JPH0245900A - Method for detecting vehicles caught in traffic jam - Google Patents

Abstract

PURPOSE:To correctly detect a vehicle in a traffic jam by detecting a moving vehicle with an image in which a difference image between two frame images having a time difference is binarized and, thereafter, contracted, and executing an edge detection in the difference image. CONSTITUTION:At the time of extracting the difference image, the difference image is made into a binarized image by a prescribed threshold by a first binarized image preparing part 13, and thereafter, the bright part of the binarized image is contracted at prescribed times by a first contracted image preparing part 14. In such a way, the bright part related to the moving vehicle is left. Thereafter, by a moving vehicle detecting part 15, based on the image outputted from the first contracted image preparing part 14, the existence and the number of white picture elements in an object area are measured, and when the white picture elements exist in a definited number or above, the fact that the moving vehicle exists is judged. When the number of white picture elements is the definite number or below, the fact that the moving vehicle does not exist is judged. In such a way, by detecting the moving vehicle and investigating the moving speed, the degree of a traffic jam is detected.

Description

DETAILED DESCRIPTION OF THE INVENTION FIELD OF INDUSTRIAL APPLICATION The present invention relates to a method of detecting vehicles stopped or congested on a road using image processing technology.

Problems to be Solved by the Prior Art and the Invention In recent years, due to congestion on roads, there has been a growing need to accurately grasp traffic conditions, particularly vehicle stops and traffic jam conditions, and promptly provide them to drivers and the like.

However, in the past, the only methods available were to estimate the length of a line of stopped or congested vehicles using ultrasonic vehicle detectors, etc., but it was not possible to accurately detect the situation of stopped or congested vehicles. Ta.

Therefore, an object of the present invention is to provide a method that can accurately detect vehicles that are stuck in a traffic jam.

Means for Solving the Problems In order to achieve the above object, a method according to the present invention includes fixing a television camera at a position overlooking a detection area on the road, sequentially storing images captured by the television camera in an image storage unit, A difference image between two frame images stored with a predetermined time difference is created, and after the difference image is binarized, it is shrunk a predetermined number of times, and the number of white pixels in the shrunk image is calculated as follows: If the number of white pixels is more than a certain number, it is determined that there is a moving vehicle, and if the number of white pixels is less than a certain number, it is determined that there is no moving vehicle, and when it is determined that there is no moving vehicle, the difference Edges are detected on the image, the edge-detected image is further binarized and then shrunk, and the presence or absence of edges is investigated in the binarized and shrunken image, and if an edge is found, it is determined that there is a ``stagnant vehicle''. At the same time, the position of the stationary vehicle is determined, and if there is no edge, it is determined that there is no vehicle.

In the difference image between the two images stored in the working image storage unit,
Since a moving vehicle appears as pixels with gradation, if the differential image is binarized and compressed to remove various noises, the moving vehicle can be clearly detected.

Furthermore, when there is no moving vehicle, the difference image becomes an image where nothing exists. Therefore, in order to determine whether the difference image in which nothing exists means that there is no vehicle in the detection area at all, or whether it means that the vehicle is stationary, Perform detection. In this way, if an edge is detected,
It is determined that there is a ``stagnant vehicle''.

Example The present invention will be described below based on embodiments shown in the drawings.

FIG. 2 shows the apparatus used to carry out the invention,
The device comprises an image operator (3) equipped with a television camera (2).
), first and second image storage units (4°5), and a central processing unit (10) using a computer.

The television camera (2) is fixed at a position looking down on the detection target area on the road from diagonally above and projects an image of the road in real time. The image processing unit (3) converts the image captured by the television camera (2) into a digital signal corresponding to the luminance gradation and outputs it to the image storage unit (4, 5). In this way, the first and second image storage units ('4
．． In step 5), input images are sequentially stored alternately at constant time intervals. Hereinafter, each image stored in this manner will be referred to as a "frame image."

In order to detect a moving vehicle, the central processing unit (10) includes a difference image creation unit (12), a value image creation unit (13),
A first contraction image creation section (14) and a moving vehicle detection section (15).
). Furthermore, the central processing unit (10) is configured to perform stagnation (
In order to detect the presence or absence of a vehicle (stopped and not moving), an edge detection section (1B), a second binary image creation section (17), a second contraction image creation section (1B), and a stationary vehicle detection section ( 19)
and has.

The method of the present invention will be explained below with reference to FIGS. 1 to 8.

First, the image captured by the television camera (2) is stored in each of the first and second image storage units (4, 5), and the differential image creation unit (12) stores the image in both the storage units (4, 5). A difference image between two stored frame images with a predetermined time difference (for example, 1 second) is created.

