基于图像分解和多重校正融合的车牌图像去雾算法  

LICENSE PLATE IMAGE DEHAZING ALGORITHM BASED ON IMAGE DECOMPOSITION AND MULTIPLE CORRECTION FUSION

作　　者：王巧月 陈树越[1] 赵双 Wang Qiaoyue;Chen Shuyue;Zhao Shuang(School of Information Science and Engineering,Changzhou University,Changzhou 213164,Jiangsu,China)

机构地区：[1]常州大学信息科学与工程学院,江苏常州213164

出　　处：《计算机应用与软件》2020年第8期202-206,共5页Computer Applications and Software

基　　金：江苏省研究生科研创新基金项目(KYCX19_1770)。

摘　　要：针对大雾天气下,车牌识别困难,智能交通系统工作效率低等问题,提出一种基于图像分解和多重校正融合的车牌图像去雾算法。将雾图分解为纹理层和结构层,通过优化纹理层强化图像轮廓及车牌边缘纹理信息;对结构层采用多重伽马校正,并采用拉普拉斯金字塔融合方法选取最优对比度和饱和度,融合多重校正图像实现去雾,恢复车牌颜色和清晰度;融合处理后的纹理层和结构层,得到纹理信息和清晰度良好的无雾车牌图像。实验表明,该算法主观去雾效果明显且自然,能有效提高车牌定位和识别效果。Aiming at the difficulties in license plate recognition and low efficiency of intelligent transportation system in foggy weather,this paper proposes a license plate image dehazing algorithm based on image decomposition and multiple correction fusion.The hazy image was decomposed into a texture layer and a structure layer,and the texture layer was optimized to enhance the image contour and the license plate edge texture information.The multiple gamma correction was applied to the structure layer,and the Laplacian pyramid fusion method was used to select the optimal contrast and saturation.The multi correction image was fused to realize defogging and recover the color and definition of license plate.Finally,by fusing the processed texture layer and the structure layer,we could get the fog-free license plate image with good texture information and good definition.The experimental results show that the subjective dehazing effect of this algorithm is obvious and natural,which can effectively improve the license plate location and recognition effect.

关 键 词：图像分解 多重校正 图像去雾 车牌定位识别 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]