Generative self-supervised hyperspectral image target detection method based on spectral mask

By using spatial spectral masks and generative self-supervised learning, combined with a lightweight Transformer encoder, the problems of high label sample requirements and poor generalization ability in hyperspectral target detection are solved, achieving efficient target detection and improved accuracy.

CN117746235BActive Publication Date: 2026-06-23CHINA UNIV OF GEOSCIENCES (WUHAN)
2 Cites 0 Cited by

Patent Information

Application Number
CN202311694328.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-12-08
Publication Date
2026-06-23
Estimated Expiration
2043-12-08

AI Technical Summary

Technical Problem

Existing hyperspectral target detection methods rely on a large number of labeled samples and have poor generalization ability, making it difficult to effectively transfer to different scenarios and tasks, and resulting in low computational efficiency.

Method used

A spatial spectrum masking strategy is used to mask hyperspectral remote sensing images. Combined with generative self-supervised learning and a lightweight Transformer encoder, feature representations are extracted from unlabeled data through self-supervised learning and fine-tuned using prior spectra to construct a target detection network.

Benefits of technology

It significantly reduces the need for labeled samples, improves the model's generalization ability, and enhances detection accuracy and efficiency, enabling it to better accomplish hyperspectral image target detection tasks.

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

Abstract

The application provides a generative self-supervised hyperspectral image target detection method based on a spectral mask, features of hyperspectral remote sensing data are encoded and decoded through a generative model and a spectral mask strategy, a general feature representation is learned in the absence of labeled samples, a transfer learning model based on a lightweight transformer encoder, a full connection layer and a Sigmoid layer is constructed, a detector composed of a full connection layer and a Sigmoid layer is fine-tuned in combination with a small amount of training samples and a weighted binary cross-entropy loss, rapid training of the transfer learning model is realized, and the pre-training model is effectively transferred to different target detection tasks. The application has the beneficial effects that: the hyperspectral remote sensing image itself features are fully utilized, the general feature representation is learned from the unlabeled data according to the self-supervised learning theory by using the generation task, and the pre-training model is transferred to various detection tasks, so that the detection precision and time efficiency are improved.
Need to check novelty before this filing date? Find Prior Art