Enhancement of Over-Exposed and Under-Exposed Images Using Hybrid Gamma Error Correction Sigmoid Function

Mohd Azau, Mohd Azrin (2007) Enhancement of Over-Exposed and Under-Exposed Images Using Hybrid Gamma Error Correction Sigmoid Function. Masters thesis, Universiti Putra Malaysia.

[img] PDF
1119Kb

Abstract

The demands to improve the visibility quality of the captured images in extremes lighting conditions have emerged increasingly important in digital image processing. The extremes conditions are when there is lack of reasonable lightnings termed as underexposed and too much of light termed as overexposed. The popular enhancement technique currently used is the contrast enhancement through contrast stretching, histogram equalization, homomorphic filtering and contrast adjustment. The adjustments are to transform the less useful images to more meaningful images when the post image processing operations are carried out. This thesis is motivated to deal with the problems concerning image capturing in these two extremes conditions. The sigmoid function is used to adjust the contrast with two controlling parameters. The parameters adjust the contrast locally and globally. The gamma function is commonly used to correct the non-linear error in the images due to the camera lenses. This thesis combines the functions' properties and developed a hybrid algorithm to improve the quality of the poorly captured images by adjusting the contrast and compensating the gamma error. The sigmoid and gamma function are coded in MATLAB 6.0 in which testes are made over the selected images. The sample images are taken using different type of cameras transformed to grayscaled input images. The luminosities of the surroundings are also measured using a light meter. The derivations of the parameters' ranges are done by calculating the root mean square error or the standard deviation. The suggested ranges are used in the hybrid system which has two variants, Variant I and Variant 11. The first variant, combines the sigmoid function inside the gamma compensation function while the second variant combines the gamma compensation function inside the sigmoid function. Based on the test results, the proposed algorithm significantly improves the contrast of the images. For the underexposed image samples, the percentages of the intensity lesser than 0.1 decreases as more of the intensities reside at higher values. For the overexposed image samples, the percentages of intensity greater than 0.9 decreases as more of the intensities reside at lower values. With the suggested range deduced, the images are contrast enhanced with the reduction of percentage of pixels residing he intensity less than 0.1 and greater than 0.9. The comparative analyses are made by comparing the suggested hybrid system with the existing adaptive homomorphic filtering, adaptive histogram equalization and adaptive contrast enhancement.

Item Type:Thesis (Masters)
Subject:Gamma functions
Subject:Imaging systems
Chairman Supervisor:Associate Professor Ishak bin Aris, PhD
Call Number:FK 2007 12
Faculty or Institute:Faculty of Engineering
ID Code:6166
Deposited By: Nur Izyan Mohd Zaki
Deposited On:12 May 2010 08:23
Last Modified:27 May 2013 07:27

Repository Staff Only: item control page

Document Download Statistics

This item has been downloaded for since 12 May 2010 08:23.

View statistics for "Enhancement of Over-Exposed and Under-Exposed Images Using Hybrid Gamma Error Correction Sigmoid Function "


Universiti Putra Malaysia Institutional Repository

Universiti Putra Malaysia Institutional Repository is an on-line digital archive that serves as a central collection and storage of scientific information and research at the Universiti Putra Malaysia.

Currently, the collections deposited in the IR consists of Master and PhD theses, Master and PhD Project Report, Journal Articles, Journal Bulletins, Conference Papers, UPM News, Newspaper Cuttings, Patents and Inaugural Lectures.

As the policy of the university does not permit users to view thesis in full text, access is only given to the first 24 pages only.