Crop classification method for multi-temporal remote sensing images based on spatio-temporal attention U-shaped network

By constructing a spatiotemporal attention U-shaped network model, the problems of insufficient utilization of the temporal characteristics of multi-temporal remote sensing data and the impact of cloud cover are solved, achieving high-precision crop classification and good generalization ability, and reducing the deployment threshold.

CN120808040BActive Publication Date: 2026-06-09DALIAN MARITIME UNIVERSITY
View PDF 3 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN Β· China
Patent Type
Patents(China)
Current Assignee / Owner
DALIAN MARITIME UNIVERSITY
Filing Date
2025-07-29
Publication Date
2026-06-09

AI Technical Summary

Technical Problem

Existing crop classification methods struggle to fully utilize the temporal characteristics of multi-temporal remote sensing data, and temporal remote sensing data is easily affected by factors such as cloud cover, resulting in low classification accuracy and high deployment barriers.

Method used

A multi-temporal remote sensing image crop classification method based on a spatiotemporal attention U-shaped network is adopted. By constructing a spatiotemporal attention U-shaped network model, a convolutional block attention module, a lightweight temporal attention encoder module, a dynamic upsampling module, and an adaptive feature fusion module are integrated to adaptively process temporal noise and achieve high-precision classification.

Benefits of technology

It achieves high-precision crop classification without relying on complex preprocessing, and improves computational efficiency and the generalization ability of the classification model.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120808040B_ABST
    Figure CN120808040B_ABST
Patent Text Reader

Abstract

The application belongs to the technical field of remote sensing image processing, and relates to a multi-temporal remote sensing image crop classification method based on a space-time attention U-shaped network, which classifies crops in multi-temporal remote sensing images by constructing a space-time attention U-shaped network model, integrating a convolution block attention module, a lightweight time attention encoder module, a dynamic up-sampling module and a self-adaptive feature fusion module, adaptively capturing crop key growth period features through a space-time attention mechanism during the training process of the network model, effectively suppressing cloud cover interference, and fully utilizing time series information and spatial feature information.Under the condition that the number of training sets is sufficient, accurate classification of farmland crops in multi-temporal remote sensing data is realized, and good results are achieved.The method exhibits strong environmental adaptability in spatial generalization and temporal generalization, and has obvious advantages in spatial generalization.
Need to check novelty before this filing date? Find Prior Art