Full-floating raft culture area extraction method based on polarimetric SAR (Synthetic Aperture Radar) image

Through the fully floating raft farming area extraction method based on polarized SAR images, using deep learning and feature network model, the problems of low accuracy and high noise influence in the existing methods are solved, and the extraction effect with higher accuracy and integrity is achieved.

CN119741651AActive Publication Date: 2025-04-01HOHAI UNIV
5 Cites 1 Cited by

Patent Information

Application Number
CN202411722626.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-11-28
Publication Date
2025-04-01
Estimated Expiration
2044-11-28

AI Technical Summary

Technical Problem

The existing fully floating raft farming area extraction method based on SAR images has problems with low accuracy and serious impacts of coherent spot noise.

Method used

A fully floating raft farming area extraction method based on polarized SAR images was designed. Through pre-processing, polarization decomposition, band combination and slice processing, a feature network model was constructed. Combined with the advantages of deep learning, the features of the fully floating raft farming area were extracted and the impact on noise and background environment was reduced.

Benefits of technology

The accuracy of extraction in the fully floating raft farming area is improved, the impact on noise and background environment is reduced, and the integrity of the extraction results is enhanced.

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

Abstract

The invention discloses a polarimetric SAR image-based full-floating raft culture area extraction method, which comprises the following steps of: performing preprocessing, polarimetric decomposition and wave band combination on each sample polarimetric SAR image to obtain a multi-dimensional channel characteristic image corresponding to each sample polarimetric SAR image, performing slicing processing, taking a multi-dimensional channel characteristic slice image as input, and extracting a multi-dimensional channel characteristic slice image corresponding to each sample polarimetric SAR image; taking the full-floating raft cultivation local area in the multi-dimensional channel feature slice image as output, and training the constructed feature network model to obtain a full-floating raft cultivation area extraction model; for a to-be-analyzed polarimetric SAR image, applying the full-floating raft culture area extraction model to obtain a corresponding full-floating raft culture area; according to the design method, polarization features of full-floating raft culture in a polarization SAR image are fully utilized, influences of different polarization features on full-floating raft culture area extraction are considered, feature optimization is performed on the different polarization features, a feature set suitable for full-floating raft culture area extraction is constructed, deep learning advantages are combined, and the full-floating raft culture area extraction efficiency is improved. The full-floating raft culture area extraction task is realized, and the influence of speckle noise and background environment on the full-floating raft culture area extraction task is reduced.
Need to check novelty before this filing date? Find Prior Art