恶劣天气环境下车牌识别方法研究与仿真

Vehicle License Plate Recognition Method and Simulation in Adverse Weather Conditions

研究恶劣天气下车牌识别的问题。由于图像像素丢失,影响识别的精度。针对传统PCA检测识别车牌的方法,光照条件恶劣的时候,存在图片像素丢失,导致算法无法准确识别车牌的问题。提出了一种运动车牌识别的优化算法。通过将相邻的3帧图像序列进行垂直方向和水平方向投影作差,对车辆图像完整地分割,避免了仅对单个像素的读取,解决了像素丢失的问题,实现对车牌的准确识别。实验表明,改进算法实现了恶劣天气环境下对车辆图像的正确识别,取得满意的效果,为识别研究提供了依据。

Traditional PCA vehicle video detection methods lost image pixels under bad weather conditions,which leaded to the problem of poor identifying accurate of license plates.In order to improve the accuracy of license plate recognition,this paper presented a movement license plate recognition algorithm.By three image sequences adjacent to the vertical and horizontal directions for the poor projection,full segmentations of images of vehicles were carried out to avoid only reading a single pixel,which solved the problem of missing pixels and achieved accurate license plate recognition.Experimental results show that in adverse weather conditions,using this algorithm can complete correct identification and obtain satisfactory results.