Multispecies detection and identification of African mammals in aerial imagery using convolutional neural networks
Bibliographic record
Abstract
Abstract Survey and monitoring of wildlife populations are among the key elements in nature conservation. The use of unmanned aerial vehicles and light aircrafts as aerial image acquisition systems is growing, as they are cheaper alternatives to traditional census methods. However, the manual localization and identification of species within imagery can be time‐consuming and complex. Object detection algorithms, based on convolutional neural networks (CNNs), have shown a good capacity for animal detection. Nevertheless, most of the work has focused on binary detection cases (animal vs. background). The main objective of this study is to compare three recent detection algorithms to detect and identify African mammal species based on high‐resolution aerial images. We evaluated the performance of three multi‐class CNN algorithms: Faster‐RCNN, Libra‐RCNN and RetinaNet. Six species were targeted: topis ( Damaliscus lunatus jimela ), buffalos ( Syncerus caffer ), elephants ( Loxodonta africana ), kobs ( Kobus kob ), warthogs ( Phacochoerus africanus ) and waterbucks ( Kobus ellipsiprymnus ). The best model was then applied to a case study using an independent dataset. The best model was the Libra‐RCNN, with the best mean average precision (0.80 ± 0.02), the lowest degree of interspecies confusion (3.5 ± 1.4%) and the lowest false positive per true positive ratio (1.7 ± 0.2) on the test set. This model was able to detect and correctly identify 73% of all individuals (1115), find 43 individuals of species other than those targeted and detect 84 missed individuals on our independent UAV dataset, with an average processing speed of 12 s/image. This model showed better detection performance than previous studies dealing with similar habitats. It was able to differentiate six animal species in nadir aerial images. Although limitations were observed with warthog identification and individual detection in herds, this model can save time and can perform precise surveys in open savanna.
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".