JPH0869596A - Traffic monitoring device - Google Patents

Abstract

PURPOSE: To automatically detect an event like a snarl, an accident or a fire by correcting positions, areas, and changes with time on a picture face of individual moving bodies extracted from an input picture by a distance table to calculate the speeds and sizes of individual moving bodies. CONSTITUTION: A moving body running on a road is photographed by a camera 2, and the observation face obtained from the actual position of the photographed background picture and the position on the picture and the picture face are allowed to correspond to each other by the distance table generated by a distance table generation means 35. Positions, areas, and changes with time on the picture face of individual moving bodies extracted from the input picture obtained by the camera 2 are corrected by the distance table. After this correction, speeds and sizes of individual moving bodies are calculated by a calculation means 6. Thus, an event like a snarl, an accident, or a fire is automatically detected.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a traffic monitoring device for monitoring traffic of a moving body.

[0002]

2. Description of the Related Art Conventionally, when observing a vehicle passing on a road, an observer watches a monitor image to monitor the state of the vehicle. The state is monitored by measuring the number and speed of vehicles by image processing.

[0003]

In the conventional traffic monitoring by image processing, when the camera installation position, the installation angle, and the lens angle of view change, a distance table that associates the road surface with the image surface based on those data. Was calculated geometrically and it was troublesome.

When detecting a traffic volume or an abnormal event by image processing when a visible camera is used at night, it is detected based on lights such as headlights and taillights. However, there was a problem that the measurement system of a vehicle with multiple lights turned on deteriorated.

The present invention solves these problems.
After extracting and calculating the number of moving objects from the images of the moving objects taken by the camera, the actual size and speed are calculated by calibration, and based on this, the number of moving objects passing, the speed, the measurement between vehicles, and traffic congestion The purpose is to realize automatic detection of events such as accidents and fires.

[0006]

FIG. 1 shows a block diagram of the principle of the present invention. In FIG. 1, a traffic monitoring device 1 monitors a moving object traveling on a road, and is composed of a camera 2, an image processing device 3, and the like.

The camera 2 captures an image of a moving body. The image of the moving body taken by the camera 2 is input to the image processing device 3 in real time, or
Recorded once on video tape 23 by 23, and later VT
It may be reproduced by R24 and input to the image processing apparatus 3.

The image processing device 3 is a distance table creating means 3 for creating a distance table based on an image from the camera 2.
In addition to 5, the input image is image-processed to calculate the speed, size, number of passing vehicles, etc., various events are detected and notified to the administrator, various alarms are displayed, etc. , And 35, 4 to 7.

The distance table creating means 35 creates a distance table that associates the observation surface with the road surface from the positions on the image of the markers whose actual positional relationship is known in the background image.

The cutting-out means 4 cuts out an image within a preset mask setting range from the image taken by the camera 2. The moving body extraction means 5 extracts the moving body and the size which move on the image cut out by the cutting means 4.

The calculating means 6 calculates the speed of the moving body extracted by the moving body extracting means 5. The determining means 7 determines the inter-vehicle distance based on the position of the moving body and the positions of other moving bodies, the calculated number of passing moving bodies per unit time, images taken by a plurality of cameras 2, and the like. Various kinds of events such as determining traffic congestion, determining the beginning of traffic congestion, and the like are determined.

[0012]

According to the present invention, as shown in FIG. 1, with respect to a background image containing a marker whose actual positional relationship is photographed by the camera 2, the observation surface and the image surface are associated from the position of the marker on the image surface. A distance table is created by the distance table creating means 35, and an image of the moving body taken by the camera 2 is created, and the cutout means 4 cuts out an image within a preset mask setting range from this image to extract the moving body. A moving body that the means 5 moves on the cut-out image,
The size and the number of passing vehicles are extracted, and the calculating means 6 calculates the speed of the extracted moving body.

