JP4697761B2 - Queue detection method and queue detection apparatus - Google Patents

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates to a matrix detection method and a matrix detection device, and more particularly, to a vehicle detection method and a matrix detection device for detecting a generation status of a monitoring target queue caused by traffic congestion or the like on a road. About.
[0002]
[Prior art]
Conventionally, various proposals have been made in order to grasp the traffic congestion state of vehicles on the road. One of them is to arrange a vehicle detector on the road, detect the temporal transition of the presence or absence of the vehicle at the observation point, and try to grasp the traffic congestion state of the vehicle on the road (hereinafter referred to as “conventional example 1”). "). As the vehicle detector in the conventional example 1, one that detects the presence or absence of a vehicle at an observation point by measuring the distance to an object below by an ultrasonic sensor or the like has been proposed.
[0003]
In addition, a technique for recognizing a traffic jam situation of a vehicle on the road by recognizing a pattern of each vehicle from a result of imaging on the road by a video camera and detecting a movement and a stop situation of each vehicle is also proposed. (Hereinafter referred to as “conventional example 2”).
[0004]
In addition, a technique has been proposed that tries to grasp the traffic jam situation of the vehicle on the road by detecting the movement and stop status of the vehicle from the difference between the imaging result on the road by the video camera and the background image obtained in advance. (Hereinafter referred to as “conventional example 3”).
[0005]
[Problems to be solved by the invention]
In the technique of Conventional Example 1 described above, the number of observation points is limited, and only the time occupation ratio of the vehicle at the observation points can be measured. Therefore, for example, even when there is only one observation point and the vehicle happens to stop at that observation point position, it is erroneously detected that there is a traffic jam, that is, a vehicle queue. In order to prevent such false detection, it is necessary to arrange a large number of vehicle detectors over a wide area and simultaneously process signals from a large number of vehicle detectors. It is necessary to construct a system that has such a system.
[0006]
In the technique of Conventional Example 2, since individual vehicles are extracted from the imaging result by pattern recognition, a very large computational resource is required. Such a system having a large-scale computing resource inevitably has a large-scale and complicated configuration.
[0007]
Further, the technology of Conventional Example 3 may be able to detect the occurrence and cancellation of a vehicle queue with a simple configuration in an environment where the background is statically determined. However, in the outdoors where the background state changes greatly, an effective background image cannot be determined, which is not practical.
[0008]
The present invention has been made under such circumstances, and an object of the present invention is to detect a queue that can accurately detect the occurrence or cancellation of a queue to be monitored such as a vehicle with a simple configuration. It is to provide a method and a queue detection device.
[0009]
[Means for Solving the Problems]
  The queue detection method of the present invention is a queue detection method for detecting an occurrence status of a queue to be monitored,The queueAlong a specific directionThe monitored objectA length that is sufficiently longer than the first specific length, which is a typical length, along the specific directionMonitoring areaIn each of the extracted monitoring areas, after extracting the monitoring area in the imaging result obtained by the imaging each time the imaging is performed, periodically imaging the area including theMore than twice the second specific length, which is a length obtained by adding a typical distance between the monitoring targets adjacent in the queue to the first specific length when the queue is generatedA determination region extraction step of extracting a determination region having a length corresponding to the first image direction corresponding to the specific direction; and the determination region along the first image direction,SaidDivide by image specific length corresponding to the second specific lengthdidAn object region specifying step for specifying an object region in which the monitoring target is estimated after estimating the background density in each of the divided regions; an imaging result obtained by the imaging and a previous imaging A moving object region specifying step of comparing the imaging result and specifying a moving object region; detecting the occurrence of the monitoring target queue based on the object region and the moving object region in each of the divided regions; A queue detection step;The object region specifying step is executed, and the density value of the pixels arranged along the second image direction corresponding to a direction orthogonal to the specific direction while changing the position of the first image direction in the determination region. When the variance and change width of the image are obtained, a background density having the background density as a representative value of the density values of pixels arranged in the second image direction in which the variance is minimum and the change width is equal to or less than an allowable value. Calculating a point in the first image direction in which a difference between a representative value of density values of pixels arranged in the second image direction and the background density is not less than a first predetermined value in each of the divided regions; An object area calculating step for specifying, as the object area, an area in which the image object points are arranged in the first image direction for a predetermined length or more, and the moving object area specifying step is obtained by the imaging. Imaging results A difference image calculation step of calculating a density value difference for each pixel with respect to an imaging result obtained by the previous imaging; and a predetermined number or more of pixels having an absolute value of the density value difference equal to or greater than a second predetermined value. A moving object area calculating step of obtaining, as the moving object area, an area in which a group of pixels arranged along the second image direction is present at a predetermined density or more along the first image direction. The process includes a queue determination step of determining a queue area based on a ratio occupied by each of the object area and the moving object area for each of the divided areas.This is a queue detection method.
