Methods, devices, equipment, and media for identifying highway landslide hazards based on multi-source image fusion and cross-modal collaboration

By performing seasonal-weather joint enhancement and modal normalization stitching on optical remote sensing images, and combining cross-modal channel weights and spatial attention weighting, the improved U-Net network was input to solve the problem of weakened terrain structure information in multi-source satellite image fusion, and high-precision landslide disease identification under complex landforms was achieved.

CN121904541BActive Publication Date: 2026-05-26CHANGSHA UNIVERSITY OF SCIENCE AND TECHNOLOGY
View PDF 2 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
CHANGSHA UNIVERSITY OF SCIENCE AND TECHNOLOGY
Filing Date
2026-03-26
Publication Date
2026-05-26

AI Technical Summary

Technical Problem

Existing multi-source satellite image fusion methods fail to effectively handle the characteristics of different modal data in landslide disease identification, resulting in weakened DEM topographic structure information. Furthermore, the feature extraction and reconstruction methods of the Unet model lead to insufficient representation of topographic structure, making it difficult to meet the high-precision identification requirements under complex terrain.

Method used

Enhance optical remote sensing images using seasonal-weather joint data, modal scale normalization and stitching, combined with cross-modal channel admission control and spatial attention weighting, input into an improved U-Net network for feature extraction, filtering and reconstruction, including dual-path heterogeneous downsampling, gated skip connections and decoder feature calibration.

Benefits of technology

It improves the accuracy of identifying landslide hazards on highways in complex terrain environments and enhances the precision of boundary identification. It preserves terrain structure information, suppresses feature redundancy noise, and achieves high-precision landslide target identification.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121904541B_ABST
    Figure CN121904541B_ABST
Patent Text Reader

Abstract

This application discloses a method, device, equipment, and medium for identifying highway landslide hazards based on multi-source image fusion and cross-modal collaboration, relating to the field of satellite landslide disaster technology. The method includes: acquiring registered optical remote sensing images and digital elevation model data; performing seasonal-weather joint data enhancement on the optical images; obtaining a four-channel fused feature map through modal normalization and stitching; further enhancing the features through cross-modal channel admission control and spatial attention weighting; inputting the enhanced features into an improved U-Net network containing dual-path heterogeneous downsampling, gated skip connections, and decoder feature calibration; and achieving highway landslide hazard identification through feature extraction, filtering, calibration, reconstruction, pixel classification, and boundary refinement. This method preserves terrain structure information, suppresses feature redundancy noise, and can accurately identify landslide targets in complex terrain, improving the accuracy of highway landslide hazard identification and boundary recognition in complex terrain environments.
Need to check novelty before this filing date? Find Prior Art