JP4209826B2 - Traffic flow monitoring device - Google Patents

Description

  The present invention relates to a traffic flow monitoring device that detects by image processing the presence of a vehicle that has stopped running within a monitoring range in an automobile traffic flow, the presence of a fallen object, and the like.

As one of the avoidance vehicle detection methods in the traffic flow measurement, there is a method of detecting an individual vehicle and determining avoidance from the amount of movement for a certain time.
However, in this method, it is necessary to recognize the individual vehicle, and there is a problem that the determination accuracy is lowered if the individual vehicle cannot be recognized.
In addition, in falling object detection, if the size of the falling object cannot be secured to some extent in the image, it is judged as image noise or a shadow due to the structure, and a minute falling object in the back of the screen cannot be found. was there.
In Patent Document 1, the motion vector of an image captured by a remote monitoring camera and the direction of the motion vector are calculated for each block, and the average of the direction of the motion vector and the direction of the normal motion vector detected in advance is calculated. If there is a block that is shifted by a predetermined amount or more, it is described that the road is abnormal such as overtopping.

JP 2001-357484 A (pages 3 to 4, FIG. 1)

In the conventional avoidance vehicle / falling object determination method, in order to detect the avoidance vehicle, it is necessary to measure the movement trajectory of the individually recognized vehicle.
Moreover, in the fallen object detection, there is a problem that a fallen object with a minute pixel at the back of the screen is overlooked as noise.
Moreover, in the thing of patent document 1, since the reference | standard of road abnormality detection was made into the average of the direction of the normal motion vector detected in advance, when the average value is not constant, motion vector accumulation | storage at the beginning of operation start When the amount is small, the variance value of the motion vector is likely to fluctuate easily, resulting in a problem that the detection accuracy is lowered.

  The present invention has been made to solve the above-described problems, and an object of the present invention is to provide a traffic flow monitoring device that can efficiently detect a avoiding vehicle or a stopped vehicle without recognizing an individual vehicle. ing.

In the traffic flow monitoring apparatus according to the present invention, an image capturing means for capturing a road state at a predetermined cycle, and an object for tracking the moving direction of the object for each block obtained by dividing the image captured by the image capturing means into blocks. Tracking means, basic object moving direction setting means for presetting a basic object moving direction corresponding to the road shape for each block of the image, basic object moving direction for each block set by the basic object moving direction setting means, and object Object movement direction comparison means that compares the movement direction of the object for each block tracked by the tracking means for each block, and sets a block whose movement direction difference is equal to or greater than the first value as an escape block , this object movement direction comparison When the block adjacent to the escape block set by the means is the escape block, the escape block connecting the adjacent escape blocks Binding processing unit, and that the number of blocks of避走connecting block connected by the避走block linking unit is provided with a避走object determination unit determines the object as避走object when a predetermined time period, more than the predetermined one It is.

As described above, according to the present invention, the image capturing unit that captures the road state at a predetermined cycle, and the object tracking that tracks the moving direction of the object for each block obtained by dividing the image captured by the image capturing unit into blocks. Means, basic object moving direction setting means for presetting a basic object moving direction corresponding to the road shape for each block of the image, basic object moving direction for each block set by the basic object moving direction setting means, and object tracking The object movement direction comparison means for comparing the movement direction of the object for each block tracked by the means for each block, and setting a block having a movement direction difference equal to or greater than the first value as the escape block , and this object movement direction comparison means When the block adjacent to the escape block set by is a escape block, the escape block connection processing unit that connects adjacent escape blocks And the number of blocks is a predetermined period避走connecting block connected by the避走block linking unit, because with a避走object determination unit determines the object as避走object when less than a predetermined one, monitoring the road If the basic object moving direction for each block according to the road shape is set, the escape object can be detected at low cost.

