Evaluation of Various Deep Learning Algorithms for Landslide and Sinkhole Detection from UAV Imagery in a Semi-arid Environment
Bibliographic record
Abstract
Abstract Sinkholes and landslides occur due to soil collapse in different slope types, often triggered by heavy rainfall, presenting challenges in the semi-arid Golestan province, Iran. This study primarily focuses on the detection of these phenomena. Recent advancements in unmanned aerial vehicle (UAV) image acquisition and the incorporation of deep learning (DL) algorithms have enabled the creation of semi-automated methods for highly detailed soil landform detection across large areas. In this study, we explored the efficacy of six state-of-the-art deep learning segmentation algorithms—DeepLab-v3+, Link-Net, MA-Net, PSP-Net, ResU-Net, and SQ-Net—applied to UAV-derived datasets for mapping landslides and sinkholes. Our most promising outcomes demonstrated the successful mapping of landslides with an F1-Score of 0.95% and sinkholes with an F1-Score of 89% in a challenging environment. ResUNet exhibited an outstanding Precision of 0.97 and Recall of 0.92, culminating in the highest F1-Score of 0.95, indicating the best landslide detection model. MA-Net and SQ-Net resulted in the highest F1-Score for sinkhole detection. Our study underscores the significant impact of DL segmentation algorithm selection on the accuracy of landslide and sinkhole detection tasks. By leveraging DL segmentation algorithms, the accuracy of both landslide and sinkhole detection tasks can be significantly improved, promoting better hazard management and enhancing the safety of the affected areas.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame machine prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. The Gemma side is a direct model label for every work in the frame, read from the title-only record. The Codex side is a classifier learned from the 10,348 direct Codex labels and calibrated to design-weighted sample rates; fields without enough sample support carry no Codex call. Candidate is the union of the two sides; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels.
Distilled classifier scores by category (both heads)
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.001 | 0.003 |
| Meta-epidemiology (narrow) | 0.002 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.001 |
| Insufficient payload (model declined to judge) | 0.001 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one source (direct Gemma or distilled Codex), not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".