针对车道线检测算法计算量大,耗时长的问题,为提高算法的实时性和鲁棒性,把通过CCD获取的道路图像由视频转码模块TVP5158从模拟信号转化为数字信号,在进行工作区域的限定的基础上,进行灰度化,并利用改进的滤波算子进行图像的滤波处理;为利于后续步骤的处理,然后利用对噪声的低敏感性和更有利边缘的显示的Sobel算子进行边缘分割;然后在搭建DSP硬件平台上,结合改进的Hough算法,提高检测与识别的准确性和实时性,将处理后的图像送到LCD显示屏上显示最终检测的结果.%Given that the lane detection algorithm is time-consuming, in order to improve the real time androbustness of lane detection algorithm, the main innovatory work following has been done in this paper. Theroad image obtained by CCD is converted into digital signal from analog signal by video transcoding moduleTVP5158, and the working area is defined. Based on this, the gray level is improved, and the filtering operatoris processed by the improved filtering operator. Then, the Sobel operator with low sensitivity to noise and morefavorable edge display is used to segment the edge, which is beneficial to the subsequent processing. Then, on theDSP hardware platform, combined with the improved Hough algorithm, the accuracy of real-time detection andrecognition improved, the processed image then was sent to the LCD display to display the final detection results.
展开▼
