Quality enhancement is an important issue to be solved in image processing. There are numerous works which has been conducted for enhancing the quality of underwater images. Due to absorption and scattering of light the underwater image gets blurred. In addition, Time complexity (TC) using existing techniques are also higher. In order to overcome these limitations, Biorthogonal Wavelet Transform Principle Component Analysis Based Non-linear Bilateral Filtering (BWTPCA-NBF) technique is proposed. The aim of BWTPCA-NBF technique is to reduce the scattering effect of underwater image by enhancing the image pixels by extended approach of image enhancement. Initially, BWTPCA-NBF technique performs decomposition process where it applies 9/7 Biorthogonal filter based wavelet transformation to segment the input underwater images into a number of sub-bands. Subsequently, BWTPCA-NBF technique identifies the noise pixels that are impacted by scattering and light absorption effects in each sub-band with application of Kernel Principal Component Analysis (KPCA). Finally, BWTPCA-NBF technique increases the quality of input underwater images by normalizing each identified pixel with the weighted average of intensity values from near pixels. As a result, BWTPCA-NBF technique attains improved underwater images quality with a minimal time. This provides better result of underwater image enhancement than existing methods. The simulation analysis of BWTPCA-NBF technique is conducted using parameters such as Mean Square Error (MSE), Peak Signal to Noise Ratio (PSNR) and TC. Simulation result confirms that BWTPCA-NBF technique improves PSNR and minimizes TC of underwater image enhancement as compared to existing works.
Volume 11 | 07-Special Issue
Pages: 364-375