A system and method for capturing images of running a red light under low-illumination robustness
A red-light running and low-illumination technology, applied in the field of red-light capture systems, can solve the problems of missed or wrong shots, robustness, etc., and achieve the effect of preventing missed or wrong detections
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
Embodiment 1
[0040] Embodiment 1, a robust red light running capture system under low illumination.
[0041] The red light running capture system of the present invention includes:
[0042] 1. The camera unit is used to shoot video images of the detected area; in order to meet the shooting definition requirements, the camera unit can use an integrated high-definition network camera with more than 7 million pixels.
[0043] 2. The red light detector is used to count the proportion of red light areas judged to be super red pixels.
[0044] The judgment rule of the red light detector is: count the proportion of super red pixels in the red light area. When the proportion of super red pixels is greater than 90%, it is judged as red light, otherwise it is not red light.
[0045] The judgment rule of the super red pixel is: R>G+50&&R>B+50&&R>150&&G<100&&B<100, wherein R, G, B are the RGB value components of the pixel respectively, && means satisfying the conditions at the same time.
[0046] 3....
Embodiment 2
[0051] Embodiment 2, a kind of robust red light capture method under low illumination, refer to the attached figure 1.
[0052] The snapping method for running a red light of the present invention comprises the following steps:
[0053] 1) Obtain a frame of image SrcImg from the camera unit;
[0054] 2) Calculate the grayscale image Gray_cur of the frame image SrcImg and the data RGBImg of the RGB space;
[0055] 3) Use the current grayscale image and the previous frame grayscale image to make a frame difference to obtain the foreground;
[0056] 4) If the proportion of foreground pixels is greater than a certain threshold (10%), start the license plate detector and the tail light detector in the foreground area at the same time to detect the license plate and tail lights;
[0057] 5) According to the position relationship between the license plate and the vehicle, the range of the license plate is expanded to obtain the target position of the vehicle;
[0058] 6) Track th...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 
