Research On Underwater Image Restoration Based On Imaging Model

Underwater image is an important source of information for human to explore the underwater world.However,due to the absorption and scattering of light by water,underwater images often have problems such as color distortion,low contrast,low brightness and so on.The quality degradation seriously affects the applications of underwater images in research and engineering,therefore it is important to improve the quality of underwater images.In recent years,underwater image restoration methods based on imaging model have received extensive attention,but existing restoration methods still have problems such as introducing color shift,low robustness,and high complexity.This thesis research on these problems and the main work is as follows.An underwater image restoration algorithm based on gradient channel and optimized transmittance is proposed.The algorithm estimates the background light based on the texture features and color features of the background region of the image.The red channel transmittance is estimated based on the optimization principle of minimum information loss and maximum contrast.Based on the attenuation characteristic of light,the transmittance of blue and green channels is calculated by the background light and red channel transmittance.And the underwater image is reconstructed with an inverse solving imaging model.The algorithm can accurately restore the true color of the object,improve the image contrast,and has high robustness.An underwater image restoration algorithm based on depth estimation is proposed.Firstly,the depth map is estimated based on the characteristics of the selective absorption of light by water and the degree of scattering with distance.Then the farthest value in the depth map is used as the background light estimation.The red channel transmittance is globally calculated using the depth map and the red light decay rate.Finally,the transmittance of the blue and green channels is calculated and the underwater image is restored.The algorithm can not only effectively remove the color shift and enhance the image contrast,but also has the advantages of small amount of calculation and low complexity.