基于小波模极大值的磁瓦裂纹缺陷边缘检测算法

Edge Detection Algorithm of Magnetic Tile Crack Based on Wavelet Modulus Maxima

为了准确提取磁瓦表面缺陷的边缘信息,提出一种基于图像加权信息熵和小波模极大值相结合的磁瓦表面裂纹缺陷的边缘检测算法。针对磁瓦表面缺陷对比度低、背景纹理对边缘提取干扰大等特点,设计了一种自适应改变截止频率的BHPF滤波器。利用图像梯度方差加权信息熵对背景纹理的清晰程度和复杂程度进行定量描述,拟合出信息熵同截止频率的非线性函数关系,自适应改变滤波器参数。为避免在多尺度下将缺陷的边缘信息丢失,采用分解尺度判别函数获取小波变换的最优分解尺度。为保证裂纹缺陷边缘连续性和定位准确性,采用双阈值对小波模极大值进行判定求得边界点。实验结果表明,该方法对磁瓦裂纹缺陷边缘的检测优于传统的Canny和Sobel边缘检测算子,可用于磁瓦其他缺陷的提取,为实现缺陷的自动识别奠定了基础。

In order to accurately extract edge information of magnetic tile surface defect, an edge detection algorithm based on image weighted information entropy and wavelet modulus maxima is proposed. Because the magnetic tile surface with low contrast and textured background has a negative influence on edge extraction, a new BHPF filter with adaptive changing cutoff frequency is designed. The clarity and complexity of textured background are quantitatively described by weighted information entropy of image gradient variance. The filter changes its parameter through matching the non-linear relationship between information entropy and cutoff frequency. To prevent the losing of edge information, the best decomposition scale is obtained by the level determination function. In order to ensure the edge continuity and veracity, wavelet modulus maxima is judged through a double threshold to get the edge point. Experimental results show that the algorithm outperforms the conventional canny and sobel algorithms in detection of magnetic tile crack edge. This edge detection algorithm can also detect other defects.