带钢表面缺陷在线检测系统的图像滤波算法

Steel strip image filtering plays quite significant role in steel strip surface defects online detection system. On one hand, unnecessary texture background in digital images could be removed by filtering. On the other hand, filtering effects would directly affect the following image processing and analysis. The defects cannot be detected correctly is one of the most serious problems existing in current steel strip surface defects online detection system. This may result from texture background of the image which hasn’t been filtered effectively. Aiming to solve this problem, a novel wavelet-based anisotropic diffusion image filter method has been proposed. Firstly, the original image is decomposed into the low frequency and high frequency components by wavelet transform. Then, the high frequency components are regularized by wavelet diffusion coefficients. Finally, the filtered image is reconstructed by taking the inverse wavelet transform. To achieve a satisfactory filtering result, wavelet-based anisotropic diffusion is often performed iteratively. Experimental results indicate that this new approach could not only filter the unnecessary texture background but also preserve valuable detail information effectively. With more favorable filtering combination properties, this approach will lay a solid foundation for the following image processing, e.g. image edge detection, image auto-segment, etc.