Image noise removal and detection

• Main Author: Edirachcharige Sahan Ruchira Jinadasa
• Other Authors: Kai-Kuang Ma
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2021
• Subjects: Engineering::Electrical and electronic engineering::Applications of electronics
• Online Access: https://hdl.handle.net/10356/153770
• Institution: Nanyang Technological University

In this paper, the effect of different types of noises on images are shown as wells as the various method of detecting and removing noise. Here MATLAB is used as the image processing tool to detect and remove unwanted noises. Most images captured are affected by random noises by some way or the other. Although this noise might seem random to the untrained eye, what lies beneath is a pattern and a root cause. It is crucial for us to understand the different types of noises and their origins before diving into noise detection and de-noising. Therefore, these random noises are identified and explained. Further clarification is given as to how and why these noises are present in images. After grasping the concept of noise, the report will continue by demonstrating the different tools and filters used to overcome these noises. These de-noising methods will be demonstrated on grey-scale images as well as RGB images. Further research was done into the most prevalent noise which is Gaussian noise and how we can rectify images affected by it. After understanding the different types of noises one simple method of detecting noise in image is discussed. Histogram analysis is done on the image to identify the type of noise affecting the image. Next the research was conducted on noise found in digital photography. Two types of noises affecting digital photography is introduced in this report. The negative and positive repercussions of these noises are demonstrated by illustrations here. Here a powerful de-noising software called Light-room is discussed and compared with a much simpler program using MATLAB. These two entities were tested on de-noising corrupted digital images as well as its limitations when enhancing underexposed images. Lastly research was done on the niche area of underwater imagery. Several research and papers has been published regarding this field and in this chapter few of these methods is discussed in detail. Furthermore, a much simpler but productive method of enhancing underwater imagery using MATLAB is discussed. These two methods are compared regarding their effectiveness in enhancing the image.