基于双层连通性检测的路面裂缝图像去噪算法  

A Denoising Algorithm for Pavement Cracking Images Based on Bi-layer Connectivity Checking

作　　者：彭博[1] 罗楠欣 蒋阳升[2,3,4] 陈成 K C P WANG 

机构地区：[1]重庆交通大学交通运输学院,重庆400074 [2]西南交通大学交通运输与物流学院,四川成都610031 [3]西南交通大学综合运输四川省重点实验室,四川成都610031 [4]西南交通大学综合交通运输智能化国家地方联合工程实验室,四川成都610031 [5]School of Civil&Environmental Engineering,Oklahoma State University,OK,74078,USA

出　　处：《公路交通科技》2015年第11期7-13,26,共8页Journal of Highway and Transportation Research and Development

基　　金：国家自然科学基金基金项目(51108391,61170041);中央高校基本科研业务费专项资金项目(A0920502051208-99)

摘　　要：为了消除路面裂缝图像中的孤立噪声并保护裂缝边缘的细致特征,本文基于像素和裂缝子块两个层级的连通性检测设计了去噪算法,主要包括3个部分：（1）像素连通性检测,剔除连通性差的像素;（2）孤岛噪声消除算法,删除相对孤立的目标区域;（3）8×8裂缝子块去噪算法。然后,提出了算法流程并探讨参数的选择。最后,应用Visual Studio 2008和Open CV编程对本文算法、常用的中值滤波去噪法、长线段和黑色像素比例法进行性能测试,结果表明,本文算法可有效消除孤立噪声并保持边缘连续性,具有更高的准确率（85.06%）和召回率（85.80%）,综合性能指标F1值比后两者分别高0.74%~19.19%和0.20%~12.06%。In order to remove the isolated noises in pavement cracking images and protect the fine features of cracking edges,a denoising algorithm is designed based on the connectivity checking of pixel level and cracking subimage level,which has 3 main parts：（ 1） pixel connectivity checking,to eliminate pixels with bad connectivity;（ 2） a denoising algorithm to wipe out insular target regions;（ 3） 8 × 8 cracking subimage denoising algorithm. Then,the algorithm procedure is put forward and the parameter selection is discussed.At last,the performance tests of the algorithm and 2 commonly used methods（ median filtering and the algorithm based on long line and black pixel percentage） are conducted using Visual Studio 2008 and Open CV programming. The result indicates that the proposed algorithm can remove isolated noises effectively and keep continuity of cracking edges with higher precision（ 85. 06%） and recall rate（ 85. 80%）,and can achieve better scores of comprehensive performance index F1： 0. 74%- 19. 19% and 0. 20%- 12. 06%higher than those of the other 2 methods respectively.

关 键 词：道路工程 路面裂缝 去噪算法 图像处理 连通性检测 

分 类 号：U416.23[交通运输工程—道路与铁道工程]