Embodiment 1 FIG.
Embodiment 1 of the present invention will be described with reference to the drawings.
1 is a block diagram showing a configuration of a traffic flow monitoring apparatus according to Embodiment 1 of the present invention.
In FIG. 1, an image capturing unit 10 (image capturing unit) captures an image with a predetermined cycle by an image sensor such as a monitoring ITV camera, and stores the image as digital data. The basic object traveling direction setting unit 11 (basic object traveling direction setting means) is a block region including a road lane that needs to be monitored on the captured image. When the moving object moves along the lane, the basic object traveling direction setting unit 11 The object moving direction is set, and since the setting value does not change when the installation condition of the camera is determined, the information is determined in advance. The method for determining the basic object moving direction S (i, j) of the (i, j) th block B (i, j) is as follows.
As basic information, based on the known information of block information and road alignment information, the center coordinates of block B (i, j) are set to (x, y), and the vertical width of B (i, j) is set to H (B (i, j). )), The horizontal width of B (i, j) is W (B (i, j)), the left end x coordinate of the lane where the coordinates (x, y) exist is LE (y), and similarly the right end x coordinate is RE (y y), the average slope of the curve connecting the lane edge coordinates LE (y1) and LE (y2) is LS (y1, y2), and similarly the average slope of the curve connecting the lane edge coordinates RE (y1) and RE (y2) If it is obtained as RS (y1, y2),
The x-coordinate fraction BR (i, j) in the lane at the block center
BR (i, j) = (x-LE (y)) / (RE (y) -LE (y))
Then, the basic object moving direction S (i, j) of each block B (i, j) is
S (i, j) = BR (i, j) * LS (y-K1 * H (B (i, j)), y + K2 * H (B (i, j)))
+ (1-BR (i, j)) * RS (y-K1 * H (B (i, j)), y + K2 * H (B (i, j)))
(K1 and K2 are constant values)
It becomes.
Note that the image blocks used in the basic object traveling direction setting unit 11 are divided in advance as blocks of a fixed size.

The object tracking unit 12 (object tracking means) performs tracking processing for all blocks including road lanes that require monitoring on the captured image, calculates the object movement direction for each block from the block tracking result, In addition, by combining with the camera installation information, the moving speed of the object in each block is also calculated. Object tracking processing for each block by the object tracking unit 12 at this time is based on matching processing in a two-dimensional grayscale black and white image, and matches how much pixel amount each block has moved in which direction in the next processing cycle. Tracking is performed by the processing, and the moving direction and moving speed with respect to the road are calculated from the moving direction and moving amount, and camera installation information.
The object movement direction comparison unit 13 (object movement direction comparison unit) follows the temporary movement direction on the road lane given by the basic object traveling direction setting unit 11 and the oblique line from the object movement direction obtained by the object tracking unit 12. An object that is not moving, that is, an object that is performing an evasive travel is extracted.

  The avoidance block connection processing unit 14 and the avoidance object determination unit 15 constitute an avoidance object determination unit. The escape block connection processing unit 14 connects similar block information of the avoidance traveling direction and the traveling speed between adjacent blocks. The avoidance object determination unit 15 determines whether or not an escape is in the lane from the connection avoidance block information obtained by the avoidance block connection processing unit 14, and determines a traffic event of avoidance.

FIG. 2 is a plan view showing the provisional travel direction of each block set by the basic object travel direction setting unit of the traffic flow monitoring apparatus according to Embodiment 1 of the present invention.
FIG. 2 shows the basic traveling direction of an object that is uniquely determined according to the road shape set by the basic object traveling direction setting unit 11. An arrow is shown for each block B.
FIG. 3 is a plan view showing the moving direction of the moving object tracked by the object tracking unit of the traffic flow monitoring apparatus according to Embodiment 1 of the present invention.
In FIG. 3, the moving direction of the object is indicated by an arrow for each block.
FIG. 4 is a flowchart showing the escape event detection process of the traffic flow monitoring apparatus according to Embodiment 1 of the present invention.
FIG. 5 is a plan view showing the moving direction comparison of the moving object processed by the object moving direction comparing unit of the traffic flow monitoring apparatus according to Embodiment 1 of the present invention.
In FIG. 5, the moving object of each block on the left side shows the moving direction exceeding the angle error range.

Next, the avoidance event determination operation will be described with reference to FIG.
FIG. 4 is a processing flow of the object tracking unit 12, the object movement direction comparison unit 13, and the traffic event determination unit 14. First, in step S11, the object tracking unit 12 performs object tracking processing individually for each block obtained by dividing the captured image into blocks. The tracking result corresponds to FIG. In step S12, the moving direction and moving speed of the object of each block are calculated from the tracking information of the object obtained in step S11. The movement speed of the block object is a value that can be calculated from information obtained from the lens focal length, depression angle, elevation angle, and installation height of the camera, and the block movement distance per unit.
In step S13, the object movement comparison unit 13 determines the movement direction of the object in the block obtained in step S12 and the temporary movement direction (basic object movement direction) as shown in FIG. ) And a block in which the difference in the moving direction is more than the angle error ± α1 degree (first value) is set as the escape block. FIG. 5 shows a state in which the escape block setting is performed by comparing the basic object movement direction set for each block with the object movement direction obtained by movement tracking. There are four escape blocks on the left.