[0010]
  According to this, after extracting the monitoring area (for example, the area of a specific lane on the roadway) in the imaging result in the determination area extraction step based on the imaging result of each time, in the extracted monitoring area,More than twice the second specific length, which is a length obtained by adding a typical distance between the monitoring targets adjacent in the queue to the first specific length (for example, the length of a passenger car) when the queue is generatedIs extracted along the first image direction corresponding to the specific direction (for example, the traveling direction of the vehicle in the specific lane). Subsequently, in the object region specifying step, the determination region is along the first image direction,The second specific length(For example, the length of a passenger car plus the typical inter-vehicle distance in a traffic jam) divided by the specified image lengthdidEstimate the background density inside each divided area. Then, using the estimated background density as a reference density, an object region that is estimated to have a monitoring target is specified based on the imaging result. When the imaging result is a perspective image, the second specific length varies depending on the position in the depth direction in imaging of the imaging target.
  Here, in the object region specifying step, first, in the background density calculating step, the position in the first image direction is changed in the determination region, and the second image direction corresponding to the direction orthogonal to the specific direction is changed. When the dispersion and change width of the density values of the pixels arranged along the line are obtained, a representative value of the density values of the pixels arranged in the second image direction in which the dispersion is minimum and the change width is equal to or less than an allowable value is obtained. The background density is used. Then, in the object region calculation step, in each of the divided regions, the first image direction in which a difference between a representative value of density values of pixels arranged along the second image direction and the background density is equal to or greater than a first predetermined value As an image object point, an area in which the image object points are arranged for a predetermined length or more in the first image direction is specified as the object area.
[0011]
  Next, the imaging result obtained by the current imaging is compared with the imaging result obtained by the previous imaging, and the mode of change from the previous to the current imaging result is analyzed to identify the moving object region. In addition, although it is preferable that the time interval of each imaging is set shorter than the shortest time during which the monitoring target may move by the second specific length, it is not limited to this.
  When specifying such a moving object region, first, in the difference image calculation step, a density value difference for each pixel is calculated for the imaging result obtained by the imaging and the imaging result obtained by the previous imaging. In the moving object region calculating step, a pixel group arranged along the second image direction in which a predetermined number or more of pixels having an absolute value of the density value difference equal to or greater than a second predetermined value exists is equal to or higher than a predetermined density. A region present along the image direction is obtained as the moving object region.
[0012]
  And based on the object area and the moving object area in each divided area,Each divided areaThe ratio of the object area and the moving object area in each is calculated, and the generation status of the monitoring target queue is detected by analyzing how the ratios are in each divided area.
[0013]
Therefore, since it is not necessary to recognize the pattern of each monitored vehicle, a huge amount of computing resources are not required, and since a fixed background pattern is not used, it is possible to accurately monitor the object to be monitored even outdoors where the background state changes greatly. The occurrence status of the queue can be detected.
[0014]
In the queue detection method of the present invention, the imaging can be performed with a visual field viewed from above the monitoring area and obliquely intersecting the specific direction.
[0015]
In the queue detection method of the present invention, the second specific length is a length obtained by adding a typical distance between the monitoring targets adjacent in the queue to the first specific length when the queue is generated. be able to.
[0019]
In addition, the queue detection method of the present invention may further include a display step of displaying the occurrence status of the queue detected in the queue detection step.
[0020]
Here, in the display step, at least one of graphic display and voice display can be performed with respect to the occurrence status of the queue.
[0021]
  In the queue detection method of the present invention, the monitoring target can be a vehicle on the road.
