Automating offshore infrastructure extractions using synthetic aperture radar & Google Earth Engine
Why this work is in the frame
A frame that forgets how it found something cannot be audited. These are the routes that admitted this work.
Bibliographic record
Abstract
Although Land Use and Land Cover (LULC) change is primarily focused on the types, rates, causes, and consequences of land change, increased anthropogenic development on the ocean's surface, such as offshore oil extraction, offshore wind energy, aquaculture, and coral reef conversion to military outposts, suggest that LULC change not only pertains to historically terrestrial space, but also new lands created on top of ocean surfaces. Therefore, similar human disturbance analyses are necessary for these transformed marine environments, but the lack of accurate, accessible, and up-to-date location information about these spatially dispersed changes significantly limits examination of their environmental impacts. Subsequently these dynamic changes across the oceans are poorly documented. Here, we developed a cloud-native geoprocessing algorithm to automatically detect and extract offshore oil platforms in the Gulf of Mexico using synthetic aperture radar and Google Earth Engine. Cross-validated results indicate our top model identified offshore infrastructure with a probability of detection of 98.70%, an overall accuracy of 96.09%, a commission error rate of 2.68%, and an omission error rate of 1.30%. Its generalizability was tested across wind farms in waters of China and the United Kingdom, which resulted in an overall accuracy of 97.00%, a commission error rate of 2.07%, and omission error rate of 0.97%. These generalization capabilities indicate our model can be potentially used to map global offshore infrastructure. Such increased ocean transparency could allow for improved marine environmental management by bringing objectivity, scalability, and accessibility.
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.
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) | 0.002 | 0.001 |
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 it