) are each composed of an optic with a focal length that will be a function of the distance between the camera and the area to be analyzed. Thus, for example, for cameras installed at 3-4 meters high, the focal length will be between the margins of 2.5-4mm. The cameras (4) have a solid state sensor that can be both CCD and CMOS. The positioning of the cameras (4) can be defined to obtain images of a single waiting or crossing area, as well as two waiting areas and a crossing area together. The cameras (4) are interconnected with the image processor (7) via Firewire or USB or Camera Link or GigE Vision. To obtain images with sufficient contrast in low lighting conditions, there is an artificial lighting system (5) that can be both infrared and visible. Said lighting system (5) will be activated automatically by means of a photocell that monitors the current lighting conditions at all times. The processor (7) is connected to a storage system (8) of hard disk type data or solid state memory preferably. The processor (7) is responsible for processing the images captured by the cameras (4), and it is an industrial PC or an electronic card
    5
    10
    fifteen
    twenty
    25
    30
    35
    40
    Four. Five
    fifty
    55
    based on DSPs or FPGAs. The image processor (7) implements a specific procedure for the recognition of pedestrians, vehicles and other road objects. The device also consists of an information system for drivers of vehicles (2), based on an information signal installed on a pole or pole and which is fully visible from the position of a driver of a vehicle at a minimum distance of 50 meters . A driver information system (2) is installed for each direction of vehicle traffic. In the case of roads with a single traffic direction with more than two lanes, an information system (2) is installed on each side of the road to optimize their perception by vehicle drivers. The device also has an interaction system with traffic light controllers (3) that regulate crosswalks. This system is implemented on a microcontroller or microprocessor-based card, and its mission is the modulation of the green crosswalk time at pedestrian crossings regulated by traffic lights.
    The procedure for monitoring and managing safety at crosswalks is characterized by obtaining one or more stereoscopic images (9) of the waiting areas at marked crosswalks, said waiting areas being generally located in a sidewalk area or in The median of an avenue or promenade. In the same way, one or several stereoscopic images of the pedestrian crossing zone at signposted crosswalks are obtained. Next, a three-dimensional image of geometric disparity is created using the technique called SemiGlobal Block Matching (SGM) or another similar technique of calculating geometric disparity image from stereoscopic images (10). This allows obtaining the three-dimensional elements of interest present in the analyzed scene from the three-dimensional representation of the waiting areas and pedestrian crossing (11) using background subtraction techniques, height filtering and temporal integration with the three-dimensional elements detected in moments before. The discrimination between three-dimensional elements of the scene corresponding to pedestrians, vehicles and fixed objects of the infrastructure is then carried out using techniques of geometric analysis, speed calculation and machine learning to recognize the shape of pedestrians and vehicles, distinguishing them from such objects as street lamps, traffic light posts or traffic signs, or bollards, using a trained classifier with both real and synthetic samples (12). The next step is the detection of the presence of pedestrians in the waiting areas and in the crossing areas in marked crosswalks, as well as the monitoring of their trajectories using Bayesian filtering techniques (13). Likewise, the presence of vehicles is detected in the vicinity of the pedestrian crossing areas and monitoring their trajectories using Bayesian filtering techniques. The result obtained allows to interact with an information system (2) that alerts the drivers of the vehicles (15) about the presence of one or more pedestrians waiting in the waiting area or traveling along the passage zone of Pedestrians of the crosswalk. Alternatively or in addition, the optimal duration of the green traffic light time is regulated, which regulates the time of pedestrian crossing based on the number of pedestrians present at the crosswalk, their traffic speed and safety conditions. From this calculation, we proceed to interact with an information management system (3) that allows modifying the green time duration of the traffic light (16) that regulates the pedestrian crossing time based on the optimal time needed to increase Safety at the crosswalk.
    INDUSTRIAL APPLICATION
    The patent object of this invention has its field of application in the intelligent transport systems industry, as well as in the field of companies in charge of traffic control and regulation in urban areas where the management of pedestrian crossings and traffic lights results essential to increase road safety and the flow of urban traffic.
