基于数学形态学和改进Otsu的单板缺陷检测

在经典检测算子中,针对单板灰度图像的缺陷检测存在缺陷边缘检测不清晰甚至出现伪边缘的问题,提出了一种Otsu改进算法与数学形态学相结合的单板缺陷检测算法.在HSI彩色空间中,对利用数学形态学滤波后的H、S、I分量采用本文算法进行分别处理,并将三分量的检测结果进行叠加.结果表明,该算法能够准确检测出单板的一个或多个缺陷边缘.与经典边缘检测算子的检测结果对比可知,该算法无论对于缺陷的定位还是边缘的提取均优于其他方法.

Defect detection of veneer based on mathematical morphology and improved Otsu algorithm

In the classical detection operators, in order to solve the problem that the detect edges are not clear and the pseudo-edge even forms in the defect detection of gray level images of veneer, a defect detection algorithm for veneer in combination with both improved Otsu algorithm and mathematical morphology was proposed. In the HSI color space, the H, S and I components filtered with mathematical morphology were separately dealt with the proposed algorithm, and the detection results of three components were superimposed. The results show that the proposed algorithm can accurately detect one or more defect edges. Compared with the detection results of classical edge detection operators, it is revealed that the proposed algorithm is better than other methods not only in defect positioning but also in edge extraction.