Water Body Extraction from Sentinel-2 Imagery with Deep Convolutional Networks and Pixelwise Category Transplantation
Bibliographic record
Abstract
A common task in land-cover classification is water body extraction, wherein each pixel in an image is labelled as either water or background. Water body detection is integral to the field of urban hydrology, with applications ranging from early flood warning to water resource management. Although traditional index-based methods such as the Normalized Difference Water Index (NDWI) and the Modified Normalized Difference Water Index (MNDWI) have been used to detect water bodies for decades, deep convolutional neural networks (DCNNs) have recently demonstrated promising results. However, training these networks requires access to large quantities of high-quality and accurately labelled data, which is often lacking in the field of remotely sensed imagery. Another challenge stems from the fact that the category of interest typically occupies only a small portion of an image and is thus grossly underrepresented in the data. We propose a novel approach to data augmentation—pixelwise category transplantation (PCT)—as a potential solution to both of these problems. Experimental results demonstrate PCT’s ability to improve performance on a variety of models and datasets, achieving an average improvement of 0.749 mean intersection over union (mIoU). Moreover, PCT enables us to outperform the previous high score achieved on the same dataset without introducing a new model architecture. We also explore the suitability of several state-of-the-art segmentation models and loss functions on the task of water body extraction. Finally, we address the shortcomings of previous works by assessing each model on RGB, NIR, and multispectral features to ascertain the relative advantages of each approach. In particular, we find a significant benefit to the inclusion of multispectral bands, with such methods outperforming visible-spectrum models by an average of 4.193 mIoU.
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 distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 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 teacher head, 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".