    5
    10
    fifteen
    twenty
    25
    30
    35
    40
    Four. Five
    fifty
    1. Procedure for monitoring and safety management in crosswalks characterized by comprising the following stages:
    to. Obtaining one or more stereoscopic images (9) of the waiting areas at marked crosswalks, said waiting areas being generally located in a sidewalk area or in the median of an avenue or promenade.
    b. Obtaining one or several stereoscopic images (9) of the pedestrian crossing zone at marked crosswalks.
    C. Creation of a three-dimensional image of geometric disparity using the technique called SemiGlobal Block Matching (SGM) or another similar technique of calculating the image of geometric disparity from stereoscopic images (10).
    d. Obtaining the three-dimensional elements of interest present in the scene analyzed from the three-dimensional representation of the waiting areas and pedestrian crossing using background subtraction techniques, height filtering and temporal integration with the three-dimensional elements detected in previous moments of time (eleven).
    and. Discrimination between three-dimensional elements of the scene corresponding to pedestrians, vehicles and fixed objects of the infrastructure using techniques of geometric analysis, speed calculation and machine learning to recognize the shape of pedestrians and vehicles, distinguishing them from objects such as lampposts, traffic light posts or traffic signs, or bollards, using a trained classifier with both real and synthetic samples (12).
    F. Detection of the presence of pedestrians in the waiting areas and in the crossing areas in marked crosswalks, as well as monitoring their trajectories using Bayesian filtering techniques
    (13).
    g. Detection of the presence of vehicles in the vicinity of pedestrian crossing areas and monitoring of their trajectories using Bayesian filtering techniques (13).
    h. Interaction with an information system that allows vehicle drivers to be alerted to the presence of one or more pedestrians waiting in the waiting area or passing along the pedestrian crossing area of the crosswalk (15).
    i. Calculation of the optimal time duration in green of the traffic light that regulates the time of pedestrian crossing based on the number of pedestrians present at the pedestrian crossing, their traffic speed and safety conditions (16).
    j. Interaction with an information management system that allows modifying the duration of the green traffic light that regulates the pedestrian crossing time based on the optimum time required to increase safety at the crosswalk (16).
    2. Device for the monitoring and management of safety in pedestrian crossings which, using the method of claim [1], is characterized in that it comprises one or several monitoring systems (1) of waiting areas or crosswalks crossing areas pedestrians, an information transmission system (2) to vehicle drivers and an interaction system with traffic light controllers (3).
    权利要求:
    Claims (7)
    [1]
    5
    10
    fifteen
    twenty
    25
    30
    35
    [3]
    3. Device for monitoring waiting areas or pedestrian crossing areas in pedestrian crossings according to claims [1] and [2] characterized in that it comprises two cameras (4), an artificial lighting system (5), a system of synchronism (6), a processor (7) and a storage system (8).
    [4]
    4. Device for monitoring waiting areas or pedestrian crossing areas at crosswalks according to claims [1], [2] and [3], characterized by an artificial lighting system (5) of visible or infrared nature which is activated from the signal sent by the synchronization system (7), in coherence and synchronization with the acquisition of images made by the two cameras (4).
    [5]
    5. Device for monitoring waiting areas or pedestrian crossing areas at crosswalks according to claims [1], [2] and [3], characterized by a processor (7) that adapts to interact with the system of information (2) for the warning to drivers about the presence of pedestrians in the waiting or crossing areas of a pedestrian crossing, as well as with the information management system (3) that allows modifying the green time for the passage of pedestrians at crosswalks controlled by traffic lights.
    [6]
    6. Device for monitoring waiting areas or pedestrian crossing areas at crosswalks according to claims [1], [2] and [3], characterized by a processor (7) that is adapted so that the result of the detection of vehicles and pedestrians, as well as the stereoscopic images corresponding to the waiting areas and pedestrian crossing of crosswalks are stored in a storage device (8) for later use.
    [7]
    7. Device for monitoring waiting areas or pedestrian crossing areas in pedestrian crossings according to claims [1] and [2] characterized in that it comprises an information system (2) that allows a visual warning message to be sent to the drivers of vehicles approaching a pedestrian crossing, informing those drivers about the presence of pedestrians in the waiting area or in the passage area of said pedestrian crossing.
