Vehicle detection method based on textural features

Abstract

A vehicle detection method based on textural features comprises the steps of reading a current frame according to a shot vehicle video stream, performing gray scale and automatic binarization, searching binarized contours, learning an appropriate threshold value according to the number of the contours, updating the threshold value in real time, judging that a vehicle exists if the number of the contours of the current frame is smaller than the threshold value, and otherwise, judging that no vehicle exists. The method for determining and updating the threshold value in real time by the learningsystem can cope with the influence of different parking lots and environmental changes, and is relatively high in adaptability; in addition, the texture contour method is high in processing speed andcan meet the requirement for real-time video processing.

Description

【Technical field】

[0001] The invention relates to detection and determination technology for the presence or absence of moving target items, in particular to a vehicle detection method based on texture features. 【Background technique】

[0002] With the vigorous development of China's economic construction, the urban population and motor vehicle ownership are also increasing rapidly, the traffic flow is increasing day by day, traffic congestion is becoming more and more serious, and traffic accidents occur from time to time. The traffic problem has become a major social problem in urban management, hindering and restricting the development of urban economic construction. Therefore, in-depth research on solving urban traffic problems has extremely important practical significance. To solve urban traffic problems, it is necessary to accurately grasp vehicle traffic information. At present, the common domestic traffic management and monitoring methods include manual monitoring...