Deep learning for urban land use category classification: A review and experimental assessment
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
Mapping the distribution, pattern, and composition of urban land use categories plays a valuable role in understanding urban environmental dynamics and facilitating sustainable development. Decades of effort in land use mapping have accumulated a series of mapping approaches and land use products. New trends characterized by open big data and advanced artificial intelligence, especially deep learning, offer unprecedented opportunities for mapping land use patterns from regional to global scales. Combined with large amounts of geospatial big data, deep learning has the potential to promote land use mapping to higher levels of scale, accuracy, efficiency, and automation. Here, we comprehensively review the advances in deep learning based urban land use mapping research and practices from the aspects of data sources, classification units, and mapping approaches. More specifically, delving into different settings on deep learning-based land use mapping, we design eight experiments in Shenzhen, China to investigate their impacts on mapping performance in terms of data, sample, and model. For each investigated setting, we provide quantitative evaluations of the discussed approaches to inform more convincing comparisons. Based on the historical retrospection and experimental evaluation, we identify the prevailing limitations and challenges of urban land use classification and suggest prospective directions that could further facilitate the exploitation of deep learning techniques in urban land use mapping using remote sensing and other spatial data across various scales.
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) | 0.001 | 0.001 |
| Meta-epidemiology (broad) | 0.001 | 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 it