Method, system and equipment for monitoring vehicles based on vehicle taillight detection

Abstract

The invention discloses a method, a system and equipment for monitoring vehicles based on vehicle taillight detection. The method for monitoring the vehicles comprises the following steps of: sufficiently utilizing the color information and the motion information of vehicle lights; pairing tail lights by adopting a space symmetry strategy; identifying a single target vehicle; and introducing a position prediction mechanism for preventing the shielding condition of the tail lights of the vehicles so as to effectively prevent targets from being lost. The invention solves the problems of high time consumption and low accuracy when a morphological operator is computed in the traditional algorithm, prevents the complex computation, has higher processing speed, can meet the real-time requirements, greatly decrease the iterations of the algorithm, enhance the tracking efficiency and be convenient to apply to a real-time tracking scene.

Description

technical field [0001] The invention relates to the field of communication technology, in particular to a vehicle monitoring method, system and equipment based on vehicle tail light detection. Background technique [0002] Traffic flow detection is an important part of the intelligent transportation system. The traditional traffic flow detection method uses the method of buried induction coils. Due to the disadvantages of complex construction, difficult maintenance, and damage to the road surface, it is gradually being replaced by video-based detection methods. substitute. [0003] In the video detection of traffic flow, the detection of vehicles at night has always been a difficult problem, mainly because the road light conditions at night are complex, and it is difficult to accurately obtain images of vehicles with ordinary CCD cameras. [0004] Some developed countries adopt the method of installing infrared cameras or combining CCD cameras and infrared cameras to obtain...

AI Technical Summary

Problems solved by technology

[0002] Traffic flow detection is an important part of the intelligent transportation system. The traditional traffic flow detection method uses the method of buried induction coils. Due to the disadvantages of complex construction, difficult maintenance, and damage to the road surface, it is gradually being replaced by video-based detection methods. replace

[0003] In the video detection of traffic flow, the detection of vehicles at night has always been a difficult problem, mainly because the road light conditions at night are complex, and it is difficult to accurately obtain images of vehicles with ordinary CCD cameras

[0004] Some developed countries adopt the method of installing infrared cameras or combining CCD cameras and infrared cameras to obtain night road images. This method is very effective in detecting pedestrians at night, but it will still be affected by headlights when detecting vehicles driving at night. Influenced by light, reflected light from the ground and ambient light, and infrared cameras are expensive

[0005] At present, the common method is still to use ordinary CCD cameras to take nighttime images, and detect vehicles by detecting the headlights in the images. Large bright spots, at this time it is impossible to accurately detect the vehicle, and the aperture of the camera must be manually adjusted. Obviously, this method cannot meet the needs of the development of modern traffic intelligence; and in urban traffic, this method is affected by street lights and roadsides. The impact of building lights is also great, especially in rainy days, when there are many reflections of street lights and building lights on the road, the situation is even more serious, but the calculation of morphological operators is time-consuming and cannot reach the real-time detection of traffic flow. In addition, due to the small imaging area of ​​car lights in the image, area calculation is often inaccurate, and morphological operators, such as body-to-body ratio, are more sensitive to area. For small objects such as car lights, the morphological operator's The calculation accuracy is not high, resulting in a decrease in detection accuracy

Images