KR100284596B1 - How to measure waiting length at intersection - Google Patents

Abstract

In the traffic measurement method, the video signal of the road situation detected by the CCD camera or CCTV camera is processed as a binary video signal to increase the reliability in measuring the waiting length at the intersection. The image input means maintains and installs an image input in a state where a vehicle does not exist in a predetermined state, and then sets the image into a memory area as an initial image, and then boots as the image input means while the initial image is set. After pre-processing the input video signal, the noise component is removed, the difference image is extracted by comparing the set initial image and the current image, the threshold value is set for the extracted difference image, and the binary image is extracted and extracted. The gray level and the pixel of the extracted binary image are detected and the waiting length is extracted through the calculation.

Description

How to measure waiting length at intersection

The present invention relates to a traffic volume measuring method, and more particularly, to process a video signal of a road situation detected by a CCD camera or a CCTV camera as a binary video signal to increase the reliability in measuring the waiting length at an intersection. It relates to a measuring method.

In general, as image extraction and information processing technology develop, control of remote production facilities, combined discovery of printed circuit boards (PCBs), air navigation in contaminated areas, air navigation in oilfields, and remote sensing technology Imaging technology is widely used for surveillance.

Recently, the traffic field collects and processes information about traffic flow and vehicle speed by using video input media such as CCTV cameras, and provides various types of traffic information. The scope of application is gradually being extended to overcome the limitations of the collection of information through loops and the information using infrared rays.

The collection of traffic information is a very important factor for effectively distributing traffic volume to traffic control, efficient road management and traffic planning, and rapidly changing traffic demand.

Therefore, in order to accurately grasp the traffic conditions in real time in response to the real-time signal control system, ILD embeds a loop coil on the road surface and recognizes the traffic volume and the vehicle type by detecting the magnetic field change generated by the movement of the vehicle. (inductive loop detector) Detection is the most commonly used.

In the case of the above-mentioned ILD detector, the reliability of measuring the point information is high, but the maintenance and the new installation are frequently made according to the maintenance of the road, and the unnecessary expense is frequently generated, There was a problem that can not grasp the overall flow of the road.

Therefore, in recent years, in order to minimize the problem of collecting traffic information from the ILD, an image detection apparatus capable of processing wide information and detecting multiple lanes has been used.

When measuring the information (speed, occupancy time, vehicle type) between points by using the image detection device described above, set a constant two points "x1", "x2" and the distance "d" between two points, The vehicle speed is detected by detecting information of time "t" occupying the two points, and the vehicle length is calculated by calculating the length of the vehicle from the speed of the vehicle.

However, in this case, it is not possible to accurately measure a certain distance between the two points, the measurement of the speed is inaccurate, and accordingly there is a problem that a large error in the classification of vehicles.

Therefore, in recent years, tracking is not performed to extract information of a certain point by tracking a vehicle, but to break a unit of time into an area of a set tracking zone and measure a moving distance of the vehicle during the time to calculate a speed. The method is used.

In addition, in order to measure the spatial information (waiting length) through the image detection device, the edge extraction method is commonly used, which is to extract the amount of vehicle load through the edge extraction and calculate the length based on this. However, there is a problem that the reliability of extracting the waiting length is deteriorated, and the fundamental problem of the image detection device is occlusion and the problem that the vehicle immediately behind the vehicle being tracked cannot be measured and various measurement errors are detected. There was a problem with derivation.

The present invention has been made in view of the above-described general problems, and its object is to compare the currently input image with a preset detection area in the measurement of the waiting length through image detection, and to detect the threshold value of the difference image as a binary image. It is designed to provide reliability for the measurement of waiting length, and to provide reliability for measuring traffic volume at intersections and changing signal systems.

1 is a block diagram showing a configuration of an apparatus for measuring air length according to an embodiment of the present invention;

2 is a flowchart of an embodiment for realizing a method for measuring air length in the present invention.

