智能车追寻信标灯过程的图像处理

针对自动化系统中的信标定位技术，提出了一种基于图像进行信标定位的方法，并将其应用于智能车竞赛中.综合考虑真实环境的情况，利用中值滤波算法、动态求阈值算法、交换帧丢弃算法、改进型连通域算法和改进型畸变还原算法等多种算法处理线阵CCD和CMOS摄像头采集的图像，研究该方法的误判率.经实验测试和竞赛中的应用检验，该方法能实现智能车在高速行驶时对信标灯目标的追寻，在强光(约200 lx)、弱光(约75 lx)环境下测试1 000次的误判率分别为1.4%和3.2%.研究结果对解决准确定位信标问题具有重要意义.

Image Processing during the Pursuit of Beacon Lamp by Smart Cars

To improve the beacon positioning technology in automation system, an image-based beacon positioning method is proposed and applied to smart car competition. The real environment is considered and the median filtering algorithm, the dynamic threshold algorithm, the exchange frame discarding algorithm, the improved connected area algorithm and the improved distortion reduction algorithm are used to process images captured by CCD and CMOS cameras. The error rate of this method is analysed. The results of the experiment and the application of the method in competition show that this method can achieve the goal of searching beacon lights for smart cars at high speed. The error rates of 1 000 tests under strong light(about 200 lx) and weak light(about 75 lx)environments are 1.4% and 3.2% respectively. The research results are of great significance for solving the problem of accurate beacon positioning.