Pixel-based image classification to map vegetation communities using SPOT5 and Landsat5 Thematic Mapper data in a tropical savanna, northern Australia
Bibliographic record
Abstract
Pixel-based image classification has been used to capture various components of vegetation for a number of applications and at a range of spatial scales across the world. The few studies that have attempted to capture the floristic composition of vegetation communities in tropical savanna environments, at fine spatial scales (1:25000 or less) using these methods, have found minimal success. To address this gap, we evaluated a supervised image classification process using the Maximum Likelihood Classifier and 50% of a floristic and structural (strata, cover, height, and growth form) field dataset applied to SPOT5 and Landsat5 Thematic Mapper multispectral data. Two approaches were conducted to evaluate the influence of ancillary data on classification results: (i) “image-only” (image and field data) and (ii) “integrated” (various combinations of ancillary data with the image and field data). Multivariate analysis and intuitive classification were employed to identify 22 vegetation communities within the 530 km2 study area situated on Bullo River Station, Northern Territory, Australia. Class (vegetation community) separability averaged 1.94 and 1.42 for Landsat5 Thematic Mapper and SPOT5, respectively. A standard accuracy assessment was based on the remaining 50% of the field data. Overall accuracy ranged from 30%–53% for 1:25000 and 1:100000 spatial scale products. The inclusion of ancillary data was superior to the image and field data alone. The results of this study emphasize the need for finer spatial scale maps for property management planning (≤ 1:25000) and coarser scales for regional applications (≥ 1:100000).
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) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.001 | 0.001 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.001 | 0.001 |
| 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 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".