| Image is one of the important ways for human to obtain information,and it is also an important carrier for human to understand the objective world.But in the process of image acquisition,coding,transmission and storage,the image will be polluted by the noise of various electronic devices,resulting in the problem of image degradation.Therefore,it is necessary to denoise the image to facilitate the subsequent image segmentation,recognition and other operations.In underwater image acquisition,due to the absorption and scattering of light by water body and various kinds of suspended particles contained in water body,underwater image will be seriously degraded,and the underwater image will be green or blue,with color distortion,fuzzy edge details,low contrast and other problems.In order to obtain more image information,underwater image needs to be processed Line color equalization and contrast enhancement processing to eliminate the color distortion of underwater image and enhance the clarity.The existing equipment and methods are not very accurate for the establishment of underwater image model,which will affect the results of underwater image processing.As an image processing tool,wavelet transform has been applied to the field of image processing,and has good time-frequency analysis ability.Based on wavelet transform,this paper designs an image denoising method with variable threshold(VTWD),and proposes an underwater image enhancement algorithm based on wavelet fusion wavelet fusion,UIEWF).In the VTWD method,an adaptive wavelet threshold is designed.The threshold value decreases with the increase of the number of decomposition layers by using the shrinkage factor,which is in line with the characteristics of the noise coefficient decreasing in the decomposition process.The noise coefficient and the signal coefficient can be distinguished well,so the noise coefficient can be eliminated.At the same time,a kind of continuous steady-state error and continuous(EEAC)wavelet threshold function is proposed.By controlling the size of the parameter m,the trend of the threshold function is changed,then the steady-state error and continuity are changed,and the precision of image reconstruction is improved.The simulation results show that when 56<m<59,the reconstruction accuracy of EEAC wavelet thresholdfunction is the highest.By denoising eight standard images such as Lena with ?=30 Gaussian noise concentration,the average PSNR and SSIM obtained by vtwd method are 1.029db and 0.133 higher than those by ATM method.UIEWF method is a method for underwater image enhancement.Firstly,the original underwater image is color restored to obtain the color correction image;then the color correction image is converted to CIELab color space,and the L-component is self-adaptive histogram equalization with limited contrast to obtain the brightness enhancement image;finally,the color correction image and the brightness enhancement image are wavelet fused.The results show that the algorithm can effectively improve the color distortion of the image,improve the clarity and contrast of the image,and it does not need the theoretical model of the underwater image,and is not constrained by the conditional hypothesis and prior knowledge,so it is suitable for underwater image enhancement in different environments.Compared with DCP,msrcr and DAH methods,the average uiqm and uciqe increased[3.3541,0.5469,0.3765]and[0.1302,0.0918,0.0761]respectively.Figure[35]table[4]reference[80]... |
PDF Full Text Request