基于图像识别的汽车遇障报警系统研究

A Study on Zmage Recognition Aotomobiles Obstacles Alarm Sytem

针对目前公路交通事故频发的现状,从图像识别领域研究汽车遇障报警系统,辅助汽车安全驾驶。由CCD摄像机获取图像序列,采用直方图均衡化、迭代阈值法等一系列图像预处理方法,增强图像的对比度,提取出较好的图像特征,提出了障碍物识别框的概念,很大程度地实施了系统的实时性,将激光测距技术与图像识别相结合,产生报警条件。

Aimed at the present situation in which traffic accidents frequently happened on the road,research on alarm system for automobiles running into obstacles based on the image recognition is carried out,which can assist drivers in driving automobiles safely.Image sequence is obtained by CCD camera through such a series of preprocessing approaches as histogram equalization and Iterative threshold method,which can enhance contrast of the image and extract the comparatively good characteristics of the picture.Meanwhile,the concept of obstacles’recognition frame is put forward,the real-time character of the system is guaranteed to a large degree and then the warning conditions are produced by combining the laser ranging technology and the image recognition.