摘要
针对汽车尺寸测量的应用背景,根据双目视觉原理开发了一套基于计算机视觉的汽车整车尺寸测量系统。论述了该双目视觉测量原理及传感器模型,利用图像差值法实现车身目标检测,并在此基础上利用Sobe l算法、阈值化等技术实现特征点提取。结合特征匹配与区域匹配技术实现立体匹配,并完成自动测量。试验结果表明,该系统可以实现汽车整车尺寸的自动、快速、准确测量。
Considering the background of the testing system of vehicle-body dimensions, the paper developed a testing system of vehlcle-body dimensions based on computer vision according to binocular stereo vision theory. The testing principle and sensor model are expounded, and the point is also focused on the techniques of vehicle-body detection of difference and feature extraction using technique of improved Sobel algorithm,threshold segmentation, etc. Stereo matching is realized with the combination of feature matching and area matching. The testing results show that the testing system achieves automatic, fast and accurate testing of vehicle dimensions.
出处
《交通与计算机》
2006年第3期22-26,共5页
Computer and Communications
基金
教育部博士点基金项目资助(批准号:20030183032)
