A Benchmark Study of Segmentation Models and Adaptation Strategies for Landslide Detection from Satellite Imagery
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866915943361282048 |
|---|---|
| author | Kowsher, Md Zhan, Weiwei Chen, Chen |
| author_facet | Kowsher, Md Zhan, Weiwei Chen, Chen |
| contents | Landslide detection from high resolution satellite imagery is a critical task for disaster response and risk assessment, yet the relative effectiveness of modern segmentation architectures and finetuning strategies for this problem remains insufficiently understood. In this work, we present a systematic benchmarking study of convolutional neural networks, transformer based segmentation models, and large pre-trained foundation models for landslide detection. Using the Globally Distributed Coseismic Landslide Dataset (GDCLD) dataset, we evaluate representative CNN- and transformer-based segmentation models alongside large pretrained foundation models under consistent training and evaluation protocols. In addition, we compare full fine-tuning with parameter-efficient fine-tuning methods, including LoRA and AdaLoRA, to assess their performance efficiency tradeoffs. Experimental results show that transformer-based models achieve strong segmentation performance, while parameter efficient finetuning reduces trainable parameters by up to 95% with comparable accuracy to full finetuning. We further analyze generalization under distribution shift by comparing validation and held-out test performance. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2604_16663 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | A Benchmark Study of Segmentation Models and Adaptation Strategies for Landslide Detection from Satellite Imagery Kowsher, Md Zhan, Weiwei Chen, Chen Computer Vision and Pattern Recognition Landslide detection from high resolution satellite imagery is a critical task for disaster response and risk assessment, yet the relative effectiveness of modern segmentation architectures and finetuning strategies for this problem remains insufficiently understood. In this work, we present a systematic benchmarking study of convolutional neural networks, transformer based segmentation models, and large pre-trained foundation models for landslide detection. Using the Globally Distributed Coseismic Landslide Dataset (GDCLD) dataset, we evaluate representative CNN- and transformer-based segmentation models alongside large pretrained foundation models under consistent training and evaluation protocols. In addition, we compare full fine-tuning with parameter-efficient fine-tuning methods, including LoRA and AdaLoRA, to assess their performance efficiency tradeoffs. Experimental results show that transformer-based models achieve strong segmentation performance, while parameter efficient finetuning reduces trainable parameters by up to 95% with comparable accuracy to full finetuning. We further analyze generalization under distribution shift by comparing validation and held-out test performance. |
| title | A Benchmark Study of Segmentation Models and Adaptation Strategies for Landslide Detection from Satellite Imagery |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2604.16663 |