Abstract: In order to solve the inaccuracy problem of asphalt pavement crack image edge extraction, based on multi-scale morphology, asphalt pavement crack detection algorithm is proposed accoording to characteristics of crack shape irregularity and weak signal. The proposed algorithm employs operators of noise immunity edge detection with multiple structure elements, and according to the variable probability of multiple structure elements fill up edges of crack, an algorithm of adaptive weight is brought forward, it has not only a good ability to smoothing filter, but also a strong ability to extract the multiple image edge. Compared with the traditional edge detection algorithm, the proposed algorithm has restrained noise jamming effectively while detecting the all kinds of asphalt pavement crack accurately, theoretical analysis and experimental results show that the algorithm is efficient and feasible.

Key words: mathematics morphology, multiple structure elements, edge detection, adaptive weight

摘要： 针对沥青路面裂缝图像边缘提取不精确的问题，基于形态学多尺度的思想，结合沥青路面裂缝图像中裂缝形状不规则，裂缝信息比较弱小的特点，提出了一种沥青路面裂缝图像检测算法。算法构造了多结构元素的抗噪型边缘检测算子，且依据不同形状的结构元素对裂缝边缘填充的几率不同，设定了一种自适应确定权重的算法，使得此算子在平滑滤波的同时又检测到各种类型的边缘。与传统裂缝图像检测算法相比，该算法有效抑制了各种噪声的影响，实现了多种类型边缘的精确提取；理论分析和实验结果表明该算法是可行且有效的。

关键词: 数学形态学, 多结构元素, 边缘检测, 自适应权重