  Further, in the queue detection method of the present invention, in the current imaging result, a collection of regions composed only of pixels having density values for which the difference from the background density obtained in the background density calculation step is equal to or less than a predetermined allowable value. Current background image creation step for creating a current background image; cumulative background image update for creating a new cumulative background image by superimposing the current background image on the cumulative background image created from the previous imaging results And a step of replacing the object region specifying step each time the image is taken after the number of pixels of the new accumulated background image has reached a predetermined ratio or more with respect to the number of pixels of the determination region. The second object region specifying step for specifying the object region is executed based on the comparison result between the cumulative background image created from the previous imaging results and the current imaging result image. Can.
[0022]
  The queue detection device of the present invention is a queue detection device that detects the occurrence status of a queue to be monitored,The queueAlong a specific directionThe monitored objectA length that is sufficiently longer than the first specific length, which is a typical length, along the specific directionMonitoring areaImaging means for periodically imaging an area including: after extracting the monitoring area in the imaging result obtained by imaging by the imaging means, in the extracted monitoring area,More than twice the second specific length, which is a length obtained by adding a typical distance between the monitoring targets adjacent in the queue to the first specific length when the queue is generatedA determination area extracting means for extracting a determination area having a length corresponding to the first image direction corresponding to the specific direction; and the determination area along the first image direction;SaidDivide by image specific length corresponding to the second specific lengthdidAn object area specifying means for specifying an object area where the monitoring target is estimated to exist after estimating a background density in each of the divided areas; an imaging result obtained by the imaging and a previous imaging A moving object area specifying means for comparing the imaging result and specifying a moving object area; detecting the occurrence status of the monitoring target queue based on the object area and the moving object area in each of the divided areas; Queue detection means;And the object region specifying unit is configured to change density values of pixels arranged along a second image direction corresponding to a direction orthogonal to the specific direction while changing a position in the first image direction within the determination region. When determining the variance and the change width, after setting the representative value of the density values of the pixels arranged in the second image direction in which the variance is minimum and the change width is equal to or less than the allowable value as the background density, In each of the divided areas, a point in the first image direction in which a difference between a representative value of density values of pixels arranged along the second image direction and the background density is a first predetermined value or more is used as an image object point. An area in which the image object points are arranged in a first image direction for a predetermined length or more is specified as the object area, and the moving object area specifying unit is configured to acquire an imaging result obtained by the imaging and an imaging obtained by the previous imaging. As a result After calculating the density value difference for each pixel, a pixel group arranged along the second image direction in which a predetermined number or more of pixels whose absolute value of the density value difference is the second predetermined value or more exists is a predetermined density or more. The area existing along the first image direction is obtained as the moving object area, and the queue detecting means waits based on the ratio of the object area and the moving object area in each divided area. Determining the matrix regionA queue detection device.
[0023]
  According to this, after every determination by the imaging unit, the determination region extraction unit extracts the monitoring region in the imaging result, and then in the extracted monitoring region,More than twice the second specific length, which is a length obtained by adding a typical distance between the monitoring targets adjacent in the queue to the first specific length when the queue is generatedIs extracted along the first image direction corresponding to the specific direction. Subsequently, the object area specifying unit divides the determination area along the first image direction by an image specific length corresponding to a second specific length having a predetermined relationship with the first specific length.didAn object region in which a monitoring target is estimated to exist within each divided region is specified.
  Here, the object region specifying means changes the position of the first image direction in the determination region, and the density value of the pixels lined up along the second image direction corresponding to the direction orthogonal to the specific direction. After determining the representative value of the density values of the pixels arranged in the second image direction in which the variance is minimum and the change width is equal to or less than an allowable value, as the background density, In each of the divided areas, a point in the first image direction in which a difference between a representative value of density values of pixels arranged in the second image direction and the background density is equal to or more than a first predetermined value is defined as an image object point. A region where the image object points are arranged in a predetermined length or more in the first image direction is specified as the object region.
[0024]
  Next, the moving object area specifying unit compares the current imaging result with the imaging result obtained by the previous imaging, and specifies the moving object area.In specifying the moving object region, the moving object region specifying means calculates a density value difference for each pixel between the imaging result obtained by the imaging and the imaging result obtained by the previous imaging, A region in which a group of pixels arranged along the second image direction in which the absolute value of the density value difference is greater than or equal to a second predetermined value exists in a predetermined number or more exists along the first image direction with a predetermined density or more. And obtained as the moving object region.
  Then, the queue detection means detects the occurrence status of the queue to be monitored based on the object area and the moving object area in each divided area.When detecting the occurrence status of the queue to be monitored, the queue detection means determines the queue area based on the ratio of the object area and the moving object area for each of the divided areas.
