基于Gabor算子的人工蜂群算法大坝裂缝检测

Dam Crack Detection of Artificial Colony Algorithm Based on Gabor Operator

针对水下大坝裂缝图像的低信噪比、低对比度以及大坝裂缝形状的不规则性和方向的不确定性等特点导致的检测精确度低、通用性弱的问题,提出了一种基于Gabor算子的人工蜂群算法大坝裂缝检测方法。该方法简化了蜜蜂的搜索过程,先利用Gabor算子搜索得到一组局部最优花蜜源,根据花蜜源的"收益度"以及"朝向性"特点,从最优花蜜源的角度邻域搜索收益率较高的花蜜源,并对所有的食物源按照条件进行筛选和连接。由于减小了搜索范围,所以可以节省搜索时间,而且针对不同类型的大坝裂缝可以精确连贯地检测出大坝裂缝信息,并且可以有效地抑制噪声。

In this paper, we propose a kind of dam crack detection of artificial colony algorithm based on Gabor operator to solve problems such as low detection precision and application, caused by the low SNR and contrast of the dam crack image, and irregularity and uncertainty of the dam crack shape. This method simplifies the search process of bees. First, the Gabor operator is used to search the edge of image to get a set of local points and its direction. And then, according to the characteristics of its revenue degree and orientation, the food sourde is searched in the optimal food source neighborhood. At last, the image is searched from the local optimal points and all dam cracks information is found. We can save search time due to reducing the search scope and detect the dam crack information consistently according to the different types of dam crack. Moreover, it can suppress noise effectively.