Multi-Regional landslide detection using combined unsupervised and supervised machine learning
| Autor: | Meylin Herrera Herrera, Giorgio Santinelli, Faraz S. Tehrani |
|---|---|
| Jazyk: | angličtina |
| Rok vydání: | 2021 |
| Předmět: |
landslide
Landslide detection 010504 meteorology & atmospheric sciences k-means 0211 other engineering and technologies 02 engineering and technology Machine learning computer.software_genre 01 natural sciences Environmental technology. Sanitary engineering GE1-350 change detection TD1-1066 021101 geological & geomatics engineering 0105 earth and related environmental sciences General Environmental Science business.industry k-means clustering Landslide Random forest Environmental sciences obia machine learning HD61 General Earth and Planetary Sciences Risk in industry. Risk management Artificial intelligence business computer Change detection Geology random forest |
| Zdroj: | Geomatics, Natural Hazards & Risk, Vol 12, Iss 1, Pp 1015-1038 (2021) |
| ISSN: | 1947-5713 1947-5705 |
| Popis: | Landslide detection is concerned with delineating the extent of landslides. Most of existing works on landslide detection have limited geographical extents. Therefore, the models developed out of these studies might perform poorly when applied to regions with different characteristics. This study investigates an Object-Based Image Analysis methodology built on unsupervised and supervised Machine Learning to detect the location of landslides occurred in multiple regions across the world. The utilized data includes Sentinel-2 multi-spectral satellite imagery and ALOS Digital Elevation Model. In the segmentation stage, pre and post-landslide images undergo segmentation using K-means clustering. Following the segmentation stage and dataset preparation and removing highly-correlated features from the dataset, two Random Forest classifiers (RF1 and RF2) are trained and tested on two different datasets to measure the generalization level of the algorithms with RF1 dataset spanning over more geographical diversities than RF2 dataset. The results show that the RF models can successfully detect landslide segments with test precision = 0.96 and recall = 0.96 for RF1 and test precision = 0.90 and recall = 0.87 for RF2. Further validation shows that, compared to RF2, RF1 results in less mislabelled non-landslide segments. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|