Design and Development of Laplacian Pyramid Combined with Bilateral Filtering Based Image Denoising

This paper mainly deals with image denoising algorithm by applying bilateral filter in Laplacian subbands using raspberry pi. Bilateral filter is used for reducing various noises especially Additive White Gaussian Noise which occur more in standard test images. The important feature of this nonlinear bilateral filter is the preservation of the edges, while reducing the noise in the images. The main idea is to replace the pixel's intensity value by a weighted average of adjacent pixel intensity value. Euclidean distance and radiometric differences of pixels, are used for weight calculation. This calculation mainly preserves sharp edges by looping through each pixel and adjusting weights to the nearest pixel values. Our project aims to reduce cost and power consumption using Raspberry pi. It consists of series of microcomputer packed onto a single circuit board. These low power computers are mass produced at very low prices. The performance of Raspberry pi is equivalent to a personal computer. In order to perform noise reduction in Raspberry pi, python2 and opencv package is installed in real-time Raspbian Linux operating system and algorithm is executed using python2 programming language. The denoising method using Laplacian subbands provides better denoised images compared to Gaussian filter and Bilateral filter that applied in standard test images.