An Image Filtering Algorithm Based On The Granular Computing Theory Of Quotient Space

Image is the most important source for people to gain and exchange information, so the high-quality image is the prime condition to important information acquisition and decision making. However, the imaging will be affected by equipment location, posture, environmental conditions and other subjective or objective factors, resulting in poor image quality can not meet the need, Therefore, the necessary of the image processing is particularly important.The denoising is a classical problem in image processing, however, there is no universal filtering algorithm, especially in the facing of the higher intensity noise. In this paper, Summary analysis of the classic image filtering algorithms, the granular inverse harmonic mean filter algorithm is proposed based on granular computing, can filter out high intensity degree noise better. In this paper, are as follows:Overview of image filtering algorithms to explore the advantages and shortcomings of the classic algorithms in image filtering, and point out that the method can be improved.Image filtering algorithm based on granular computing, in the noise detection stage, by treatment of the detection window for granular segmentation according to certain principles, at the same time make the appropriate noise detection, according to principle of falsity preserving to judge whether another division to the classification mark of the final testing of noise pixels to give the upper limit of noise and lower noise. It can effectively overcome the inadequacies of the traditional filtering algorithm. In the noise detection stage, the emergence of false detection due to interference between the noise pixels, missed and so onIn the noise filtering stage, Traditional filtering algorithm in a single mode, by the inverse harmonic mean filter for the filter characteristics, noise were marked for different types of noise pixels in the corresponding detection granularity space in the implementation of the noise filter, and is conducive to reduce interference, divisible noise, protect the edges of the image, texture, details and other features.Full use of high-speed of the granular computing theory method and efficient features, with the advantages of the inverse harmonic mean filter the granular inverse harmonic mean filter to be formed based on granular computing,has harmonic filter the advantages of fast and efficient filtering purposes, has a good robustness and real-time, rich and developed the theory of image filtering algorithms. The test results show that different types of image and noise filtering algorithm proposed in this paper and the granular inverse harmonic mean filtering has the best results.