Gabor滤波器在彩色纹理表面缺陷检测中的应用

Application of Gabor Filtering Methods to Auto-defect-detecting of Color Texture Surface

提出了彩色纹理表面自动缺陷检测的Gabor滤波方法,以同时测度图像中的颜色和纹理偏差。提出的方法不依赖于纹理特征的提取,它基于Gabor滤波器和两个颜色特征复数的彩色图像卷积的能量响应,将彩色纹理图像缺陷检测的复杂问题转换成了滤波图像中的简单的二值化问题。对纺织品、木材等许多实际彩色纹理表面的实验结果表明了该方法的有效性。

This paper introduced the application of Gabor filtering method to auto-defect-detecting of color texture surface, to measure the deviations of image＇ s color and texture. The method was based on the energy response of image convolution comprised by two color feature pluralities, rather than the extraction depending on texture feature, thus accomplishing the transformation between a complicated process of defect-detecting and a simple binary of filtering image. The outcomes drawn from a series of tests on a large number of particular color texture surface, such as surface of textile and wood prove the effectiveness of the method.