Assessing the accuracy of georeferenced landcover data derived from oblique imagery using machine learning
Bibliographic record
Abstract
Abstract Repeat photography offers distinctive insights into ecological change, with ground‐based oblique photographs often predating early aerial images by decades. However, the oblique angle of the photographs presents challenges for extracting and analyzing ecological information using traditional remote sensing approaches. Several innovative methods have been developed for analyzing repeat photographs, but none offer a comprehensive end‐to‐end workflow incorporating image classification and georeferencing to produce quantifiable landcover data. In this paper, we provide an overview of two new tools, an automated deep learning classifier and intuitive georeferencing tool, and describe how they are used to derive landcover data from 19 images associated with the Mountain Legacy Project, a research team that works with the world's largest collection of systematic high‐resolution historic mountain photographs. We then combined these data to produce a contemporary landcover map for a study area in Jasper National Park, Canada. We assessed georeferencing accuracy by calculating the root‐mean‐square error and mean displacement for a subset of the images, which was 4.6 and 3.7 m, respectively. Overall classification accuracy of the landcover map produced from oblique images was 68%, which was comparable to landcover data produced from aerial imagery using a conventional classification method. The new workflow advances the use of repeat photographs for yielding quantitative landcover data. It has several advantages over existing methods including the ability to produce quick and consistent image classifications with little human input, and accurately georeference and combine these data to generate landcover maps for large areas.
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".