Note that this difference image is an absolute value difference, and becomes a different image as described below depending on the vehicle situation in the detection target area. That is, 1) if no vehicle exists, an image of nothing existing; 2) if the vehicle is stationary, an image of nothing existing; 3) as shown in Figures 3 and 4, an image of nothing existing. exists,
In addition, if the vehicle is moving, an image of only the moving vehicle (the part shown by (20) in FIG. 5).

However, even with images 1) and 2) above, due to the shaking of the television camera (2) and noise in the video signal,
Since it is rare to obtain an image completely free of anything, and even in the image of 3), noise (the part indicated by (21) in Fig. 5) due to the causes mentioned above appears. Since the images of 1) and 2) and the image of 3), and the image of the moving vehicle and the noise cannot be clearly distinguished, the following steps are used to perform correction to remove the noise. For example, when the difference image shown in FIG. 5 is extracted, the difference image is first converted into a binary image as shown in FIG. In (14), the bright part of the binary image (
The white pixel portion (W) is contracted a predetermined number of times as shown in FIG. In this way, only the bright area (W) related to the moving vehicle remains. Note that the number of contractions is determined by the magnitude of the noise.

Thereafter, in the moving vehicle detection section (15), based on the image output from the first contraction image creation section (14),
Calculate the presence or absence and number of white pixels in the target area Mj
・When there are more than a certain number of white pixels, "There is a moving vehicle"
It is determined that this is the case, and the process returns to the image storage step. Also, when the number of white pixels is less than a certain number, it is determined that there is no moving vehicle, and
Proceed to the following process. Note that by detecting moving vehicles in this manner and investigating their moving speeds, the extent of traffic congestion can be detected.

When it is determined that no moving vehicle exists on the image as described above, the edge detection unit (16) Edge detection is performed on the difference image. Many edge detection methods are known. For example, after creating an image by shifting the difference image by several pixels in the vertical and horizontal directions, the difference between the difference image and the image by which the pixels have been moved is You may also use a method of taking the difference between .

This method using pixel movement has the advantage of being easy to implement in hardware using a shift register or the like, and of high processing speed.

Note that no matter which edge detection method is used, the obtained image is a multivalued image and contains various types of noise.
In the second binary image creation section (17), the image is binarized using a predetermined threshold, and then the second contracted image creation section (18
), noise is removed by contracting bright areas of the binarized image. The image obtained in this way is 1) an image without any edges when there is no vehicle in the detection area, and 2) an image without any edges when there is a stationary vehicle in the detection area, as illustrated in FIG. In this image, the edges of the stationary vehicle are detected. Finally, the stationary vehicle detection unit (19) investigates the presence or absence of an edge in the image, and if an edge is present, it is determined that there is a stationary vehicle, and the location of the stationary vehicle is determined. Furthermore, when there is no edge, it is determined that no vehicle exists in the detection target area.

Effects of the Invention As described above, the method according to the present invention uses image processing technology, and uses a contracted image after binarizing a differential image between two frames with a time difference to detect a moving vehicle. The presence or absence of a parked vehicle is detected by detecting edges in the differential image.

With such a configuration, the present invention has the effect of being able to accurately detect stopped vehicles and greatly contributing to alleviating road congestion.

[Brief explanation of the drawing]

FIG. 1 is a flowchart of the method of the present invention, and FIG.
A block diagram of an apparatus used to carry out the method; FIGS. 3 and 4 are diagrams illustrating two images stored in the first and second image storage units; FIG. ■１
6 is a diagram showing a binary image created from the difference image, FIG. 7 is a diagram showing a contracted image of the binary image, and FIG. 8 is a diagram showing a stopped vehicle. FIG. 3 is a diagram illustrating an image in which edges are detected for detection. (2): TV camera, (4): First image storage section, (5): Second image storage section. Figure: To the Commissioner of the Japan Patent Office 1゜Indication of the case Patent Application No. 197364/1983 2゜Name of the invention Method for detecting stopped and congested vehicles 3゜Relationship with the case

Claims (1)

[Claims] [1] A television camera (
2) is fixed, the images projected by the television camera (2) are sequentially stored in the image storage section (4, 5), a difference image between the two frame images stored with a predetermined time difference is created, and the difference image is After the image is binarized, it is shrunk a predetermined number of times, the number of white pixels in the shrunken image is measured, and if the number is greater than a certain number, it is determined that there is a moving vehicle, and the number of white pixels is a certain number. In the following cases, it is determined that there is no moving vehicle, and when it is determined that there is no moving vehicle, edge detection is performed on the difference image, and the edge-detected image is further binarized and then shrunk. , Investigating the presence or absence of an edge in the binarized and shrunk image, and when an edge is present, determining that there is a "stagnant vehicle" and determining the position of the stalled vehicle, and when there is no edge, determining that there is "no vehicle". A method for detecting stopped and congested vehicles.