[0025]
That is, the queue detection device of the present invention detects the occurrence status of the queue to be monitored using the queue detection method of the present invention. Therefore, it is possible to accurately detect the occurrence of the queue even with the simple configuration even when the background environment is greatly changed.
[0026]
The queue detection device according to the present invention may further include a display device for displaying the occurrence status of the queue detected by the queue detection means.
[0027]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, an embodiment of the present invention will be described with reference to FIGS. FIG. 1 is a block diagram illustrating a configuration of a queue detection device 100 according to an embodiment. The queue detection device 100 is a device that detects the occurrence status of a vehicle queue in a specific lane of a roadway, that is, the occurrence status of a traffic jam of a vehicle.
[0028]
As shown in FIG. 1, the queue detection device 100 detects a situation in which a vehicle queue is generated by processing a video camera 10 serving as an imaging unit that captures an image on a road and an imaging result of the video camera 10. The processing device 20 includes a screen display device 43 and a voice display device 45 as display devices that display the occurrence status of the queue. The queue detection device 100 further includes an input device 41 such as a keyboard and a mouse for inputting instructions to the processing device 20 by the operator. The screen display device 43 is also used to display input information from the input device 41 for confirmation and to display the operation status of the queue detection device 100.
[0029]
As conceptually shown in FIG. 2 (A), the video camera 10 is installed above the roadway R, and the vehicle OBJ1 in the lane RL (see FIG. 2 (B)) to be monitored on the roadway R obliquely from above. , OBJ2 and OBJ3 are observed on the roadway R with a visual field in which the traveling direction is expected. As a result, in the field of view of the video camera 10, the state of the vehicles OBJ1, OBJ2, OBJ3 and their surroundings on the roadway R can be captured with a perspective view as shown in FIG. . Further, it is assumed that the field of view of the video camera 10 is set so that the roadway R is viewed at an angle at which the roadway surface between vehicles can be observed at least on the near side even in a traffic jam.
[0030]
In the following description, the traveling direction of the vehicles OBJ1, OBJ2, and OBJ3 in the lane RL of the roadway R is referred to as the Y direction, and the width direction of the roadway R is referred to as the X direction. Here, on the image (ie, the monitor coordinate system (X_M, Y_M), The traveling direction of the vehicles OBJ1, OBJ2, and OBJ3 in the lane RL varies depending on the position in the X direction. However, the traveling direction of the vehicles OBJ1, OBJ2, and OBJ3 in the lane RL is collectively referred to as the Y direction. .
[0031]
Further, the roadway R is one lane on one side, and the separation of the lanes is indicated by white lines WL1 to WL5. Further, it is assumed that there is a step between the roadway R and the sidewalk, and the roadway R and the sidewalk can be identified by the wall surface of the step.
[0032]
Returning to FIG. 1, the processing device 20 includes (a) a control device 29 that performs overall control of the queue detection device 100, (b) an imaging data collection device 21 that collects imaging result data from the video camera 10, ( c) An image analysis device 28 that performs image analysis based on the collected imaging data, and (d) a storage device 30 are provided.
[0033]
The image analysis device 28 includes (i) a determination region extraction device 22 as a determination region extraction unit that extracts a determination region JGA (see FIG. 2B) from a captured image, and (ii) each of the determination regions JGA. And an object area calculating device 23 as object area specifying means for specifying an object area in the divided area. Further, the image analysis device 28 includes (iii) a moving object region calculation device 24 as a moving object region specifying unit that specifies a moving object region in each divided region in the determination region JGA, and (iv) an object region in each divided region. A queue calculation device 25 is provided as queue detection means for detecting the occurrence status of the queue based on the information and the moving object area information. Furthermore, the image analysis device 28 includes a display instruction device 26 that displays the detected queue generation status via the screen display device 43 or the voice display device 45.
[0034]
The storage device 30 includes an imaging data storage area 31, a determination area data storage area 32, an object area information storage area 33, a moving object area information storage area 34, and a queue information storage area 35.
[0035]
In FIG. 1, the data flow is indicated by a solid arrow, and the control flow is indicated by a dotted arrow. The operation of each device of the processing device 20 configured as described above will be described later.
[0036]
In the present embodiment, the processing device 20 is configured by combining various devices as described above. However, the processing device 20 is configured as a computer system, and functions of each device except the storage device 30 are built in the processing device 20. It can also be realized by a program.
