Similar-physical-property-model-based intelligent intersection peccancy detection method and system thereof

Abstract

The invention relates to a similar-physical-property-model-based intelligent intersection peccancy detection method and a system thereof. The method that is applied to the video-based intersection pedestrian red-light-running event detection field in urban traffic is characterized in that a similar physical property model is used for detecting a pedestrian rapidly and a detection accuracy is improved by using a track filtering method; and with combination of signal lamp state detection, determination of the pedestrian red-light-running event is completed. According to the method, an image of a pedestrian crossing a street at an intersection is collected, the coordination image of the pedestrian is calibrated manually, and training is carried out at a given data set to obtain a pedestrian similar physical property model; the pedestrian in the video image is detected rapidly by using the similar physical property model and a pedestrian movement locus is established for the detected pedestrian according to a relation of a distance and an angle; and on the basis of combination of the signal lamp state and the pedestrian locus feature, a pedestrian red-light-running event is determined and image capturing and voice prompting are carried out.

Description

technical field [0001] The invention relates to the fields of intelligent traffic and pattern recognition, in particular to a technical field of detecting pedestrians running red lights. Background technique [0002] At present, the incidents of pedestrians running red lights at intersections in major cities are relatively serious, which brings great hidden dangers to traffic safety. Pedestrian detection based on video processing is a key technology in intelligent transportation systems. Compared with traditional detection methods such as gratings, video detection has low installation cost and good scalability. In practical applications, due to the complexity of the scene, the change of perspective and scale, the diversity of human body poses and clothing, and partial occlusion, it is difficult to detect pedestrians. [0003] The traditional method of searching for objects in an image is called the sliding window method, that is, continuously traversing exhaustively differ...

AI Technical Summary

Problems solved by technology

In practical applications, due to factors such as the complexity of the scene, the change of perspective and scale, the diversity of human body posture and clothing, and partial occlusion, pedestrian detection is more difficult.

[0003]The traditional method of searching for objects in an image is called the sliding window method, that is, continuously traversing exhaustively different areas on the image, and then judging the specific content of this area. method is very inefficient

Various existing video-based detection methods for pedestrians running red lights still have some shortcomings in terms of real-time and accuracy, so they cannot meet the detection requirements of pedestrians running red lights at intersections in real traffic scenarios.

Images