Removal of salt and pepper noise from images using hybrid filter (HF) and fuzzy logic noise detector (FLND)

Abstract

SummaryImage restoration is the practice of removing or reducing the degradation. Degradation occurs due to image acquisition, out of focus, and image transfer over the internet. Image restoration tries to recover images that have been degraded. The restored image is not an original image; it is an approximation of the actual image. Image restoration is the preprocessing task done before other image processing tasks such as image segmentation, image compression, etc. This research implements a restoration algorithm using hybrid filter (HF) and fuzzy logic noise detector (FLND) for the removal of impulse noise from images. The proposed technique consists of two stages. In the first stage, we split the image into a number of windows, and each window applies the hybrid filter (Mean Filter and Adaptive Median Filter). The output from the first stage is given as an input to the second stage in which FLND generates the fuzzy rules, the rules used to classify the pixel as noisy and noise free. If the pixel is considered as noisy pixel, then it is restored by median filter. The noise‐free pixels were left unchanged. Increasing the (peak‐signal‐to noise ratio) PSNR of an image is the foremost objective of this research work. The proposed method is evaluated on standard lena image and the PSNR value at different noise level are presented in this research paper. The results are compared with Mean Filter (MF), Adaptive Median Filter (AMF), Weighted Median Filter (WMF), Decision‐Based Filter (DBF), Weighted Fuzzy Mean Filter (WFM), Adaptive Fuzzy Mean Filter (AFM) and Wavelet Filter (WF). The drawbacks in existing methods are rectified in our proposed method. It suppresses noise, protects image characteristics, and also increases visual quality and PSNR value.