[0037]
Hereinafter, the queue detection operation by the queue detection apparatus 100 of the present embodiment will be described with reference to other drawings as appropriate along the flowchart shown in FIG.
[0038]
Note that, as a premise, the processing related to the imaging by the video camera 10 and the imaging result is performed at least once, and the determination area data storage area 32 contains the image data of the determination area specified from the previous imaging result. It is assumed that
[0039]
First, in step 111 in FIG. 3, the imaging data collection device 21 collects imaging data sent from the video camera 10 for one screen. Then, the imaging data collection device 21 stores the collected imaging data in the imaging data storage area 31.
[0040]
Next, in step 112, the determination area extraction device 22 reads out the imaging data from the imaging data storage area 31, and extracts the determination area JGA (see FIG. 2B). In extracting the determination area JGA, the determination area extraction device 22 first recognizes the above-described white lines WL1 to WL5 and the boundary wall surface between the roadway R and the sidewalk from the read imaging data, and the area of the lane RL that is the monitoring area. Is identified. Then, a determination area JGA near the center of the lane RL is extracted from the identified area of the lane RL. As the determination area, even if the vehicle exists as the Y position (vehicle traveling direction position), the vehicle does not exist in the X direction (width direction of the lane RL) at the Y position. An area is selected such that the road surface appears to some extent (for example, 1/3 or less of the width in the X direction of the determination area).
[0041]
Thus, by extracting the determination area JGA based on the white line or boundary wall surface that does not move originally, the instability of extraction of the determination area JGA in the captured image due to the shake of the video camera or the like is eliminated. The determination area extraction device 22 stores the extracted image data in the determination area JGA in the determination area data storage area 32 as the current determination area data.
[0042]
Next, in step 113, the object area calculation device 23 reads the current determination area data from the determination area data storage area 32, and specifies the object area where the vehicle exists. In specifying the object area, the object area calculation device 23 first adds a length of the current determination area image to the typical length of the vehicle along with the average inter-vehicle distance in a traffic jam along the Y direction. A plurality of divided areas are defined by dividing the image according to the length of the image. Note that, when the image is obtained as a perspective image as in the present embodiment, the length of the divided region in the Y direction varies depending on the Y position.
[0043]
Subsequently, the object area calculation device 23 calculates the median value, density change width, and variance value of the density values of a plurality of pixels arranged in the X direction at each Y position for each divided area. Then, the Y position where the density change width of the imaging result of the lane RL is equal to or smaller than the allowable value and the variance is minimum is specified. Then, the object area calculation device 23 estimates that the Y position is a Y position where only the road surface exists along the X direction in each divided area, and adopts the median value as the road surface density value, that is, the background density value.
[0044]
Next, the object region calculation device 23 compares the obtained background density value with the median value at each Y position calculated above for each divided region. Then, the object area calculation device 23 recognizes the Y position where the difference between the median value and the background density value is equal to or greater than a predetermined first predetermined value as an object point, and the object point is determined as a first predetermined value in the Y direction. An area that continues for more than the number of pixels is specified as an object area. The object area calculation device 23 stores the object area information in each divided area thus obtained in the object area information storage area 33.
[0045]
Next, in step 114, the moving object area calculation device 24 reads the current determination area data and the previous determination area data from the determination area data storage area 32, and the vehicle moves from the previous imaging to the current imaging. The moving object area thus identified is identified. In specifying such a moving object region, the moving object region calculation device 24 calculates a difference image between the determination region image captured last time and the determination region image captured this time.
[0046]
Subsequently, the moving object region calculation device 24 calculates a Y position where the number of pixels equal to or greater than a predetermined second predetermined value among a plurality of pixels arranged in the X direction at each Y position in the difference image is equal to or greater than a certain ratio. Recognize as a moving object point. Then, the moving object area calculation device 24 identifies an area in which the moving object points continue for a second predetermined number of pixels or more in the Y direction as the moving object area. The moving object area calculation device 24 stores the moving object area information thus obtained in the moving object area information storage area 34.
