Image Salt and Pepper Noise Adaptive Filtering Algorithm Based on Fuzzy Reasoning

In order to tackle conflicting issues of image salt and pepper noise smoothing and image detail preservation, a new detail protection filter algorithm---Adaptive Fuzzy Directional Median (AFDM) filter algorithm is proposed based on fuzzy reasoning system.The main techniques include fuzzy partitioning of image pixels, construction of noise map, and noise filtering of adaptive filtering. In the process of filtering, the fuzzy membership function for creating the noise map is established by calculating the direction index and the mean deviation. For the noise pixels divided into labels 1 and 2, according to the relative relationship between the direction index and the adaptive threshold, the algorithm automatically selects to perform median filtering or median filtering in a particular direction. The simulation results show that the algorithm can get satisfactory results in filtering effect and detail preservation ability.