Research On Underwater Image Fusion Algorithm Based On Color Compensation And Brightness Equalization

The theory and technology of marine information collection,transmission and processing is the key to the rational development and utilization of marine resources.Visual technology has attracted much attention because of its ability to carry rich information.Therefore,scholars have been trying to obtain high-quality seabed images for use in different underwater applications.However,the quality of seafloor images varies due to their special physical and chemical properties,which not only seriously reduces the underwater image quality obtained by traditional imaging systems,but also reduces the ability to obtain effective information in underwater images.In view of the visual effect,information density and algorithm complexity of underwater degraded images,a method of color compensation and luminance balance is proposed to enhance the multi-scale fusion of underwater images.The main content of this thesis includes the following aspects:(1)Aiming at the problem of color deviation caused by selective attenuation of color channel of underwater image,a preliminary enhancement scheme based on 3C-HF(Color Channel Compensation-Homomorphic Filtering)operator is proposed.First,the color compensation and equalization of the original image is carried out by the white balance method based on color channel compensation,and then the locally homomorphic filtering operation is applied to the color-complemented image to remove the inhomogeneity and ambiguity of the image,and a preliminary enhanced underwater image is obtained.(2)The preliminary enhanced image is further operated to form two different images after brightness correction and detail enhancement.For the problem of uneven luminance distribution caused by uneven underwater illumination,the adaptive gamma correction algorithm based on weighted distribution is applied to the preliminary clear image to complete the brightness correction;for the problem of detail loss of underwater image,the adaptive histogram equalization method with limited contrast is used to realize the detail enhancement in Lab space.The preprocessing operation of underwater image brightness and detail enhancement is completed,and two images to be fused are obtained.Then the weight of the fused image is calculated to realize the multi-scale fusion of the underwater image,and the exposure weight,local and Laplace contrast weight and saturation weight are selected to calculate the normalized weight of the detail enhanced image and the luminance correction image.In this thesis,the multi-scale fusion technique is used to decompose and reconstruct the image,and the enhancement processing based on image fusion is realized.(3)This thesis uses several underwater images for experiment and analysis,and evaluates the algorithm from both subjective and objective angles.At the same time,it also carries on the practical application of high-order tasks to the experimental results.Compared with the existing enhancement methods,this method can not only effectively improve the clarity of the image,but also ensure the authenticity of underwater color.It also enhances our visual perception and verifies the robustness and practicability of the algorithm.