Real-Time Video Noise Reduction Techniques Implementation on FPGA

Abstract

The Gaussian filter and the Median filter are two types of noise re duction filters, and this thesis provides a complete presentation and analysis of each of these filters. This was a conscious choice that was motivated by the large amount of study that was undertaken on FPGA implementation for grayscale photos, in contrast to the minimal attention that was paid to color images. The filters have been particularly developed to work on grayscale image formats. Because of the rise in data dimensions, the processing of color im ages necessitates much more complex computations. Furthermore, the similarities in edge characteristics between grayscale and color pictures have led to color image processing being a field that has received comparatively less research than grayscale image process ing. This research makes a substantial contribution to the development of FPGA-based video processing by delivering efficient noise re duction capabilities as well as a unique camera interface solution, both of which are backed by aspecific VLSI architecture. The afore mentioned technological improvements have substantial ramifica tions in a variety of fields, including robotics, surveillance, and au tomation, all of which place a premium on being able to interpret video in real-time