一步异相关与最相关滤波算法(英文)测试技术英文学报20120702.docVIP

The Most Relevant Weighted Filtering with One-step Singular Correlation Recognition Liu Chao(刘超)1，Cai Jianchao(蔡建超)2 (1, School of Mechanics, Jinzhong University, Jinzhong 030600; 2, College of Physics and Electronics Engineering, Shanxi University, Taiyuan 030006) Abstract: Based on the recognition of one-step singular correlation and remedying method obtained before, the correlation properties of the neighborhood pixels and the characteristics of de-noising for images were further analyzed. A kind of the most relevant weighted filtering based on one-step singular correlation recognition (OSSC-MRWF) method was put forward. The simulation experiments and comparisons with some current common methods under the salt-and-pepper noises were done. The results show that this method not only could effectively recognize the salt-and-pepper noises and mend up the noise points, but also protect the original information such as the edge details very well. The accuracy and performance indicators are further improved considerably. Keywords：correlation function; image filtering; one-step singular correlation; singular value recognition; weighted filtering; salt-and-pepper noise. Manuscript Number: dio: 1 Introduction It is unavoidable to be disturbed by noises in imaging, transform, receive, and processing, etc. which causes the image lack fidelity or vagueness and affects the later stages for images recognition and restoration, segmentation, and feature extraction. Therefore，the noise filtering always is one of important tasks for image processing. The salt-and-pepper noise commonly exists in general practice and the median filtering is widely used for these kinds noise, which could satisfy ordinary requirements in a certain degree. So far, there are many improved methods based on the median filtering or weighted filtering, etc. The optimized algorithm studied by Brest and Lukac based on multi-point search in the mass, which offers a good ideal for improving the median