Deepsea imaging

Abstract

This final year report of Deepwater Imaging, primarily mastering of the software MatLab have to come into place before I could kick start the whole research of using algorithm to do any form of colour balancing mainly to increase the quality of images taken under the sea. For instance histogram equalization which usually have been used to tackle issues on grayscale images. However, it could be also used on colour images. A detailed comparison of the 2 types of images that I have taken into consideration for carrying out the research. 1. Grayscale Image and 2. Colour Image. Subsequently, understanding how the discrete intensity level range does matter as it varies in terms of the results as the discrete functions do draw an association with every intensity level. Moving on, the 3 steps to understanding the transformation of histogram from normalization to equalization and lastly the segmentation. Putting the Gaussian or Normal distribution rule into place with mean and standard deviation to ensure everything is in place. After which the construction of the histogram with comparison of histograms of the 2 types of images and understanding how intensity could affect the clarity of the image with the existence of the Red, Blue & Green channels. It appears that bin sizes and there is a difference when comes to blending in the images and also just applying the algorithm individually to the various image channels to do enhancement to the quality of these poor images. The aim of this project is to do a comparative scrutiny of different enhancement methods. And they are done based on judging the visual quality which will be further explained.