基于图像处理的路面裂缝特征提取方法

Pavement Crack Feature Extraction Based on Image Processing

摘要:

裂缝是一种常见的路面损坏形式,及时、准确地检测和评估路面裂缝的状态,对于路况监测和养护决策具有重要的意义。文章提出了一种基于图像处理的路面裂缝特征提取方法,首先,对分割得到的路面二值图像进行连通域分析,利用裂缝区域和干扰区域之间的特征差异,筛选出真正代表裂缝的连通域；然后,采用改进的快速并行细化算法提取出裂缝骨架,以端点所在分支的长度为标准来去除骨架中的毛刺,针对骨架断裂和向内缩短的问题,通过端点方向以及端点间距离来进行骨架的连接和生长。实验结果表明,该方法能够有效去除分割结果中的干扰区域,提取出清晰完整且无毛刺的裂缝单一像素骨架,较好地反映了裂缝的主要结构和形态特征。

Abstract:

Cracks are a common type of pavement damage. Thus, the accurate and timely detection and evaluation of the status of pavement cracks are crucial for road condition monitoring and maintenance decision-making. This study proposes a pavement crack feature extraction method based on image processing. First, we performed a connected-domain analysis on the segmented pavement binary image, whereby the connected domains representing cracks were selected based on the feature differences between the crack and interference regions. Subsequently, we applied an improved fast parallel-thinning algorithm to extract the crack skeleton, and the branch length of the endpoint was used as the standard to remove burrs in the skeleton. To address the problems of skeletal fracture and inward shortening, we used the direction of the endpoints and the distance between them to connect and grow the skeleton. The experimental results show that this method can effectively remove interfering regions from the segmentation results, thereby extracting a clear and complete single-pixel skeleton of cracks without burrs to better reflect the main structural and morphological characteristics of the crack.