In step S14, the escape block connection processing unit 14 performs region connection of the escape block obtained in step S13. When a block connected to an arbitrary escape block is an escape block within a speed error V1 km / h and a movement angle difference R1 degree, a connection process between the escape blocks is performed.
In step S15, the avoidance object determination unit 15 detects the avoidance event from the avoidance connection block information obtained in step S14. During the T1 cycle, when the number of connected escape blocks is continuously generated by B2 or more, it is detected that there is an escape moving object on the lane, and the escape event determination is performed. In this step S15, the avoidance determination is performed simply based on the number of consecutive occurrences of the escape block at an arbitrary location in the lane, but the determination criterion is performed in more detail based on the occurrence location of the escape block connection region. May be.

  As described above, according to the first embodiment, by comparing the moving direction of the object in the monitoring lane with the ideal moving direction of the object obtained from the shape of the lane, it is possible to recognize the inside of the lane without recognizing the individual vehicle. It is possible to detect the escape object.

  According to the first embodiment, it is possible to detect a runaway event that could not be detected without recognizing an individual vehicle until now from the flow of the object in the lane. Since the ideal direction of travel is calculated as known information, it is possible to provide a traffic flow monitoring device that determines a avoidance event at a lower cost than the avoidance event determination using information accumulation processing such as statistical processing. It becomes possible.

Embodiment 2. FIG.
A second embodiment of the present invention will be described with reference to FIGS.
FIG. 6 is a block diagram showing a configuration of a traffic flow monitoring apparatus according to Embodiment 2 of the present invention.
In FIG. 6, 11 to 15 are the same as those in FIG. In FIG. 6, a moving block connection processing unit 16 and a reverse running avoidance object determination unit 17 are provided, and the moving block connection processing unit 16 and the reverse running avoidance object determination unit 17 constitute a reverse running avoidance object determination unit.

FIG. 7 is a flowchart showing reverse runaway event detection processing of the traffic flow monitoring apparatus according to Embodiment 2 of the present invention.
Next, reverse run avoidance event detection processing will be described with reference to FIG.
In step S21, the moving block connection processing unit 16 performs region connection of blocks moving at a similar direction and similar speed from the block information tracked by the object tracking unit 12. When the moving direction angle error between adjacent blocks is within ± α2 degrees and the moving speed error is V2 km / h, it is created as a connected block.
In step S22, the lane in which the connection block area exists and the moving direction of the connection block area are confirmed. If the difference between the forward travel direction of the lane where the connected block area exists and the moving direction of the connected block area is (180 ± α3) degrees (second value, where α3 <90) or more, the moving object is a reverse running object Set as.
In step S23, from the connected block area information set in step S22, when reverse running connected blocks are continuously generated during the T2 cycle, it is detected that there is a backward running object on the lane, and the reverse running object is detected. The event is judged as a runaway object in the forward lane. In this step S23, the reverse running determination is simply performed based on the number of consecutive reverse running blocks occurring at an arbitrary location in the lane, but the determination criterion is performed in more detail based on the occurrence location of the escape block connecting region. May be.

  As described above, according to the second embodiment, in the face-to-face lane road, when there is an obstacle that obstructs traveling on the own lane, the obstacle is avoided by traveling on the opposite lane. The reverse running object is detected as the avoidance object in the opposite lane. As a result, it is possible to detect a moving object that does not take the avoidance movement in the surveillance camera on the two-way traffic lane road as an avoidance vehicle, and a wider traffic flow can be measured.

Embodiment 3 FIG.
A third embodiment of the present invention will be described with reference to FIGS.
FIG. 8 is a block diagram showing a configuration of a traffic flow monitoring apparatus according to Embodiment 3 of the present invention.
In FIG. 8, 10 to 17 are the same as those in FIG. In FIG. 8, a movement stop object determination unit 18 (movement stop object determination unit) that receives the output of the movement block connection processing unit 16 and determines a movement stop object is provided.

FIG. 9 is a flowchart showing the moving stop object detection process of the traffic flow monitoring apparatus according to Embodiment 3 of the present invention.
Next, the process of the movement stop object determination unit 18 will be described with reference to FIG.
In step S31, a low-speed moving connection block is detected from the speed information of the connection block information obtained by the moving block connection processing unit 16. During the T31 cycle, B3 or more connected blocks are continuously connected at a speed within V31 km / h, and thus set as a low-speed moving connected block.
In step S32, a stop connected block is detected based on the low speed connected block information obtained in step S31. If a moving object in the block area set as a low-speed moving connection block continues for a period of T32 cycles (predetermined period) and within V32 km / h (predetermined speed), the state is changed from a low-speed moving block to a moving stop block. Set.
This makes it possible to efficiently detect an object that has moved in the lane and stopped.