At this time, the moving body extracting means 5 calculates a difference image between the cut-out image and the background image registered in advance, and then calculates a projection value in a direction perpendicular to the moving direction to end the projection value. Alternatively, the head is extracted as the position of the moving body, the size of the moving body is calculated from the magnitude of the projection value, and the number of moving bodies (the number of passing vehicles) is calculated from the number of the projection values.

Further, a marker image such as a lamp of a moving body is photographed by the camera 2, the cutting means 4 extracts a marker which moves on the marker image, and the moving body extracting means 5 extracts the distance between the extracted markers. Is a predetermined value or less, it is determined that they are on the same moving body and they are integrated. On the other hand, when is a predetermined value or more, it is determined that they are on different moving bodies, and the calculating means 6 determines the marker or the integration. The velocity of the moving body composed of the plurality of markers is calculated.

Further, when it is detected that the speed of the moving body calculated by the event judging means 7 is gradually decreasing, the congestion occurrence prediction display is performed. Therefore, after extracting the image of the mask setting range from the image of the moving body taken by the camera 2 and extracting and calculating the number of moving bodies, it is possible to calibrate and calculate the actual size and speed. It will be possible to measure the number of vehicles passing through the vehicle, speed, distance between vehicles, and automatic detection of events such as traffic jams and accidents.

[0016]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, the construction and operation of an embodiment of the present invention will be described in detail with reference to FIGS.

FIG. 2 shows a flowchart for explaining the operation of the present invention. This is a procedure when the moving body traveling on the road is photographed by the camera 2 and the image processing is performed to calculate the number, size, and speed of the moving body under the configuration of FIG.

In FIG. 2, S1 captures an image of a moving body. This captures an image of a moving object moving on the road by the camera 2 of FIG. In S2, the image in the mask setting range is cut out. This cuts out only the image within the mask setting range preset as the range in which the moving body travels from the image including the moving body on the road captured in S1.
For example, as shown in FIG. 3 to be described later, only the image within the mask setting range for the left lane processing on the left side and the right lane processing for the right side is cut out across the central broken line, and the image processing range is set to the necessary minimum image processing. Reduce the amount.

In step S3, the moving body is extracted and the quantity is calculated. This is a difference image ((b) of FIG. 5) obtained by taking a difference between the image cut out in S2 shown in (a) of FIG. )
Then, the projection value obtained by projecting this difference image in the direction perpendicular to the traveling direction (horizontal direction in FIG. 5B) is obtained, and the head or the tail (or the average of both) of this projection value can be used for the moving body. The position is set in the moving body position table 9 every predetermined time, and the length of the projection value is set to the size of the moving body. From these, the position, size, and number of the moving body on the image are obtained.

In step S4, the size and movement of the moving body are calibrated. This calibrates the position and size of the moving body on the image obtained in S3 to the actual position and size on the road. As shown in FIG. 4, which will be described later, this calibration is performed by determining the pixel and the length of the mark in the image captured by the camera 2 of a previously known mark on the road (for example, a lateral line drawn at a predetermined distance on the road). The correspondence of is asked. For example, as shown in FIG. 4, a calibration value is obtained and registered in the distance table 8 (see FIG. 9). The length (number of pixels) on the image is calibrated to the actual length (m) by referring to the distance table 8 to obtain the actual size and movement of the moving body.

In step S5, the speed and size are calculated. This calculates the speed and size of the actual moving body calibrated in S4. As described above, the moving body moving on the road is photographed by the camera 2, only the mask setting range is cut out from the photographed image, and the difference image between the cut-out image and the background image is obtained. Find the value,
The head, the tail, or the average of these projection values are registered as the position of the moving body in the moving body position table 9 in association with the time,
The speed, size and number of the moving body on the image are obtained and calibrated to obtain the size and speed of the moving body on the actual road. As a result, it is possible to capture an image of a moving body traveling on the road by the camera 2 and detect the number of passing moving bodies, the size, and the speed. Details will be sequentially described below with reference to FIGS. 3 to 11.

