路面裂缝图像处理算法研究

Pavement Crack Diseases Recognition Based on Image Processing Algorithm

为了避免传统人工视觉裂缝检测方法的耗力、耗时、不精确、影响交通、危险、花费高等缺点,提出了一种新的基于图像处理技术的路面裂缝类病害自动识别算法。识别分为两个步骤:首先以一个5×5的窗口为基准,在这个窗口中确定9种不同的掩膜模板,对有噪音的路面图像进行平滑和增强;然后基于阈值分割理论,采用最大类间、类内距离准则对图像进行阈值分割,提取图像上的裂缝特征。最后对采集的200幅路面裂缝图像进行了平滑和分割试验研究,和Robison等常用的平滑模板相比,对图像进行增强的同时较好地保护了裂缝边缘。在对平滑后的图像进行分割当中,和Hough变换、数学形态学等分割算法进行了对比研究,结果表明了该算法在精度、速度和可靠性方面具有一定的优势。

Conventional visual and manual road crack detection method is labor-consuming, non-precise, dangerous, costly and also it can affect transportation.To avoid these shortcomings, an automatic road crack detection., algorithm based on image process was presented. It includes two steps： firstly, using nine different templates in a 5 × 5 window to smooth and enhance noise pavement image; secondly, based on threshold segmeentafion method, adopting criterion of maximum class distance to segment processed image and extract crack character. In the end, the foregoing smooth and segmentation algorithm on 200 pavement crack images were tested. Comparing with other accustomed templates, the images are enhanced and the crack edge is protected. Further comparing with segmentation methods of Hough transform and mathematic morphology, the presented method shows definite advantage in precision, processing speed and reliability.