A study on color model selection for underwater color image preprocessing

Abstract

Underwater captured images suffer from quality degradation and blurring due to light absorption and scattering. Different color models combining with various preprocessing methods are used to overcome such problems, performing varying degrees of effect. Our goal is to analyze and evaluate the various color models in underwater images preprocessing. Three existing common underwater image preprocessing methodologies include contrast limited adaptive histogram equalization, homomorphic filtering and wavelet threshold denoising are applied to measure the performance of each color model in terms of three objective parameters including mean square error (MSE), peak signal to noise ratio (PSNR) and structural similarity index measure (SSIM). Experimental results demonstrate that the color models applied in different preprocessing techniques has various processing results while HSI and YUV color models relatively perform better in the underwater color image preprocessing. The both models give less MSE error and high PSNR ratio and high SSIM index.