FIG. 3 shows an example of the mask setting range of the present invention. This is a mask setting range set in advance from an image on the road taken by the camera 2, here, for extracting the center broken line, for right lane processing, and for left lane processing. Cut out only the image. In the following description, the image for the right lane processing and the image for the left lane processing are used to perform image processing to extract a moving object, reduce the image to be processed to a necessary minimum, and increase the processing speed.

FIG. 4 shows an explanatory diagram of the image calibration of the present invention. This is a calculation of the calibration value by obtaining the correspondence between the length on the image taken on the road by the camera 2 and the actual length on the road. Here, as shown,
Length on the image (number of pixels) and actual length on the road (m)
Find the correspondence with. The correspondence is determined by taking a line on the road at a predetermined interval, for example, a line drawn at a predetermined interval on the road with the camera 2, and measuring the length of the line interval (number of dots) on the taken image. And the actual length of the line interval on the road are found. For example, the actual length (m) on each position on the image is calculated and registered in the distance table 8 like the length (dots) actual length (m) 58 on the image shown. Keep it. Accordingly, if the length (dot number) of the moving body is known on the image, it can be calibrated to the actual length (m) on the road.

FIG. 5 is an explanatory diagram for detecting a moving body. Figure 5
(A) indicates an image taken by the camera. FIG. 5B shows an image after the moving body in the left lane is extracted. This is an image obtained by cutting out only the image within the mask setting range for left lane processing of FIG. 3 from the image of FIG. This is an image after obtaining a difference image of the difference from the background image that was present and extracting only the moving body in the left lane.

FIG. 5C shows the projection value in the horizontal direction. This is for the image for left lane processing in FIG. 5B, which is projected in the direction perpendicular to the traveling direction, here in the horizontal direction,
The projection value after binarizing the value after the projection with a predetermined threshold value is shown. The portion 1 has a moving body, and the portion 0 has no moving body. The head position, the tail position, or the average position of the two values is determined as the position of the moving body in the moving direction of the projection value. In addition, the length of the moving body is represented by the length in the range where the projection value is 1. The actual length of the moving body on the road is calibrated to the actual length of the moving body on the road using the values (distance table 8) registered in FIG.

FIG. 6 shows an example of the installation position of the camera of the present invention. FIG. 6A shows a side view. The camera 2 is installed, for example, at a height of 10 m on the overpass across the road, adjusted to the angle as shown in the drawing toward the traveling direction of the moving body, and the observation section is set to about 100 m. To do.

FIG. 6B shows a top view. Camera 2
Is installed right above the central broken line on the overpass with a height of about 10 m as shown in the figure. Here, the moving body traveling on each lane of two lanes is photographed at the same time for the left lane and the right lane. Then, the two are image-processed independently, and the number of moving bodies, length, speed, etc. are calculated in parallel.

FIG. 7 shows a flow chart of image cutting of the mask setting range according to the present invention. This is a detailed flowchart of S2 in FIG. In FIG. 7, in S11, an image of the moving body is captured. This captures an image of a moving body traveling on a road by the camera 2 shown in FIG.

S12 is displayed on the display. S
13 detects the central broken line and side line on the road. This detects the central broken line and side lines on the road, for example, a white line on the screen imaged in S11.

In step S14, the area between the central broken line and the side line is determined as the mask setting range. This is a mask setting range for the area between the central broken line and the side line of the white line detected in S13 (mask setting range for left lane processing and right lane processing).
To decide.

In step S15, a confirmation screen is displayed on the display. This is because the mask setting range for cutting out the image for left lane processing and the mask setting range for cutting out the image for right lane processing are displayed on the display. Confirm whether or not. In the case of YES, since it has been found that the operator has instructed that the automatically set mask setting range is correct, in S18, the mask setting of the mask setting range for the left lane processing and the right lane processing (further for extracting the central broken line) Register each range. On the other hand, in the case of NO, the operator corrects in S17 and registers in S19.

As described above, the mask setting ranges for the left lane, the right lane, and the central broken line extraction are registered from the image captured by the camera 2. After that, only the image of the registered mask setting range becomes the image for extracting the central broken line, for the left lane processing, and for the right lane processing.