  From the above processing, a minimum (T31 + T32) cycle is required for moving stop object determination. For example, if T31 = 0.5 seconds, B3 = 2, V31 = 10.0 km / h, T32 = 2.0 seconds, V32 = 5.0 km / h, an object moving at 10 km / h or less is a lane. In the case of stopping within, at least (T31 + T32) = 2.5 seconds later, it is detected as a moving stop object.

  As described above, according to the third embodiment, it is possible to detect an object that moves and stops in a road lane, and a wider traffic flow can be measured.

Embodiment 4 FIG.
Embodiment 4 of the present invention will be described with reference to FIG.
FIG. 10 is a block diagram showing the configuration of a traffic flow monitoring apparatus according to Embodiment 4 of the present invention.
10, 10 to 18 are the same as those in FIG. In FIG. 10, a traffic jam determination unit 19 (traffic jam determination unit) that receives traffic from the moving block connection processing unit 16 and determines traffic jam is provided.

Next, the process of the traffic jam determination unit 19 will be described.
First, the lane average speed is calculated from the connecting block moving speed obtained from the moving block connecting processing unit 16. The accumulated speed Vx of each traveling lane is determined by the accumulation of block moving speeds having speeds other than the movement stop block. Therefore, the object moving average speed Vav in the lane is Vav = Vx / (number of moving blocks + number of stopped blocks).
Next, when the lane average speed Vav is V4 km / h or less during the T4 cycle, it is determined that there is a traffic jam.
As described above, in the fourth embodiment, the traffic jam is determined only by the lane average speed. However, the traffic lane occupation rate is obtained using the difference image or the like, and the traffic jam event is determined from the lane occupation rate and the lane average speed. In this case, the determination accuracy is further improved.

  As described above, according to the fourth embodiment, it is possible to efficiently detect a traffic jam event even under a condition where an individual vehicle is not recognized, and a wider traffic flow can be measured.

Embodiment 5 FIG.
A fifth embodiment according to the present invention will be described with reference to FIGS.
FIG. 11 is a block diagram showing a configuration of a traffic flow monitoring apparatus according to Embodiment 5 of the present invention.
In FIG. 11, 10 to 19 are the same as those in FIG. In FIG. 11, a video multiplexing unit 20 (video multiplexing unit) that switches a captured image of the image capturing unit 10 is provided.

FIG. 12 is a conceptual diagram showing video multiplexing of the traffic flow monitoring apparatus according to Embodiment 5 of the present invention. Next, the video multiplexing unit 20 will be described with reference to FIG. FIG. 12 is an example of a video obtained by time-division multiplexing video from four cameras.
The video multiplexing unit 20 switches video at a rate of once per fixed period TC seconds, and after (TC seconds × number of multiplexed cameras) seconds, cyclic video multiplexing is performed so that the video at the same location is input. Is what you do. Thereby, traffic flow measurement of a plurality of ITV cameras can be performed with one traffic flow monitoring device.

  As described above, in the fifth embodiment, it is possible to measure traffic flow of a plurality of ITV camera images at the same time with one input image, and it is possible to measure traffic flow with reduced costs.

In the fifth embodiment, any one camera video is in a state where there is no video for TC × (number of multiplexed cameras−1) seconds, that is, a non-monitoring state. Events such as avoidance and traffic jams occur continuously regardless of whether they are unmonitored or monitored, so there is no particular problem, but for stopped vehicles, it is possible to detect stopped vehicles that occurred during unmonitored operations. Can not. Therefore, the stopped vehicle is detected indirectly by considering a stopped vehicle that already exists when the camera is switched as a part of an obstacle such as a falling object.
The fallen object detection method is determined based on the presence or absence of the escape vehicle as follows.
If TC x (number of multiplexed cameras-1) and the process returns to its own camera image is defined as one monitoring period, it will move if a runaway event is detected in the monitoring period of H5 times or more during the P5 monitoring period. The fallen object is determined on the assumption that there is a fallen object that causes the object to escape. Since this includes a stopped vehicle that has occurred during the non-monitoring period, it is possible to indirectly determine the stopped vehicle. Thereby, a more accurate traffic flow monitoring apparatus can be provided by also detecting a fallen object and a stop event indirectly from the frequency of occurrence of a runaway event.

  As described above, according to the fifth embodiment, it is possible to monitor more road images by using the video multiplexing device, and other traffic events that occur in the non-monitoring time zone can be monitored. By indirectly detecting from information on continuously occurring traffic events, the traffic flow monitoring device can be provided without limiting the detected events.

