改进高通滤波的圆柱滚子缺陷图像非线性反锐化掩模

摘要

针对现有反锐化掩模方法在滚子缺陷区域增强中存在的问题,通过在非线性反锐化掩模方法中引入改进的高通滤波模板,构建了一种新的反锐化掩模方法.以滚子周面缺陷图像为例,对线性、非线性反锐化掩模以及本文方法进行了图像增强及缺陷区域提取效果对比试验,并以分布分离指标DSM对图像增强效果进行量化评价.试验结果表明:使用基于改进高通滤波模板的非线性反锐化掩模方法可以有效改善缺陷区域的增强效果,与线性及非线性反锐化掩模方法相比,该方法增强后缺陷图像的DSM值最高,缺陷提取区域大小最接近真实缺陷.%Aiming at existing problems of current unsharp masking method in enhancement of defect area of rollers, a new unsharp masking method is introduced through introducing improved high-pass filter template in nonlinear unsharp masking method. The enhancement for defect images of circumferential surface of rollers and extraction effect of defect area are comparatively tested by using linear, non-linear unsharp masking methods and method in this paper. The distribution separation measure is used to quantitatively evaluate enhanced effect of images. The test results show that the nonlinear unsharp masking method based on improved high-pass filter template is able to effectively improve enhanced effect of defect area. Compared with linear and nonlinear unsharp masking methods, the DSM value of enhanced defect images is the highest, and the size of defect extraction area is closest to real defect.