3 is a flowchart of an embodiment of extracting an waiting length from a binary image extracted in the present invention.

The present invention for achieving the above object is a method for measuring the traffic volume, the image input means is provided to maintain a predetermined angle at a predetermined position of the road is a predetermined image to be input in the state that the vehicle does not exist Processing to a state and then setting the memory area as an initial image;

Preprocessing an image signal input from the image input means in a state where an initial image is set, removing noise components, and extracting a difference image by comparing the set initial image with a current image;

Setting a threshold value for the extracted difference image and extracting a binary image;

And detecting the gray level and the pixel of the extracted binary image, and then extracting the waiting length through an operation.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

As can be seen in FIG. 1, the apparatus for measuring air length according to the present invention includes an image input unit 10, an image signal processing unit 20, an information analyzer 30, and a display unit 40. 10) is a basic element for collecting image data, which is composed of CCD carrera or CCTV camera, and is installed at a predetermined position on the road keeping the predetermined angle to acquire the image at a predetermined height. Process it as an electrical signal and output it.

The video signal processing unit 20 converts an analog video signal applied from the video input unit 10 into a digital signal and outputs the digital signal.

The image signal processor 20 has a resolution of at least 320 × 240 pixels and supports 256 gray levels per pixel and approximately 30 frames per second.

The information analyzer 30 preprocesses an image signal applied as a digital signal from the image signal processor 20, and then detects a threshold value by comparing a value extracted through preprocessing with a set value, and an image of a threshold value. Binarize to perform the wait length measurement.

The display unit 40 is provided with a picture in picture (PIP) function, and displays information on a state of the currently measured waiting length and image information of a currently input road.

The display unit 40 uses a 3 × 3 matrix, and arranges the nine pixels in the region in the order of increasing density, so that the median filter having the center, that is, the concentration of the fifth pixel, as the density value at the center position. filter) was used.

Referring to FIG. 2, the operation for measuring the waiting length in the present invention having the function as described above is as follows.

When the video information about the situation of the road is input through the video input unit 10, the video signal processing unit 20 converts the input video information of the analog state into a digital signal and then applies it to the information analyzer 30 (step) 101, the information analyzing unit 30 determines whether the information on the initial image of the road is set in its memory area (step 102).

If the information on the initial image is not set in the memory area, it is determined that the currently input information is to set the initial image for the measurement of the waiting length, that is, the input image, that is, the image of only the road without a vehicle. The noise contained in the information inputted through the filter is removed through a predetermined filter (step 103), and then set as an initial image (step 104).

If the detected image is in the state for updating the initial image due to the change of the gray level value of the surroundings with time, the background gray level between the currently detected image is compared with the initial image previously set in the memory area. The smallest error value is detected and summed with the initial image stored in the memory area to update the initial image.

The image to be updated may be a "+" value or may be a "-" value (negative).

When the initial image is detected in the memory area of the information analyzer 30 as described above and the input of the real image information that the vehicle is present on the road is detected, the information analyzer 30 is input through the set program or filtering means. After removing the noise component included in the image information (step 105), the difference image is detected by comparing the gray level value of the image information currently detected with the initial image set in the memory area (step 106).

The gray level detection above I ₁ [i] [j] = abs (I ₂ [i] [j]-I ₃ [i] [j]) The absolute value of each pixel is obtained from the equation Therefore, even if the vehicle has a lower gray level than the road, the gray level of the vehicle can be obtained.

Subsequently, a threshold of a dark portion and a threshold of a bright portion are set in the difference image detected through the gray level comparison (step 107), and then a binary image is extracted (step 108).

