欢迎访问《图学学报》
分享到：
摘要：针对轴类零件表面图像的特点，提出一种基于图像处理的表面裂纹检测算法。算
法首先采用空域和小波域混合滤波对图像去噪，提出一种图像分块自适应模糊集增强方法，以
提高裂纹区和背景区的对比度；然后应用Canny 边缘检测算子和数学形态学操作进行图像分割，
提取裂纹区域；最后通过计算裂纹连通域的圆形度和长宽比特征判断零件的表面图像中是否有
裂纹存在，实现裂纹检测。试验表明，该算法即使在信噪比不高的图像中也能实现对目标裂纹
的检测，证明了算法的有效性。
Abstract: According to the characteristic of the image of a shaft component surface, an algorithm to#br# detect surface crack is proposed based on image processing. The algorithm firstly adopts complex spatial#br# and wavelet method to de-noising, comes up with an adaptive fuzzy enhancement of block matrix to#br# improve the contrast of crack and background. Canny edge detection and morphology operation are used#br# for image segmentation. At last, the circularity and length-width ratio are obtained to judge the existence#br# of crack, and crack detection is done eventually. The result of test manifests the accurate detection of#br# crack even if the signal to noise ratio is not very high, proving the effectiveness of algorithm.
