Underwater image enhancement method based on progressive feedback network

Abstract

The invention provides an underwater image enhancement method based on a progressive feedback network, and the method comprises the following steps: S1, carrying out the pairing processing of underwater image data for training, carrying out the data enhancement and normalization processing, and obtaining a to-be-trained paired image; s2, inputting the to-be-trained paired images into a multi-stage progressive image enhancement network capable of enhancing the images in combination with discrete wavelet transform and an attention feedback mechanism at each stage, training an image enhancement model capable of enhancing underwater images, and correcting the stages of the network by using an attention supervision module; s3, setting a target loss function of the image enhancement network; s4, using the paired training images to enhance network convergence to Nash equilibrium; s5, performing normalization processing on an underwater image to be enhanced, inputting the trained image enhancement model, and outputting an enhanced image; according to the invention, the underwater image quality can be improved.

Description

technical field [0001] The invention relates to the technical fields of image processing and computer vision, in particular to an underwater image enhancement method based on a progressive feedback network. Background technique [0002] Underwater image enhancement technology has attracted much attention due to its significance in the fields of ocean engineering and underwater robots. The quality of underwater imaging has a great impact on underwater operations. For example, vision-dependent tasks such as seabed exploration and underwater target detection have high requirements on the quality of underwater images. Low-quality underwater images will lead to the inefficiency of these tasks. and accuracy drops severely. The enhancement of underwater images is a challenging problem due to the complexity of the underwater environment and lighting conditions. Generally, underwater images are affected by wavelength-dependent absorption and scattering, including forward scattering...

AI Technical Summary

Problems solved by technology

However, the underwater environment is complex and the lighting and other conditions are changeable. It is difficult to estimate the parameters and the accuracy is low, resulting in the generally low quality of the enhanced image. At the same time, different environmental considerations are different, and the models that need to be established are also different. These methods have great limitations

[0004] The existing methods are usually accompanied by information loss in the process of image enhancement, which leads to blurred details in the enhanced image.

Images