Image denoising in wavelet and spatial domain

Abstract

The goal of any de-noising technique is to remove noise from an image which is the first step in any image processing. The noise removal method should be applied watchful manner otherwise artefacts can be introduced which may blur the image. In this work, three levels of Gaussian noise are used for adding noise on the original image (σ=10, σ=50, σ=100) and also (σ=15, σ=20, σ=25) to compare with previous work and analysis with it to test embedded system with median filter. Performance evaluation of the median filter, wavelet threshold de-noising techniques is provided. The techniques used are namely the median filter and wavelet threshold is used to remove noise based on raspberry pi using Python 2.7.9 with Open CV 3.2. Four methods to remove noise image are used. The first method local median (LM) which it is widely used as it is very effective at removing noise while preserving edges images and less noise, and the second method is wavelet hard threshold. The wavelet Haar is used, where the Peak signal-to-noise ratio (PSNR) is high value. The third method median before wavelet threshold and the fourth is median after wavelet threshold. The results for each method are introduced as a table for different ten images. The image camera was better than other after applying four methods for the Gaussian noise σ=10. In other hand the other images were better than image of camera for the Gaussian level 50 and 100. The results were good in median filter in wavelet threshold based on Raspberry Pi.