铸件砂眼缺陷提取的图像处理算法

Image Processing Algorithm for Sand-hole Defects Extraction in Casting Parts

针对铸件砂眼缺陷的检测问题,提出一种图像形态学重构后的差分算法。Matlab读取图像,中值滤波对图像去噪减少干扰信息,再对去噪后的缺陷图像和合格图像分别进行增强和形态学处理;随后采用差分算法求出两幅图像之间的差异信息,得出图像中灰度值最大的区域即为砂眼所在的位置;最后求出砂眼所占区域的面积、周长、圆形度,累积砂眼特征库。

An image morphological reconstruction differential algorithm was proposed to detect the defects in the cast sand-hole.The image was readed by Matlab,median filtering can reduce the interference information to the image de-noising,and then enhances and morphologically processes the defect image and qualified image after de-noising respectively.Then the difference information between the two images was obtained by the difference algorithm.Finally,the area,circumference and roundness of the area occupied by the sand hole were calculated,and the reservoir of sand hole characteristics was accumulated.