SAM4CH4: Zero-Shot Methane Plume Mapping With Segment Anything and Vision-Language Models
Bibliographic record
Abstract
Recent advances in foundation models, including large language models (LLMs) and advanced computer vision techniques, have opened new possibilities in remote sensing applications. One such model is the Segment Anything Model (SAM), which can perform image segmentation without task-specific training data. This is especially useful for detecting methane plumes in satellite imagery, where it is important to accurately separate methane column enhancements from complex background conditions. SAM's prompt-based segmentation approach helps address these challenges and reduces the need for large annotated datasets. In this study, we introduce SAM4CH4, a zero-shot segmentation framework that applies SAM for methane plume detection using Sentinel-2 imagery, with segmentation prompts automatically generated by text encoder models including Contrastive Language-Image Pre-training (CLIP), CLIP Surgery, and Grounding DINO. We evaluate the approach on both a synthetically generated benchmark dataset and real Sentinel-2 images. Results show that bounding-box prompts from the Swin-L variant of Grounding DINO, combined with the latest version of SAM (SAM2), consistently achieve high accuracy—exceeding 72% in F1-score and 95% in overall accuracy—and outperform a widely used statistical thresholding method by approximately 15% in F1-score. These results are also competitive with supervised deep learning methods, which typically require large labeled datasets and significant computational resources. By leveraging pre-trained models and removing the need for manual annotation, the proposed SAM4CH4 framework offers a zero-shot scalable and efficient solution for operational methane plume detection and monitoring.
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 | 0.002 |
| Meta-epidemiology (narrow) | 0.002 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 0.002 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.001 | 0.002 |
| Open science | 0.004 | 0.002 |
| Research integrity | 0.002 | 0.002 |
| Insufficient payload (model declined to judge) | 0.006 | 0.003 |
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".