Underwater target identification and positioning method based on physical model and deep learning fusion

By using an adaptive monocular vision method based on physical models and deep learning, the problem of insufficient stability and positioning accuracy of underwater target recognition and localization in complex waters is solved, and high-precision target recognition and localization is achieved under different turbidity and lighting conditions.

CN121811097APending Publication Date: 2026-04-07CENT SOUTH UNIV
0 Cites 2 Cited by

Patent Information

Application Number
CN202511744879.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-11-25
Publication Date
2026-04-07

AI Technical Summary

Technical Problem

Underwater target identification and positioning suffer from poor identification stability, insufficient positioning accuracy, and difficulty in suppressing refraction errors in complex waters. Existing solutions are not robust enough under changes in environmental parameters and light fluctuations, making it difficult to meet the requirements for high-reliability operations.

Method used

An adaptive monocular vision method based on the fusion of physical models and deep learning is adopted. The underwater imaging physical degradation model is used as a prior, and the dark channel prior and convolutional neural network are combined for image restoration. SuperPoint and R2D2 are used for key point detection and description. The attitude parameters are solved by combining weighted RANSAC and the eight-point method. Pixel-level fusion is performed by refraction-corrected triangulation and MiDaS depth estimation to achieve stable recognition and high-precision positioning.

Benefits of technology

It achieves stable target recognition and high-precision positioning under complex water conditions, improving the equipment's environmental adaptability and positioning accuracy, and is suitable for various types of underwater robot platforms.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention discloses an underwater target identification and positioning method based on fusion of a physical model and deep learning, and belongs to the technical field of computer vision, and the method comprises the steps: carrying out the transmissivity estimation and physical restoration of a collected underwater monocular image according to an underwater light propagation model, and carrying out the image correction; key feature points are extracted, and high-quality matching point pairs are screened in combination with the joint similarity; further deriving a basic matrix through the high-quality matching point pairs meeting the epipolar geometric constraint relation, solving an essential matrix, and obtaining relative attitude parameters between the cameras in combination with weighted re-projection-LM optimization; and finally, refraction correction triangulation is carried out according to the Snell's law, pixel-level fusion is carried out after scale normalization and space alignment are carried out on the refraction correction triangulation and dense depth output by the MiDaS, three-dimensional space coordinates of the target are inverted, and high-precision recognition and positioning of the underwater target are achieved. The system is light in structure, efficient in calculation, suitable for being integrated on various underwater autonomous or remote control robot platforms and used for tasks such as target recognition, tracking and positioning.
Need to check novelty before this filing date? Find Prior Art