Method for detecting traffic flow of planar intersection

Abstract

The invention relates to a method for detecting the traffic flow of a planar intersection, comprising the following steps: firstly, extracting the edge image of a real-time video frame image at the planar intersection; secondly, carrying out background modeling to obtain a background edge image; thirdly, subtracting the background edge image from the edge image of the current image to obtain the edge image of a moving vehicle; fourthly, solving the average value of all the pixel values in the edge image of the moving vehicle; fifthly, carrying out vehicle existence detection on the average value by adopting a double-threshold method and removing the noise interference; and sixthly, counting the traffic flow. In the invention, the background modeling is carried out by using the edge images of moving vehicles which pass through the intersection when the light turns green and the backgrounds are updated in real time by using an automatic updating algorithm of the background edges so that the accurate background edges of the moving vehicles are obtained and the shadows and the noise interferences of the background images can be effectively removed.

Description

technical field [0001] The invention relates to a technology for detecting video traffic data at road level intersections, in particular to a technology for detecting traffic flow at road level intersections, and belongs to the technical field of traffic video detection. Background technique [0002] In the field of traffic video detection technology, commonly used vehicle detection methods mainly include: background difference method and adjacent frame difference method. The background difference method needs to be based on a reliable background image, and when the ambient light changes suddenly, the situation of missed detection and false detection is more serious. Furthermore, vehicles at level intersections in road traffic will also stop and go with the change of traffic lights, which will cause vehicles parked in the detection area to be easily mistaken as Part of the background is processed, thus adversely affecting the vehicle detection results. Adjacent frame diffe...

AI Technical Summary

Problems solved by technology

Furthermore, vehicles at level intersections in road traffic will also stop and go with the change of traffic lights, which will lead to easy mistaking of vehicles parked in the detection area when modeling the background of level intersections. Part of the background is processed to adversely affect vehicle detection results

The adjacent frame difference method is to detect the moving vehicle by subtracting the corresponding pixels of two consecutive frames of images, but this adjacent frame difference method is not effective for the detection of vehicles that are too slow or too fast.

In addition, when using the adjacent frame difference method to detect vehicles, shadows and noise on the picture will also greatly affect the detection effect, resulting in false detections

Images