一种复杂车辆图像中的多车牌定位方法

针对复杂背景中多个车牌的定位问题,提出一种新的定位方法.该方法综合利用边缘检测、连通域分析、倾斜矫正等多种方法,解决了复杂背景中定位难的问题.能够准确定位杂乱背景中的车牌,对天气、光照变化、车牌在图像中的移动和旋转等,具有良好的适应能力.该方法为后续的字符分割和字符识别提供旋转角度、字符区域定位信息.

An Approach to Locate Multi-license Plates in Complex Vehicle Images

A new method is presented in this paper for locating license plates in complex background. Edge detection. Component connectivity analysis and slant correction etc manifold methods are adopted to resolve the problem of locating license plate in complex background. The capability of accurately location license plate in complex background of this method is proved by experiments. Result also shows that the method is insensitive to weather, illumination variation, motion and rotation of the license plate, and the rotation corner, character orientation etc informations are provided for the after work.