基于形态学处理与特征分析的混凝土裂缝检测研究

Research on concrete crack detection based on morphological processing an d feature analysis

混凝土裂缝的出现和发展将严重影响结构的安全性与适用性,因此对混凝土裂缝的有效检测尤为重要。传统的裂缝检测算法存在抗干扰能力弱、裂缝提取不完整等问题。提出一种结合形态学操作与对比度增强的预处理算法,以及进行阈值分割后基于裂缝几何特征的滤波算法,可以有效地实现混凝土裂缝的自动检测。预处理算法可以去除低对比度以及光照不均匀的影响,并且可以平滑背景噪声以及增强裂缝特征;基于裂缝几何特征的滤波算法可以利用面积特征、线性特征、凸性特征实现对裂缝区域的提取。试验表明,该算法可以较为精确的检测并提取出混凝土裂缝特征。

The appearance and development of concrete cracks will seriously affect the safety and applicability of the structure,therefore,the effective detection of concrete cracks is particularly important.The traditional crack detection algorithm has the problems of weak anti-interference ability and incomplete crack extraction.This paper proposes a preprocessing algorithm that combines morphological operations with contrast enhancement,and a filtering algorithm based on geometrical features of cracks,which can effectively achieve automatic detection of concrete cracks.The preprocessing algorithm can remove the effects of low contrast and uneven illumination,and can smooth background noise and enhance crack characteristics.The filtering algorithm based on the geometrical features of fractures can use the area features,linear features,and convex features to extract the fracture areas.The experiment result shows that the algorithm can detect and extract concrete crack features more accurately.