JPH0658218B2 - Rear detection device for moving body - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Use of the Invention] The present invention captures a moving body as a two-dimensional image with an ITV camera or the like, performs various image processing on the image, and detects the number of the moving body or the like. Regarding how to do.

[Background of the Invention]

In general, there is an increasing need for recognizing the number and shape of a moving body such as a vehicle by an ITV camera and an image processing device regardless of whether they are available.

Conventionally, as shown in FIG. 3, a sensor such as a special infrared ray 40, a magnetism 41, an ultrasonic wave 42 or the like was used to determine an optimum timing for taking in, but it is economical to install the sensor or the like on site, Since there are practical defects, a method for judging and incorporating only image data as described in Japanese Patent Application No. 59-98428 was invented as a solution to these problems. However, the method is to take a picture of the moving body from the front or the upper part, and if 24-hour monitoring is required, haloing will occur due to the headlight at night, etc. There was a fear that the optimal import timing would be incorrect.

[Object of the Invention]

An object of the present invention is to perform image processing of an image of a rear part of a moving body such as a vehicle, thereby enabling image capturing at night, which is a drawback of the conventional forward photography method, and enables real-time high-speed recognition during day and night. An object of the present invention is to provide a rear detection device for a movable body.

[Outline of Invention]

In order to detect the moving body entering from the lower part of the imaging visual field, the present invention examines the density change in the independent area of the lower part of the screen in order to correspond to the monitoring at night. As a result, the image is input at the optimum timing, and the day and night monitoring can be realized.

Example of Invention

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a diagram showing a configuration system of the present invention, and FIG. 2 is a diagram showing a hardware configuration according to the present invention. As shown in FIG. 1, the ITV camera 10 has a height H
Install at a distance of. Reference numeral 100 denotes a moving body, and 200 denotes an image processing device. Then, when the angle formed by the center point O on the traveling road surface and the traveling road surface at the viewing angles 2φ and 2φ of the ITV camera 10 is θ, the following relational expression is established.

L ₀ = H / sin θ L ₁ = H / tan θ L ₂ = H / tan (θ + φ) L ₃ = H / tan (θ−φ) L ₄ = L ₃ −L ₂ L ₅ = L ₃ −L ₁ Specific Example Is the angle θ between the ITV camera 10 and the traveling road surface.
= 30 °, ITV camera height H = 5m, D = 2.0m,
And, as a matter of course, the actual speed of the vehicle is not constant, so assuming that the traveling speed V = 70 km / h, L ₀ = 10.0 m, L ₁ = 8.7 m, L ₂ = 5. 7m, L
₃ = 14.8 m, L ₄ = 9.0 m, L ₅ = 6.0 m, φ = 11.
3 °, so the vertical field of view is L ₄ m,
When the vehicle passes through this period, it takes T = L ₄ / V = 128 × 10 ⁻³ sec. Further, when shooting with an ITV camera (non-interlaced mode), it takes about 16 ms to obtain one image. That is, when the vehicle passes at the above speed, there is a 1/8 screen movement until the next image is obtained.
This means that, for example, when the vehicle height is hm, eight images are obtained until the rear end of the vehicle passes through the screen. Therefore, while capturing each image one by one, ITV
Whether the moving object is input into the imaging field of view online by performing a series of processing at high speed, such as binarizing the grayscale data from the camera and taking the frequency distribution of "1" for that image. Can be determined.

An image processing apparatus as shown in FIG. 2 is conceivable as one configuration for realizing such processing.

As shown in the figure, the AD converter 20 takes in the read video signal from the ITV camera 10 and AD-converts it. Selector 2
1 normally selects the AD output of the AD converter 20. The image processing processor 22 takes in the output of the selector 21 and performs preprocessing.

The preprocessing is to remove noise and perform spatial differentiation. There are two types of images to be processed by the image processing processor 22. The first is an image for detecting whether or not the moving body 100 has input within the imaging visual field. The second is a whole image of the moving body obtained when the moving body 100 itself inputs all in the imaging field of view. This second image is captured by the timing signal obtained from the first image,
The image of the entire screen is stored in the image memory 24.

In the method of determining the input of the moving object 100 using the first image, after preprocessing such as noise removal is performed by the image processing processor 22, the binary value is binarized by the threshold value preset by the binarization circuit 23. And the binary frequency distribution (the number of pixels of "0" and "1") of the output image is obtained, and the input of the moving object is determined based on the value, or the image processing processor 22 performs preprocessing. As a result, the density frequency distribution is directly obtained by the frequency distribution accumulation processor 25, the density average level, the density dispersion, etc. are obtained, and based on the values, a similar determination method can be carried out.