[0047]
Next, in step 115, the queue calculation device 25 reads out the object area information from the object area information storage area 33 and also reads out the moving object area information from the moving object area information storage area 34 to check the occurrence status of the queue. To detect. In detecting the occurrence status of the queue, the queue calculation device 25 determines the Y direction ratio occupied by the object region and the Y direction occupied by the moving object region in each of the divided regions based on the object region information and the moving object information. Calculate the ratio. In addition, the queue calculation device 25 obtains a stationary object area that is an object area in each divided area but not a moving object area based on the object area information and the moving object information, and the stationary object area becomes the divided area. Find the percentage of the total. The queue calculation device 25 comprehensively analyzes the ratio information calculated in this way, and the current queue state of the vehicle, that is, in which Y position and in which section the queue is generated. Ask for. In this embodiment, the queue length in the determination area image is converted into a length in the real space using a coordinate conversion method, and the converted length is congested to a typical length of the vehicle. The length of the queue is converted to the number of vehicles by dividing by the length of the average distance between vehicles. The queue calculation device 25 stores the queue information thus obtained in the queue information storage area 35.
[0048]
Next, in step 116, the display instructing device 26 reads the queue information from the queue information storage area 35 and determines whether or not the queue generation status has changed since the previous queue generation status was displayed. judge. If a positive determination is made, the process proceeds to step 117. On the other hand, if a negative determination is made, the process proceeds to step 118. Hereinafter, description will be made assuming that a positive determination is made in step 116.
[0049]
In step 117, the display instruction device 26 instructs the screen display device 43 to display a display visualizing the occurrence status of the queue based on the queue information, and converts the occurrence status of the queue into the voice display device 45. Display instructions. As a result, the occurrence state of the queue and the change state thereof are provided to the operator through vision and hearing. When the display operation is thus completed, the process proceeds to step 118.
[0050]
In step 118, the control device 29 determines whether or not the input device 41 has instructed the end of the queue detection operation. If a negative determination is made here, the process proceeds to step 111, and the queue detection is continued in the same manner as described above. On the other hand, if an affirmative determination is made in step 118, the queue detection operation ends. Thus, the queue is detected until the input device 41 instructs the end of the queue detection operation.
[0051]
As described above, according to the queue detection device 100 of the present embodiment, the determination region extraction device 32 extracts the lane region RL that is the monitoring region in the imaging result based on the imaging result of the video camera 10 each time. After that, the determination area JGA is extracted in the extracted lane area RL. Subsequently, the object region calculation device 33 divides the determination region JGA into a plurality of divided regions, specifies the object region in each divided region, and the moving object region calculation device 34 uses the imaging result obtained by the current imaging. The moving object region is specified by comparing the imaging result obtained by the previous imaging and analyzing the mode of change of the imaging result from the previous time to the current time. Then, based on the object area and the moving object area in each divided area, the occurrence status of a queue of vehicles to be monitored is detected. Therefore, since it is not necessary to recognize the pattern of each monitored vehicle, a huge amount of computing resources are not required, and since a fixed background pattern is not used, it is possible to accurately monitor the object to be monitored even outdoors where the background state changes greatly. The occurrence status of the queue can be detected.
[0052]
Further, according to the queue detection device 100 of the present embodiment, the monitoring area is imaged from the diagonally upper side by the video camera 10, so that an area sufficient for detecting the occurrence of the queue is observed by one video camera 10. Can be shifted.
[0053]
Moreover, since the object area calculation device 33 obtains the density value of the background image, that is, the density value of the imaging result of the road surface every time the image is taken, it is possible to accurately detect the queue in the outdoors where the environmental conditions greatly change. it can.
[0054]
Further, since the object area and the moving object area in each divided area are specified and comprehensively determined to detect the queue generation status, the queue generation status can be detected with high accuracy.
[0055]
In this embodiment, the road is one lane on one side, but it may be a plurality of lanes.
[0056]
In this embodiment, the median value of the density values of a plurality of pixels arranged in the X direction in each divided region is used for estimating the background density value and specifying the object region. May be used.
[0057]
Further, in this embodiment, the monitoring area is imaged from the diagonally upper side by the video camera 10, but if a sufficiently wide-angle imaging device is used, the monitoring area can also be imaged from above the monitoring area. In such a case, it is possible to eliminate the troublesomeness of coordinate conversion that is derived from the perspective image.