In the above, the threshold of the dark part is to remove noise appearing in the road part in the difference between the two images at day and night, and the threshold of the bright part is the day when the sunlight is strong and strong light is reflected through the glass, making the image strange. This is to prevent the light coming out and to remove the strong light from the headlights at night, and to extract the binary image, the filtering is performed on the image detected through the following algorithm.

for (int j = Pix + 2; j <= Pix1; j ++)

{DDV [j] = CGlobal.ImageData [row] [j];

if ((DDV [j-1]> = thre && DDV [j-1] <= thre1) && (DDV [j]> thre

&& DDV [j] <= thre1) && (DDV [j-2]> = thre && DDV [j-2] <= thre1))

return TURE

As a result of experimenting on the binary image detected using the above algorithm, the threshold value of the dark part is "8" for the day and night, and the threshold value of the bright part is "150" for the day and night. Since "75" is shown as the most appropriate value, the gray level value of the vehicle is greater than "8" in the daytime and less than "150" in the daytime, and greater than "8" in the nighttime and less than "75". It was found to have a value.

When the extraction of the binary image is completed with respect to the image detected through the above process, the distance to the last part where the waiting length exists is counted by detecting the detection area divided in blocks by the step (step 109). .

Referring to FIG. 3, an operation for measuring the waiting length from the extracted binary image will be described below.

When extraction of the binary image is completed (step 301), the gray level value [i] for each set detection area is detected (step 302), and it is determined whether the value of the gray level [i] of the last detection area is "225" ( Step 303).

If the value of the gray level [i] of the last detection area is "225", the counter is increased by "1" (step 308), and the step 302 is executed to detect the gray level [i] of each detection area. If the gray level [i] of the last detection area is not "225", the pixel [j] in each detection area is detected and the pixel [j] is detected from the detection area where the vehicle does not exist to the detection area where the vehicle exists. It is determined whether has a value of "225" (step 305).

If the pixel [j] has a value of "225" from the detection area where the vehicle does not exist to the detection area where the vehicle is present, the counter is decremented by "1" (step 309), and the step 304 is executed. If the pixel [j] of each detection area is detected and it is determined that the pixel [j] to the first type detection area in which the vehicle exists does not have a value of "225", the value of the gray level [i] and the pixel detected above are detected. The waiting length is detected through the operation of [j] (step 306) (step 307).

In the above, the value of the pixel [j] for each detection region is determined by the value of the gray level [i]. The pixel [j] of each detection region is continuously reduced to the line where the vehicle exists, and thus the final pixel [j] ], The waiting length is extracted by extracting the distance and pixels of the detection area.

As described above, the present invention sets the detection area in advance, processes the detected image as a binary image, compares the binary processed image with the initial image, and then extracts the air length. Is provided.

Claims (5)

In the traffic measuring method, an image input means installed at a predetermined position on a road by processing an image input in a state where a vehicle does not exist in a predetermined state and then setting it in a memory area as an initial image. Process of doing; Preprocessing an image signal input from the image input means in a state where an initial image is set, removing noise components, and extracting a difference image by comparing the set initial image with a current image; Setting a threshold value for the extracted difference image and extracting a binary image; And detecting the gray level and the pixel of the extracted binary image, and extracting the air length through a calculation. The method of claim 1, wherein the initial image set in the memory area is updated according to a change in the gray level value of the vehicle and the surrounding environment over time. The method of claim 1, wherein the threshold of the dark portion is to remove the noise appearing in the road portion in the car image of the two images at night night, the threshold of the bright portion is reflected in the strong light through the glass in the day Method for measuring the atmospheric length at the intersection, characterized in that to prevent the image from coming out strange, and to remove the strong light due to the headlights at night. The method of claim 1, wherein the extraction of the waiting length comprises: increasing a counter by measuring a gray level [i] of a detection area set for a detected binary image; Detecting a pixel [j] of each detection area set for the binary image to execute a subtraction counter; If the extraction of the position where the vehicle exists in the binary image is completed through the above process, calculating the values of the gray level [i] and the pixel [j] to extract the waiting length at the intersection, characterized in that How to measure the waiting length of The method of claim 1, wherein the updating of the initial image comprises detecting a value having a minimum error in the comparison of the background gray level between the preset initial image and the currently detected image, and adding it to the preset initial image. How to measure the air length.