FIG. 8 shows a flow chart for extracting a moving body and calculating a quantity according to the present invention. In FIG. 8, in S21, an image of the moving body is captured. This is because the camera 2 in FIG. 1 captures an image of a moving object on the road.

In step S22, the image in the mask setting range is cut out. As shown in FIG. 5B, this cuts out an image within the mask setting range, for example, an image for left lane processing from the image.

In step S23, the difference between the previously registered background image and the input image is calculated. S24 is binarized. This binarizes the difference image calculated in S23 with a predetermined threshold value.

In step S25, the projection value is calculated. For example, as shown in (c) of FIG. 5, a projection value obtained by projecting the binarized difference image of (b) of FIG. 5 in the direction orthogonal to the traveling direction is calculated.

In step S26, the tail portion of the moving body is detected. This detection is carried out by scanning the projection value in the traveling direction from 0 to 1.
Detect points that change to. The detected position is registered in the moving body position table 9 in association with the time.

In step S27, the quantity is counted. this is,
The number of projected values, that is, the number of moving objects (the number of passing objects) is counted, and the size of the moving object (the size of the portion of the projected value of 1) is counted and recorded.

In step S28, it is determined whether or not it has ended. This determines whether the processing has been completed for all projection values. If YES, then end. In the case of NO, the next image is input for each predetermined frame in S29, and S23 and subsequent steps are repeated.

As described above, with respect to the image cut out within the mask setting range, the difference image between the image and the background image is binarized and the projection value is calculated for each predetermined frame, and the projection value is scanned to move the position of the moving body. Is detected and registered in the moving body position table 9 in association with the time, and the size of the moving body on the image is calculated. As a result, the number and size of passing mobile bodies can be obtained on the image, and the position can be registered in the mobile body position table 9 in association with the time.

FIG. 9 shows an explanatory diagram of image calibration according to the present invention. FIG. 9A shows a flowchart. In FIG. 9A, in S31, a background image with a marker is captured. This captures the background image with the side line whose distance of the white line is known as a marker.

At S32, the marker portion is cut out.
In S33, the number of pixels on the image of the marker portion is calculated.
In S34, the distance is converted into the length per pixel of the marker portion and stored in the distance table 8. This is, for example, in (b) of FIG.
As shown in, a number is given in the traveling direction of the pixel on the image displayed on the screen, and the actual position (distance) of the moving body on the road is registered in association with this pixel number. Specifically, the actual length (m) is divided by the length (the number of pixels) on the image shown in FIG. 4 described above, and the actual length of the moving body per pixel after the division is calculated as the pixel. I ₀ corresponding to the number,
Registered in the distance table 8 as I ₁ , I _2, ...

FIG. 9B shows a distance table. This is obtained by obtaining and registering the actual length (m) of the moving body on the road in association with the pixel number on the image. As described above, it becomes possible to calibrate the length on the image captured by the camera 2 to the actual length (m) of the moving body on the road by referring to the distance table 8.

FIG. 10 is a diagram for explaining the velocity calculation of the moving body of the present invention. FIG. 10A shows a flowchart.
In FIG. 10A, in S41, the time and position information for each moving body is fetched, and the moving body position table 9 is created based on the distance table 8. This is done by referring to the distance table 8 in which the information on the time and position of each moving body is fetched and the correspondence between the pixels on the image of FIG. 9B and the actual position on the moving body is registered. , The time t ₀ , t ₁ , t of the moving body
The position _{y 0, y 1, y 2} ··· mobile in ₂ ... to create and store the vehicle location table 9 of FIG. 10 (b).

In S42, the speed is calculated based on the moving body position table 9. This is based on the moving body position table 9 of FIG. 10B created in S41, as shown in FIG. 10C, at times t ₀ , t ₁ , t _2, ... The positions y ₀ , y ₁ , y _2, ... Of the moving body at that time are approximated by a straight line, and the velocity that is the inclination of the straight line is calculated.

