Vehicle Detection Based on Adaptive Background

In order to improve the accuracy of vehicle detection, and to solve the gradually changing brightness of light in the background and the movement of the objects in the background, this paper presents an algorithm that fast adapts to the background generation and updating. Focus on the objects with similar gray background, and moving object missing caused easily by segmentation, the threshold segmentation and edge sharpening--the combination of methods is used here, to strengthen the edge of moving objects. The experimental result shows that: the algorithm can adapt to changes in the background, compensate the missing moving objects after segmentation, which has high robustness and excellent real-time performance.