基于边缘检测的车牌识别算法的实现

摘要

针对传统车牌识别的不足,本文提出了基于边缘检测的车牌识别的算法.该算法首先对摄像头获取的车牌图像预处理,去除图像无用信息,然后运用Robert算子检测车牌边缘,并对车牌区域进行图像较正,用高斯滤波法去除噪声并且提取车牌信息特征,接着对车牌区域水平和竖直方向运用触点定位法分割字符,对车牌分割后与相应字符模版匹配,利用预测模型预测识别结果,最后识别出车牌字符.%We proposed the algorithm of license plate recognition based on edge detection in this paper. Firstly,the license image is preprocessed which is captured from camera. Secondly, we use the Robert operator to detect the license edge, correct the area of license image, clear the noise with the Gaussian filter and extract the license plate information characteristics. Thirdly, Using the contacting location method to locate the characters and matching with the character template. At last,the license plate can be recognized as the matched character template.