As described above, the time and position information of each moving body on the image is fetched, and the actual time and position relationship of the moving body is obtained by referring to the distance table 8 in FIG. It is possible to create the moving body position table 9 of FIG. 10B and calculate the speed of each moving body based on this moving body position table 9.

FIG. 10B shows a moving body position table. This moving body position table 9 is created by associating the actual position of the moving body with the time according to the flowchart of FIG.

FIG. 10C shows an example of the speed calculation method. This is one in which the time of the moving body position table 9 in FIG. 10B and the position of the moving body are connected by a straight line, and the inclination at that time is obtained and used as the speed.

FIG. 11 shows a flow chart for calculating the size of the moving body of the present invention. In FIG. 11, S51 calculates the projection value of the moving body at an arbitrary time. This is to calculate the projection value calculated in S25 of FIG. 8 described above, that is, the projection value obtained by projecting the difference image between the image of the moving body and the background image in the direction perpendicular to the moving direction, at any time. .

S52 is calibrated based on the distance table. This is because the projection value calculated in S51 is calculated by the distance table 8
It is calibrated to the actual size of the moving body based on. As described above, the actual size (length) of the moving body in the moving direction can be calculated from the projected value of the image.

Next, applications in the field of road management will be sequentially described with reference to FIGS. FIG. 12 shows the vehicle speed and inter-vehicle distance measurement statistics of the present invention, and an explanatory diagram of detection / occurrence prediction of abnormal events.

FIG. 12A shows a flowchart. In (a) of FIG. 12, S101 calculates the vehicle speed and the inter-vehicle distance. As described above, the vehicle speed and the inter-vehicle distance of the vehicle are calculated from the image.

In step S102, the average value at any time is calculated in real time. S103 outputs to a graph. In S104, the feature of the inclination of the graph is detected. The patterns are classified into pattern (a), pattern (b), pattern (c), pattern (d), and others according to the characteristics of the detected pattern of the inclination of the graph, and the judgment or prediction corresponding to the pattern is performed.

S105 is the time of the pattern (a), and as shown in the pattern (a) of FIG. 12B, the vehicle speed and the inter-vehicle distance are both fast and long. to decide.

Step S107 is the time of the pattern (b), and as shown in the pattern (b) of FIG. 12B, the vehicle speed and the inter-vehicle distance are both slow and short. To judge.

S109 is the time of the pattern (c), and as shown in the pattern (c) of FIG. 12B, both the vehicle speed and the inter-vehicle distance are gradually slowed and shortened. , S110 predicts the occurrence of congestion.

S111 is the time of the pattern (d), and as shown in the pattern (d) of FIG. 12B, the vehicle speed and the inter-vehicle distance suddenly become almost zero. And determine one of the following.

-End of traffic jam-Accident occurrence-Falling object occurrence-Other disaster occurrence In step S113, except for patterns (a) to (d), nothing is done here.

FIG. 12B shows a graph pattern example. Here, the solid line represents the vehicle speed, and the dotted line represents the vehicle distance. Pattern (a): a normal pattern when the vehicle speed and the vehicle distance are both fast and long.

Pattern (b): a pattern in which the vehicle speed and the distance between vehicles are both slow and short, and the traffic is congested. Pattern (c): a pattern in which the occurrence of traffic congestion is predicted when the vehicle speed and the distance between vehicles are gradually slowing and shortening.

Pattern (d): a pattern in which an abnormal situation such as an accident occurs when the vehicle speed and the vehicle distance suddenly become almost zero. FIG. 13 shows a vehicle speed and inter-vehicle distance measurement statistics and an abnormal event detection / occurrence prediction explanatory diagram of the present invention.

FIG. 13A shows a flowchart. In (a) of FIG. 13, S121 calculates the number of passing vehicles. This measures the number of projection values mentioned above,
Calculate the number of passing vehicles.

In S122, the number of vehicles in arbitrary time units is statistically calculated in real time. S123 outputs to a graph.
In S124, it is determined whether or not there is a sudden change in the statistical value. Y
In the case of ES, the occurrence of an event such as an accident is estimated in S125. On the other hand, if NO, the process proceeds to S126.

