公路路面裂缝类病害图像处理算法研究

文章在对公路路面裂缝类病害的检测过程中,通过认真分析路面病害图像特征,研究了适合于路面病害识别的图像处理算法。由于裂缝类病害的类型包括横向裂缝、纵向裂缝和不规则裂缝,边缘可能在各个角度方向存在梯度,因此,文中构造了8个方向的模板对图像进行Sobel边缘检测。边缘检测处理后,结合加权的邻域平均噪声滤除算法和Ostu图像分割算法对病害图像进行处理,处理结果相对于其他经典算法,裂缝边缘宽度较细(2个像素),并且裂缝的边缘保护很好,裂缝边缘之间断续情况较少。该算法用于对公路路面裂缝类病害的识别检测过程中,检测精度和检测效果都很好。

Image Processing Algorithm for Detecting the Pavement Crack Diseases

A new image processing algorithm is applied for detecting and analyzing the characters of pavement crack diseases.The types of the pavement crack diseases include transverse cracks,longitudinal cracks and unregulated cracks,the image of the crack edges may be not clear in different direction for the existence of edge gradient.In this paper a new Sobel algorithm is structured in eight directions to detect the crack edges.The diseases image is processed by the noise filtrating algorithm of neighborhood weighted averaging and Ostu image fractionating algorithm.The results show that the width of crack edge treated by the new algorithm of this paper is less than2picture elements.The edges of cracks are properly protected and the images of cracks are clear and continuous.The detecting precision and effectiveness of the new image processing algorithm are very good for detecting the pavement crack diseases.