An evaluation of high-resolution land cover and land use classification accuracy by thematic, spatial, and algorithm parameters
Bibliographic record
Abstract
High resolution land cover and land use classifications have applications in many fields of study such as land use and cover change, carbon storage measurements and environmental impact assessments. The wide range of available imagery at different spatial resolutions, potential thematic classes, and classification methods introduces the problem of understanding how each aspect affects accuracy. This study investigates how these three aspects affect the results of land cover classification. Results show that the maximum likelihood classifier was able to produce the most consistent results with the highest average accuracy (82.9%). Classifiers were able to identify a spatial resolution for each thematic resolution that achieved a distinctly higher overall accuracy. In addition, the effects of different land cover classifications as input to an object-based classification of land use at the parcel scale were evaluated. Results showed that land use classification requires higher resolution imagery to obtain satisfactory results than what is required for land cover classification. Also, the highest accuracy land cover classification did not produce the highest accuracy for land use, where a higher number of thematic classes performs better than fewer thematic classes. The highest accuracy LC classification by MLC with 8 classes occurred at 640 cm and achieved an overall accuracy of 83.3%. The highest accuracy LU classification was produced by the 80 cm LC with 8 classes and achieved an overall accuracy of 88.0%. Aside from the produced land cover and land use classifications, this study produces a lookup table in the form of multiple graphs for future research to reference when selecting imagery and determining thematic classes and classification methods.
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 | 0.012 | 0.033 |
| Meta-epidemiology (narrow) | 0.001 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.001 |
| Bibliometrics | 0.002 | 0.002 |
| Science and technology studies | 0.001 | 0.001 |
| Scholarly communication | 0.002 | 0.002 |
| Open science | 0.001 | 0.001 |
| Research integrity | 0.001 | 0.000 |
| Insufficient payload (model declined to judge) | 0.001 | 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 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".