[0058]
In the above embodiment, the background density value is obtained for each imaging and for each divided area, and each Y position is determined from the difference between the representative value (median value) of the pixel density value and the background density value at each Y position in the divided area. The object region is specified by determining whether or not is an object point. However, it is also possible to specify the object region as follows. That is, after obtaining the background density value obtained in the same manner as in the above embodiment for each image pickup, the image pickup result is composed of only pixels whose difference from the background density value is equal to or less than a predetermined allowable value, and the background. A group of areas estimated to be areas is created as a current background image, and a new cumulative background image is created by superimposing the current background image on the cumulative background image obtained up to the previous imaging. In creating a new cumulative background image, the density value of the current imaging result is used as it is for the pixels constituting the current background image, and does not exist in the current background image, and until the previous imaging. For the pixels constituting the accumulated background image obtained in step (b), the images are superimposed so that the density value of the accumulated background image obtained until the previous imaging is adopted.
[0059]
Then, the object region is specified in the same manner as in the above-described embodiment until the number of pixels of the accumulated background image becomes equal to or greater than a predetermined ratio with respect to the number of pixels of the determination region. On the other hand, after the number of pixels of the cumulative background image exceeds a predetermined ratio with respect to the number of pixels in the determination region, the cumulative background image is adopted as the determination background image, and the object region specification for each subsequent imaging is captured. This is performed based on a comparison image between the obtained image and the accumulated background image. Of course, the determination background image is updated every time a new image is taken in the same manner as in the case of the cumulative background image.
[0060]
In the above embodiment, the object region is specified based on the background density value, the density value of the imaging result, and the magnitude of the difference. However, in order to enhance the contour, an image of the density difference at each pixel position is used. Filter processing is performed to discriminate between pixels having a large density difference and pixels that cannot be said to have a large density difference, binarization processing is performed, and so-called expansion processing and reduction processing are performed on the binarized image. It is also possible to detect the object region after erasing the cavity in the original object region and eliminating the noise outside the object region.
[0061]
In the present embodiment, the monitoring target is a vehicle on the road, but the present invention can be applied to other objects or people as long as a queue can be generated. Further, the present invention can be applied regardless of whether it is outdoors.
[0062]
【The invention's effect】
As described above in detail, according to the queue detection method and the queue detection device of the present invention, it is possible to accurately detect the occurrence status of a queue with a simple configuration.
[Brief description of the drawings]
FIG. 1 is a block diagram showing a schematic configuration of a queue detection device according to an embodiment of the present invention.
2A is a diagram for explaining the arrangement of the video camera in FIG. 1, and FIG. 2B is a diagram for explaining a result of imaging by the video camera in FIG. 1; .
FIG. 3 is a flowchart for explaining queue detection processing by the queue detection apparatus of FIG. 1;
[Explanation of symbols]
DESCRIPTION OF SYMBOLS 10 ... Video camera (imaging means), 22 ... Determination area extraction apparatus (determination area extraction means), 23 ... Object area calculation apparatus (object area identification means), 24 ... Moving object area calculation apparatus (moving object area identification means), 25 ... Queue calculation device (queue detection means), 43 ... Screen display device (part of display device), 45 ... Voice display device (part of display device).

Claims (8)

A queue detection method for detecting an occurrence status of a queue to be monitored,
Periodic imaging of a region including a monitoring region having a length along the specific direction that is sufficiently longer than a first specific length, which is a typical length of the monitoring target along the specific direction that the queue can form Each time the imaging is performed,
After extracting the monitoring area in the imaging result obtained by the imaging, in the extracted monitoring area, a typical distance between the monitoring targets adjacent in the queue is generated in the queue when the queue is generated. A determination region extraction step of extracting a determination region having a length corresponding to at least twice the second specific length, which is a length added to the length , along the first image direction corresponding to the specific direction;
Along the determination region in the first image orientation, inside divided areas each divided in the corresponding image specified length to the second specific length is estimated that after estimating the background concentration, the monitoring target is present An object region specifying step for specifying the object region;
A moving object area specifying step of comparing the imaging result obtained by the imaging and the imaging result obtained by the previous imaging and specifying the moving object area;
Running; the division on the basis of the object area and the moving object region in each region, said a queuing detection step of detecting the occurrence of a monitored queue
The object region specifying step includes:
In the determination area, while changing the position of the first image direction, when obtaining the dispersion and change width of the density value of the pixels arranged along the second image direction corresponding to the direction orthogonal to the specific direction, A background density calculation step in which the background density is a representative value of density values of pixels arranged in the second image direction in which the variance is minimum and the change width is equal to or less than an allowable value;
In each of the divided areas, a point in the first image direction in which a difference between a representative value of density values of pixels arranged in the second image direction and the background density is equal to or more than a first predetermined value is defined as an image object point. An object area calculating step of specifying, as the object area, an area in which the image object points are arranged in a predetermined length or more in the first image direction;
The moving object region specifying step includes:
A difference image calculation step of calculating a density value difference for each pixel with respect to an imaging result obtained by the imaging and an imaging result obtained by the previous imaging;
A region in which a group of pixels arranged along the second image direction in which the absolute value of the density value difference is greater than or equal to a second predetermined value exists in a predetermined number or more exists along the first image direction with a predetermined density or more. A moving object area calculating step for obtaining the moving object area;
The queue detection step includes a queue determination step of determining a queue region based on a ratio occupied by each of the object region and the moving object region in each divided region.