It is a block diagram which shows the structure of the traffic flow monitoring apparatus by Embodiment 1 of this invention. It is a top view which shows the temporary advancing direction of each block set by the basic object advancing direction setting part of the traffic flow monitoring apparatus by Embodiment 1 of this invention. It is a top view which shows the moving direction of the moving object tracked by the object tracking part of the traffic flow monitoring apparatus by Embodiment 1 of this invention. It is a flowchart which shows the avoidance event detection process of the traffic flow monitoring apparatus by Embodiment 1 of this invention. It is a top view which shows the advancing direction comparison of the moving object processed in the object moving direction comparison part of the traffic flow monitoring apparatus by Embodiment 1 of this invention. It is a block diagram which shows the structure of the traffic flow monitoring apparatus by Embodiment 2 of this invention. It is a flowchart which shows the reverse run avoidance event detection process of the traffic flow monitoring apparatus by Embodiment 2 of this invention. It is a block diagram which shows the structure of the traffic flow monitoring apparatus by Embodiment 3 of this invention. It is a flowchart which shows the movement stop object detection process of the traffic flow monitoring apparatus by Embodiment 3 of this invention. It is a block diagram which shows the structure of the traffic flow monitoring apparatus by Embodiment 4 of this invention. It is a block diagram which shows the structure of the traffic flow monitoring apparatus by Embodiment 5 of this invention. It is a conceptual diagram which shows the video multiplexing of the traffic flow monitoring apparatus by Embodiment 5 of this invention.

Explanation of symbols

10 image capturing unit, 11 basic object traveling direction setting unit, 12 object tracking unit,
13 object movement direction comparison unit, 14 escape block connection processing unit,
15 avoidance object determination unit, 16 moving block connection processing unit,
17 reverse runaway object determination unit, 18 movement stop object determination unit, 19 traffic jam determination unit,
20 Video multiplexing unit.

Claims (6)

Image capturing means for capturing a road state at a predetermined cycle, object tracking means for tracking the moving direction of an object for each block obtained by dividing the image captured by the image capturing means, and a road for each block of the image Basic object moving direction setting means for presetting the basic object moving direction according to the shape, basic object moving direction for each block set by the basic object moving direction setting means, and for each block tracked by the object tracking means comparing the moving direction of the object for each block, the object moving direction comparing means for moving the direction difference is set a first value or more blocks as避走block, the避走block set by the object moving direction comparing means When adjacent blocks are escape blocks, the escape block connection processing unit that connects the adjacent escape blocks, and the escape block The number of blocks is a predetermined period避走connecting block connected by link unit, traffic flow monitor apparatus characterized by comprising a避走object determination unit determines as避走object to the object when the above predetermined pieces . Image capturing means for capturing a road state at a predetermined cycle, object tracking means for tracking the moving direction of an object for each block obtained by dividing the image captured by the image capturing means, and a road for each block of the image Basic object moving direction setting means for presetting the basic object moving direction according to the shape, basic object moving direction for each block set by the basic object moving direction setting means, and for each block tracked by the object tracking means Object movement direction comparison means for comparing the movement direction of the object for each block, and the result of comparison by the object movement direction comparison means is that the object is determined to be the avoidance object when the movement direction difference is equal to or greater than a first value. object determining means, and comparison result obtained by the object movement direction comparing means, the moving direction difference second value (however, the second value> 90 °) when the above reverse the object Traffic flow monitoring device you comprising the reverse-way traveling避走object determining means for determining as避走object.   The object tracking means calculates a moving speed of the object for each block of the image, and determines a traffic jam by calculating a lane average speed based on the calculated moving speed of the object for each block. 3. The traffic flow monitoring apparatus according to claim 1, further comprising means.   The object tracking unit calculates a moving speed of the object for each block of the image, and detects the moving stop object when the calculated moving speed is equal to or lower than a predetermined speed for a predetermined period. The traffic flow monitoring device according to any one of claims 1 to 3, further comprising:   The image pickup means includes a plurality of image pickup means provided and a video multiplexing means for switching the outputs of the plurality of image pickup means, and the objects from the plurality of image pickup means switched by the video multiplexing means are multiplexed. The traffic flow monitoring apparatus according to claim 1, wherein the traffic flow monitoring apparatus is input to a tracking unit.   6. The traffic flow monitoring apparatus according to claim 5, wherein the movement stop object is determined based on the occurrence frequency of the avoidance object determined by the avoidance object determination means.

Images