    [8]
    8. Device for monitoring waiting areas or pedestrian crossing areas at pedestrian crossings according to claims [1] and [2] characterized in that it comprises an information management system (3) that allows modifying the corresponding green time at the crosswalk at a crosswalk controlled by traffic lights.
    image 1
    类似技术:
    公开号 | 公开日 | 专利标题
    CN106251701B|2018-11-06|Vehicle rearview monitor and alarm system and method based on rotation zoom multi-cam
    CN101799987B|2011-11-09|Self-adaptive intelligent traffic light and control method thereof
    TWI658441B|2019-05-01|Warning sign placing apparatus and control method
    WO2015089867A1|2015-06-25|Traffic violation detection method
    US20100100268A1|2010-04-22|Enhanced clear path detection in the presence of traffic infrastructure indicator
    CN104157143B|2016-06-01|Parking offense detection system and detection method thereof
    CN105539269A|2016-05-04|Vehicle headlight system with adaptive light distribution
    KR200481229Y1|2016-09-05|Safety management system for flashing signal of road
    KR102006626B1|2019-08-06|Apparatus and method for generating a signal for extending the traffic signal for protecting a pedestrian in crosswalk
    KR102181222B1|2020-11-20|Safety management system for crosswalk pedestrian
    KR102184728B1|2020-11-30|Walk signal guide apparatus through object recognition based on image detection
    KR101388872B1|2014-04-23|Traffic safety system for pedestrian crossing
    KR20160067435A|2016-06-14|Illegal passing notice system of crosswalk
    KR101210615B1|2012-12-11|Regulation system of u-turn violation vehicle
    Czajewski et al.2013|Innovative solutions for improving safety at pedestrian crossings
    Asami et al.2014|Crosswalk location, direction and pedestrian signal state extraction system for assisting the expedition of person with impaired vision
    ES2561912B1|2016-10-11|Procedure for monitoring and safety management in crosswalks and device for its realization
    CN205491376U|2016-08-17|Road lighting and intelligent early warning system based on environment
    KR20190114347A|2019-10-10|Safety speed control system for children's protection zone
    KR20180082788A|2018-07-19|Safety management system for flashing signal of road
    KR102211552B1|2021-02-03|Method And Apparatus for Preventing Reverse Driving
    BV et al.2018|Computer vision based advanced driver assistance system algorithms with optimization techniques-a review
    de La Rocha et al.2010|Image-processing algorithms for detecting and counting vehicles waiting at a traffic light
    CN109094580B|2019-08-09|Train auxiliary driving method, system and computer readable storage medium based on image recognition
    Aubert et al.2014|Digital imaging for assessing and improving highway visibility
    同族专利:
    公开号 | 公开日
    ES2561912B1|2016-10-11|
    引用文献:
    公开号 | 申请日 | 公开日 | 申请人 | 专利标题
    KR20010103326A|2000-05-09|2001-11-23|남기식|control system of traffic signal lamp for protecting|
    EP2211321A1|2009-01-23|2010-07-28|Hella KGaA Hueck & Co.|Method and device for controlling at least one traffic light assembly of a pedestrian crossing|
    ES2370198A1|2010-02-16|2011-12-13|Universidad De Alcala|System of intelligent regulation of pedestrian crosses by means of monitoring of the waiting zones and adaptive semaforic control |
    JP2012221271A|2011-04-11|2012-11-12|Nippon Signal Co Ltd:The|Object recognition system|CN111899530A|2020-06-22|2020-11-06|厦门快商通科技股份有限公司|Traffic light control method, system, mobile terminal and storage medium|
    法律状态:
    2016-10-11| FG2A| Definitive protection|Ref document number: 2561912 Country of ref document: ES Kind code of ref document: B1 Effective date: 20161011 |
    优先权:
    申请号 | 申请日 | 专利标题
    ES201431262A|ES2561912B1|2014-08-28|2014-08-28|Procedure for monitoring and safety management in crosswalks and device for its realization|ES201431262A| ES2561912B1|2014-08-28|2014-08-28|Procedure for monitoring and safety management in crosswalks and device for its realization|
    [返回顶部]