A queue detection method characterized by the above .   2. The queue detection method according to claim 1, wherein the imaging is performed in a visual field viewed from above the monitoring area and in a direction oblique to the specific direction. The queue detection method according to claim 1 , further comprising a display step of displaying an occurrence status of the queue detected in the queue detection step. 4. The queue detection method according to claim 3 , wherein in the display step, at least one of graphic display and voice display is performed with respect to an occurrence state of the queue. The queue detection method according to claim 1 , wherein the monitoring target is a vehicle on a road. In this imaging result, a current background image creation step of creating a collection of regions consisting only of pixels having density values that are equal to or less than a predetermined allowable value in the difference from the background density obtained in the background density calculation step as the current background image When;
A cumulative background image update step of creating a new cumulative background image by superimposing the current background image on the cumulative background image created from the previous imaging results;
After the number of pixels of the new accumulated background image is equal to or greater than a predetermined ratio with respect to the number of pixels of the determination area, each time the image is taken, instead of the object area specifying step,
A second object region specifying step for specifying the object region is performed based on a comparison result between the cumulative background image created from the previous imaging results and the current imaging result image;
The queue detection method according to any one of claims 1 to 5, wherein A queue detection device for detecting occurrence status of a queue to be monitored,
Periodic imaging of a region including a monitoring region having a length along the specific direction that is sufficiently longer than a first specific length, which is a typical length of the monitoring target along the specific direction that the queue can form Imaging means for performing;
After extracting the monitoring area in the imaging result obtained by imaging by the imaging means, in the extracted monitoring area, a typical distance between the monitoring targets adjacent in the queue is generated when the queue is generated. A determination area extracting means for extracting a determination area having a length corresponding to at least twice the second specific length, which is a length added to the first specific length, along a first image direction corresponding to the specific direction;
Along the determination region in the first image orientation, inside divided areas each divided in the corresponding image specified length to the second specific length is estimated that after estimating the background concentration, the monitoring target is present An object area specifying means for specifying the object area;
A moving object area specifying means for comparing the imaging result obtained by the imaging and the imaging result obtained by the previous imaging and specifying the moving object area;
Comprising a; wherein in each divided region based on the object region and the moving object region, and queuing detection means for detecting the occurrence of said monitored queue
The object region specifying means includes
In the determination area, while changing the position of the first image direction, when obtaining the dispersion and change width of the density value of the pixels arranged along the second image direction corresponding to the direction orthogonal to the specific direction, The representative value of the density values of the pixels arranged in the second image direction in which the variance is minimum and the change width is equal to or less than the allowable value is set as the background density,
In each of the divided areas, a point in the first image direction in which a difference between a representative value of density values of pixels arranged in the second image direction and the background density is equal to or more than a first predetermined value is defined as an image object point. An area in which the image object points are arranged in a predetermined length or more in the first image direction is specified as the object area;
The moving object region specifying means includes:
After calculating the density value difference for each pixel for the imaging result obtained by the imaging and the imaging result obtained by the previous imaging,
A region in which a group of pixels arranged along the second image direction in which the absolute value of the density value difference is greater than or equal to a second predetermined value exists in a predetermined number or more exists along the first image direction with a predetermined density or more. , As the moving object region,
The queue detection means determines a queue area based on a ratio of the object area and the moving object area in each divided area.
A queue detection device characterized by the above . Queuing detection apparatus according to claim 7, further comprising a display device for displaying the occurrence of the detected queue by the queue detection means, it is characterized.

Images