Next, a specific example of the processing will be described with reference to FIGS. Third
As shown in the figure, when the moving body 100 does not exist in the screen and enters at a speed of NKm / h from below, the density average level, the density dispersion, and the area when binarized according to the color and brightness. Etc. changes from the case where it does not enter. In FIG. 3, reference numeral 40 is an infrared sensor, 41 is a magnetic sensor, and 42 is an ultrasonic sensor. Referring to FIG. 4,
The density data of the area 50 of (a) which is an independent area in the lower part of the screen changes due to the entry of the moving body. Due to this change, the input within the field of view of the moving body is detected ((b)), and then the change from (c) to (d) occurs, and the same density data as (a) is detected in the same region 50. The moment is determined as the timing for capturing the optimum image. Here, 101 is a front part of the moving body, 102 is a middle part of the moving body, and 103 is a rear part of the moving body.

By using such a method,
It is also possible to detect moving objects such as vehicles at night, and to detect moving objects more reliably.

Next, the processing flow chart will be described with reference to FIG.

First, an image is input by the image input subroutine, it is judged that the density frequency distribution of the image is above the threshold value as the input of the moving body, and the moment when the density frequency distribution after that is below the threshold value is the optimum image. Therefore, the image is captured and the image is processed later.

Next, another embodiment will be described. In order to speed up the processing, it is necessary to know the optimum image capture timing early. However, as shown in FIG. 6, the raster scanning of the current TV camera is 1 with the upper left of the screen as the starting point 60.
The line is scanned line by line, and the bottom right is the end point 61.
Therefore, in order to obtain information in a small area at the bottom of the screen,
It is delayed about 16 msec from the scanning start time (non-interlaced mode). Therefore, with the hardware as shown in FIG. 9, it takes 1/60 sec to capture an image, and t ₀ hours are required to determine the input of the moving body 100 based on the image. That is, it may be possible to make only one mobile object input determination in two frames and not be able to follow a high-speed mobile object. Therefore,
For example, when the camera body is installed upside down, the raster scanning is performed from the moving body approach direction side as shown in FIG. 7, and only the necessary portion is processed by a window as shown in FIG. As described above, it takes t ₁ hour to capture an image by hardware and t ₂ hours to determine the input of the moving body. That is, the determination can be performed once every 1/60 sec, and the determination can be performed in real time at the video rate.

In this case, the image is upside down, but in the case of recognition of characters and numbers, it can be executed by changing the method of recognition processing in advance. Further, by providing a means for storing the image data in the memory in advance so that the image can be normally displayed even when it is displayed on a monitor or the like, it is possible to always perform normal processing.

〔The invention's effect〕

According to the present invention, although it is difficult to capture an image by a headlight of a moving body such as a vehicle at night, it is possible to perform rear image capturing and further process density data in a small area. Therefore, the image can be obtained at the optimum timing at all times of the day and night. As a result, the extraction of the target area can be performed in a shorter time than the conventional method of forward imaging, which has the effects of reducing the overall processing time and achieving a high recognition rate of the numbers and shapes of moving bodies such as vehicles.

[Brief description of drawings]

FIG. 1 is a diagram for explaining a configuration system of the present invention, FIG. 2 is a diagram showing a hardware configuration according to the present invention, FIGS. 3 and 4 are diagrams showing an entry state of a moving body, and FIG. A flow chart showing the processing of the present invention, FIGS. 6, 7, and 8 are diagrams showing an example of raster scanning for explaining another embodiment of the present invention, and FIGS. 9 and 10 are other diagrams. 2 is a time chart of hardware for explaining the embodiment of FIG. 10 ... ITV camera, 100 ... Moving body, 200 ... Image processing device, 50 ... Specific target area of processing.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Morio Kanazaki 4026 Kuji Town, Hitachi City, Ibaraki Prefecture Hitate Works Ltd., Hitachi Research Laboratory (72) Inventor Mitsuya Orita 4026 Kuji Town, Hitachi City, Ibaraki Prefecture Hitsuru Corporation Hitachi, Ltd. (56) References JP-A-54-66019 (JP, A) JP-A-52-87318 (JP, A) JP-A-55-112585 (JP, A)

Claims (2)

[Claims] 1. An imaging camera installed so that a rear part of a moving body enters the imaging field of view, and a detection area smaller than the imaging field of view provided on the approaching side of the moving body in the imaging field of view. Binary frequency distribution calculating means for obtaining a binary frequency distribution of image data input to the detection area, and
A moving body provided with means for detecting a time point at which the output of the value frequency distribution calculating means exceeds the threshold and thereafter becomes equal to or less than the threshold, and means for taking in image data of the entire imaging visual field when the output of the detecting means is present. Rear detection device. 2. The rear detection device for a moving body according to claim 1, wherein the image pickup camera is installed and installed so that the detection area is located at a position where raster scanning of the image pickup camera starts.