In step S126, it is determined whether the statistical value has changed. In the case of YES, the occurrence of an event such as traffic jam is predicted in S127. On the other hand, in the case of NO, since it is normal, the process ends.

FIG. 13B shows an example of graph output. The horizontal axis represents time in units of 1 minute, and the vertical axis represents the number of vehicles passing in units of 1 minute. The solid line represents all vehicles, the dotted line represents large vehicles, and the alternate long and short dash line represents small vehicles. Here, it is determined that an accident or the like has occurred when there is a sudden change in the central statistical value and the number of vehicles passing through becomes zero.

FIG. 14 is a diagram for explaining the prediction of the head / tail position of the traffic jam of the present invention. FIG. 14A shows a flowchart. In FIG. 14A, S151 detects traffic congestion.

In step S152, it is determined whether or not images from other cameras adjacent to the front and the rear detect congestion.
If YES, S152 and S154 are repeated. On the other hand, in the case of NO, the head and tail positions of the traffic jam are estimated, and the length of the traffic jam is estimated in S156.

FIG. 14B shows an explanatory diagram. Here, a plurality of cameras 2 are installed along the traveling direction of the vehicle, and the traffic congestion described above is detected based on the images from the cameras 2. Then, it is determined that the head and the tail of the traffic jam are located between the position where the traffic is no longer detected from the image captured by any of the cameras 2 from the head and the tail in the traveling direction and the position of the camera 2 immediately before the position. As a result, the length of the traffic jam, the occurrence of an abnormal situation such as an accident in the blind spot of the camera, and the rough occurrence position are determined.

FIG. 15 is a diagram for explaining detection of a moving body by the marker of the present invention. FIG. 15A shows a flowchart. In FIG. 15A, in S171, a marker is extracted and a marker position table is created. This is created, for example, by extracting the tail lamp of a vehicle traveling at night from an image taken by the camera 2 as the marker of the moving body, and storing the time and position in the marker position table 10.

In S172, the distance between the adjacent markers at a certain time is calculated. In S173, it is determined whether it is smaller than a certain value. In the case of YES, it is determined in S174 that the marker group is on the same moving body, and S172 and subsequent steps are repeated in order to compare the distances with the adjacent markers. As a result, all the distances between the markers that are smaller than a certain value are determined to be one moving body and are integrated. For example, in the case where a plurality of lamps such as the track shown in FIG. 22A described later are turned on, the markers corresponding to the respective lamps are collectively determined as one track. On the other hand, N in S173
In the case of O, the process proceeds to S175.

In S175, it is determined that the markers are on different moving bodies. This is because the distance between the markers was found to be larger than a certain value, so it was determined that the markers were on different moving bodies.

In S176, one moving body is counted.
As described above, the marker is extracted from the image obtained by photographing the lamp of the vehicle traveling at night with the camera 2, and the marker is registered in the marker position table 10.
Then, the positions of the adjacent markers at the same time are taken out to obtain the distance between them, and when the distance is smaller than a certain value, it is integrated, and when it is found to be larger than a certain value, it is judged as that of another moving body. Repeat the process to integrate the markers. It is possible to register the position after integration (for example, the position of the leading marker, the average, or the position of the trailing marker) and the time in the moving body position table 9 described above, and count the number of passing vehicles.

FIG. 15B shows a marker position table. This is created in S171 of FIG. 15A, and the position and time of the marker are extracted and registered from the image captured by the camera 2 of the vehicle with the tail lamp lit.

FIG. 16 shows an integrated explanatory diagram of the marker of the present invention. FIG. 16A shows the input image (1). This is an image taken by the camera 2 of a truck or the like in which a plurality of lamps are lit, and the mark * indicates the marker.

FIG. 16B shows a state after the moving body of the marker is extracted from the image of FIG. In this case, as described on the right side, the distances l and m between the markers are each smaller than a certain value, so that a total of 6 markers are set to 1
This is an example in which it is integrated into a group and judged as one unit.

