Underwater Image Restoration and Target Detection Based on Monocular Depth Estimation

Abstract

The underwater image was distorted due to scattering, absorption, colour loss, diffraction, polarization and attenuation with light frequency, which led to the distortion of the underwater image and poor target detection effect. An underwater image restoration algorithm based on monodepth2 monocular depth estimation was proposed, and the target detection algorithm based on depth learning was completed. The depth map obtained by monodepth2 monocular depth estimation was used, and the image depth was integrated into the sea-thru the underwater image restoration method. The experimental results showed that the proposed algorithm achieves a good image restoration effect, which laid a foundation for target detection; The underwater image target detection based on yolov5 was realized, and the experiments verify the effectiveness of the proposed algorithm.