Underwater image enhancement method based on fractional order convolutional neural network

Abstract

The invention discloses an underwater image enhancement method based on a fractional order convolutional neural network, and the method comprises the steps: inputting an underwater image, carrying out the preliminary preprocessing of the underwater image through white balance and histogram equalization, designing an ambient light estimation network and a transmissivity estimation network of the underwater image, carrying out the parameter training of the two estimation networks, and carrying out the recognition of the parameters of the two estimation networks. And outputting the ambient light estimation network to obtain an ambient light value B, outputting the transmissivity estimation network to obtain a transmissivity parameter t, and restoring according to an underwater physical model to obtain a clear image. According to the method, on the premise that a large number of high-quality clear underwater images are not needed, the visual quality of the enhanced underwater image is obviously improved by calculating a group of IQMs to evaluate the result, and the enhancement effect is remarkable.

Description

technical field [0001] The invention belongs to the technical field of computer vision, and in particular relates to an underwater image enhancement method based on a fractional convolutional neural network. Background technique [0002] There are abundant resources in the ocean. As an important carrier of ocean information, underwater images play an irreplaceable role in ocean exploration and development. Due to the influence of the complex underwater environment, the quality of underwater images often degrades. In order to improve the low contrast, color distortion, and blurred details of the original underwater image, it is necessary to enhance the underwater image to obtain clarity. underwater image. [0003] The existing underwater image enhancement technology has made great progress, and the effect of underwater image enhancement is relatively obvious. However, restoring the visual quality of underwater images still faces great challenges. [0004] There are two main...

AI Technical Summary

Problems solved by technology

[0005] The purpose of the present invention is to address the difficulties and problems of the lack of a large number of real underwater images in reality, in order to enhance the visual quality of the original underwater images, and improve the learning rate in the network parameter training process, and provide a method based on fractional convolution Underwater image enhancement method based on neural network

Images