Detection Method of Illegal Parking Vehicles based on Computer Vision

摘要

Aimed at the shortcomings of the manual detection of illegal parking vehicles, a new detection method was proposed to detect the illegal parking vehicles based on computer vision from real-time monitoring video sequence, which saved human and material resources. In this method, firstly, background subtraction based on adaptive threshold and binarization were used to eliminate background interference after image graying; secondly, morphology processing operations including improved erosion operation and improved dilation operation were used to extract the objects; finally, illegal parting detection was used to draw the outer rectangular frame of illegal parting cars through illegal parking judgment conditions and the camera would be controlled to capture the vehicle image which could be used to recognize the vehicle license plate. Experimental results show that this method can detect the illegal vehicles successfully and has the identification that can distinguish vehicles from non-vehicles, illegal parking from passing vehicles, and can mark and count the illegal vehicles. This method can be used for illegal parking monitoring in residential car park area to effectively control the phenomenon of illegal parking in order to reduce the waste of human and material resources and remedy the transport environment