Sobel算法在车牌识别系统中的改进与实现  

Improvement and Realization of Sobel Algorithm in License Plate Recognition System

作　　者：刘锦[1] 杨顺[1] 余焕明[2] 

机构地区：[1]辽宁工程技术大学电信学院,辽宁葫芦岛125105 [2]河北科技大学信息科学与工程学院,河北石家庄050000

出　　处：《计算机系统应用》2010年第5期209-211,共3页Computer Systems & Applications

摘　　要：经典Sobel算法是图像边缘检测中常用的方法之一,但在车牌识别系统中也检测出了很多的噪声点和造成了车牌字符的断裂,加大了后期车牌提取和字符识别的难度,降低了车牌定位的精确度。针对车牌识别系统,对Sobel算子进行了改进,在实际运行中取得了良好的车牌定位效果和很强的抗干扰能力,并能在最大程度上保持字符的完整性。Classic Sobel algorithm is commonly used in edge detection. However, it has detected a lot of noise points and created broken license plate characters, which increased the difficulty of extracting license plate and characters, recognition, and reduced the accuracy of license plate location. In this paper, Sobel algorithm has been improved for license plate recognition system. The improved Sobel algorithm has achieved good results of license plate location and it has good anti-interference capability. And it can maintain the integrity of the characters to a large extent.

关 键 词：SOBEL算子 边缘检测 边缘跟踪 车牌识别 图像预处理 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]