LSDetector: An Open-Source Tool Bridging Landslide Detection Models and Practical Deployment through Three-Stage Transfer Learning
ID：65
Submission ID：69 View Protection：ATTENDEE
Updated Time：2026-07-30 17:32:14
Hits：5
Oral Presentation
Start Time：2026-08-11 11:15 （Asia/Hong_Kong）
Duration：15min
Session：[S2] Session 2 Remote Sensing of Geoenvironmental Disasters » [S2.5] Session 2 Day 3
No files
Abstract
Reusable landslide detection from remote-sensing imagery is limited by cross-region domain
shifts, scarce local labels, and weak links between model research and operational mapping.
This paper presents LSDetector, an open-source local workbench that packages LSDFormer and
LSDSAM with model-weight, dataset, adaptation, inference, evaluation, and GIS-export modules.
The workflow supports task-adaptive fine-tuning, domain-adversarial fine-tuning, and targetspecific
fine-tuning, then iteratively converts AI predictions and expert corrections into improved
regional inventories. In the Wuping rainfall-triggered landslide area, LSDSAM-H achieved the
best benchmark performance, whereas LSDFormer provided the fastest deployment option. A
large-area deployment over 17,295 km2 of PlanetScope imagery produced 40,400 candidate
landslide polygons, demonstrating the potential of LSDetector for reviewable, semi-automatic
landslide inventory construction.
shifts, scarce local labels, and weak links between model research and operational mapping.
This paper presents LSDetector, an open-source local workbench that packages LSDFormer and
LSDSAM with model-weight, dataset, adaptation, inference, evaluation, and GIS-export modules.
The workflow supports task-adaptive fine-tuning, domain-adversarial fine-tuning, and targetspecific
fine-tuning, then iteratively converts AI predictions and expert corrections into improved
regional inventories. In the Wuping rainfall-triggered landslide area, LSDSAM-H achieved the
best benchmark performance, whereas LSDFormer provided the fastest deployment option. A
large-area deployment over 17,295 km2 of PlanetScope imagery produced 40,400 candidate
landslide polygons, demonstrating the potential of LSDetector for reviewable, semi-automatic
landslide inventory construction.
Keywords
Landslide detection, remote sensing, deep learning, foundation model
Speaker
Comment submit