Performance analysis of pre-processing filters for underwater images In oceanic environment, uneven illumination, turbulence in water and floating particles make underwater image capture, a challenge. Vision sensors attached with the autonomous underwater vehicles, themselves cause light dispersion and shadows in the ocean floor. Although, several computer vision algorithms have been developed, effective analysis of the algorithms both quantitatively and qualitatively have not been done. This paper analyses the existing methods for the inherent problems and provides a framework for underwater image processing. Initially, for non-uniform illumination correction, homomorphic, anisotropic and bilateral filtering techniques are compared for contrast equalization. Contrast enhancement is done using contrast limited adaptive histogram equalization (CLAHE) with adaptive histogram clip. Finally, Haar wavelet and Symlet are compared for adaptive smoothing, elimination of remaining noise and for improving edge detection. Performance is assessed by computing peak signal noise ratio (PSNR), contrast to noise ratio (CNR), image enhancement metric (IEM), and absolute mean brightness error (AMBE). Histograms are computed before and after applying pre-processing filters, for evaluating the proposed methodology. A combination of homomorphic filtering, CLAHE and haar wavelet denoising provides better results over other methods for underwater images.