FIG. 16C shows the input image (2). This is an image obtained by photographing two passenger cars with a plurality of lamps turned on by the camera 2, and the part of * is a part of the marker.

FIG. 16D shows a state after the moving body of the marker is extracted from the image of FIG. 16C. In this case, as described on the right side, the distance n between the markers is larger than a certain value, and thus it is an example in which it is determined that the number is two.

[0078]

As described above, according to the present invention,
You can create a distance table from the images taken by camera 2,
It is no longer necessary to physically measure the camera position, angle, etc. as in the past, and the distance table is used to extract the moving objects from the images of the moving objects and calculate the quantity, and then calibrate the actual size and speed. Can now be calculated.

Also, in nighttime traffic monitoring, the number of vehicles passing through can be measured from the image of the visible camera.

[Brief description of drawings]

FIG. 1 is a principle block diagram of the present invention.

FIG. 2 is a flowchart explaining the operation of the present invention.

FIG. 3 is an example of a mask setting range of the present invention.

FIG. 4 is an explanatory diagram of image calibration according to the present invention.

FIG. 5 is an explanatory diagram of detecting a moving body of the present invention.

FIG. 6 is an example of setting positions of the camera of the present invention.

FIG. 7 is a flowchart for cutting out an image of a mask setting range according to the present invention.

FIG. 8 is a flowchart of mobile unit extraction and quantity calculation of the present invention.

FIG. 9 is an explanatory diagram of image calibration according to the present invention.

FIG. 10 is an explanatory diagram of speed calculation of a moving body according to the present invention.

FIG. 11 is a flowchart for calculating the size of a moving body according to the present invention.

FIG. 12 is a diagram illustrating statistics of vehicle speed and vehicle distance measurement and detection / occurrence of an abnormal event according to the present invention.

FIG. 13 is an explanatory diagram of statistics of vehicle speed and inter-vehicle distance measurement and detection / occurrence of abnormal events according to the present invention.

FIG. 14 is an explanatory diagram for predicting the start / end positions of a traffic jam according to the present invention.

FIG. 15 is an explanatory diagram for detecting a moving body by the marker of the present invention.

FIG. 16 is an explanatory diagram for integrating markers according to the present invention.

[Explanation of symbols]

 1: Traffic monitoring device 2: Camera 3: Image processing device 35: Distance table creating means 4: Cutting means 5: Moving body extracting means 6: Calculating means 7: Judging means 8: Distance table 9: Moving body position table 10: Marker position table

Continuation of front page (51) Int.Cl. ⁶ Identification number Office reference number FI Technical display area H04N 7/18 D

Claims (3)

[Claims] 1. A traffic monitoring device for monitoring traffic of a mobile body, wherein a camera (2) for capturing an image of the mobile body, an actual position of a background image captured by the camera (2) and a position on the image. By calibrating the distance table that correlates the observation surface and the image surface obtained from, and the position, area, and temporal change on the image surface of each moving object extracted from the input image by the traffic monitoring device with the distance table, And a calculating means (6) for calculating the speed and size of the moving body. 2. A traffic monitoring device for monitoring traffic of a moving body, wherein a camera (2) for photographing a marker image such as a lamp of the moving body, and a moving body extracting means for extracting a marker moving on the marker image ( 5) and the marker extracted by the moving object extracting means (5) are integrated on the same moving object when the distance is less than or equal to a predetermined value. A moving body extracting means (5) for judging that the moving body is on the body; and a calculating means (6) for calculating the speed of the moving body composed of the judged marker or a plurality of integrated markers. Traffic monitoring device. 3. A traffic jam is displayed when it is detected that the calculated speed of the moving body is slower than an average value, and a traffic jam prediction display is displayed when it is detected that the speed is gradually decreasing, or The event determination means (7) for determining a failure such as an accident when the speed is suddenly reduced to be extremely slow or stopped is provided, and the event determination means (7) is provided. Traffic monitoring device.