被动测距系统汽车目标提取算法研究

研究了以高速公路为背景的被动测距系统中汽车目标提取算法.该算法在采集图片中检测路面边缘，根据道路区域在图片中呈三角形的特点，提出一种新的边缘检测模板.此模板突出图片中倾斜边缘的特点，抑制垂直和水平边缘.边缘提取后利用Hough变换检测车道边缘并判断测距系统所在的车道，缩小检测区域，并在该车道内分割提取出目标汽车.解决了同一车道内存在两辆或更多车辆时，区分目标汽车与无关汽车的问题.

Research of algorithm for car detection in passive ranging system

An algorithm for detecting moving car in a passive ranging system for highway was discussed. The algorithm detected the edges of road. Based on the characteristic shape of the road areas as triangles, a new edge detection template was proposed. The template emphasized the incline characteristic of inclined edges, and restrained horizontal and vertical edges. After the edges were picked-up, Hough transform was used to determine which driveway the ranging system was in. The detected area was therefore reduced, and the object car was extracted and picked-up from the driveway areas. The algorithm can distinguish between the object car and irrespective cars when there are several cars in one driveway.