一种新型增加去雾处理的车牌检测算法在林区中的应用

Application of A Novel Algorithm for License Plate Detection with Defogging Processing in Forest Region

林区的治安防护管理一直是人们所欠缺关注的问题,林区作案、林区动植物的盗伐都透露出林区治安防护管理的短板。本文致力于机器视觉在林区的应用,为加强林区的防盗伐能力提出一种融合图像去雾处理的车牌检测算法,该方法采用限定对比度自适应直方图均衡化的技术去除大雾天、雨雪天等图像模糊的干扰,结合车牌的颜色、长宽比形态等特征对雾雨雪天气下的车牌进行检测。该方法相比普通车牌检测方法的效果有较大提升,雾雨雪天气环境下检测车牌的准确率可以达到88.00%,而普通车牌检测方法达到的准确率仅为82.67%。初步证明了本文提出的算法在林区雾雨雪天气下应用的可行性。

The management of public security protection in forest areas has always been a problem that people lack attention to.It reveals the weakness of public security protection management in forest areas that committing crimes,illegal felling of trees and poaching animals.Our group has been committed to the application of machine vision in forest areas.In order to enhance the ability of forest area to prevent illegal cutting,this paper presents a novel algorithm of license plate detection which combines image defogging processing.This method uses the technique of contrast limited adaptive histogram equalization to remove the interference of image blurring in foggy,rainy and snowy days,and combines the features of the license plate such as color,aspect ratio,etc.to detect the license plate in fog,rain and snow weather.Compared with the ordinary license plate detection method,the result of this method has been greatly improved.The accuracy of detecting license plate in fog,rain and snow weather can reach 88.00%,while the accuracy of ordinary license plate detection can only reach 82.67%.This paper preliminarily proves the feasibility of the proposed algorithm in fog,rain and snow weather of forest areas.