Landslide risk assessment method based on multi-modal learning

The invention discloses a landslide risk assessment method based on multi-modal learning, and mainly solves the problem of low landslide risk assessment precision caused by the fact that the existing method cannot process multi-modal data. According to the implementation scheme, the method comprises...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: SHI XIAOMENG, LIU RUOCHEN, LI WEIBIN, WANG KAIRUI, WANG RONGFANG, LYU HAOYUAN
Format: Patent
Sprache:chi ; eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:The invention discloses a landslide risk assessment method based on multi-modal learning, and mainly solves the problem of low landslide risk assessment precision caused by the fact that the existing method cannot process multi-modal data. According to the implementation scheme, the method comprises the following steps: extracting a digital elevation model corresponding to a high-resolution remote sensing sub-graph; semantic segmentation is carried out on the high-resolution remote sensing subgraph to extract the edge of the landslide risk point; calculating a slope map and a slope graph corresponding to the extracted digital elevation model; extracting respective corresponding feature sequences from the slope map and the slope graph; performing residential spot detection and road detection on the to-be-evaluated high-resolution remote sensing subgraph; carrying out image embedding on the high-resolution remote sensing subgraph, the slope graph and the slope graph; performing sequence embedding on the feature