Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Find Rhinos without Finding Rhinos: Active Learning with Multimodal Imagery of South African Rhino Habitats

Lucia Gordon, Nikhil Behari, Samuel Collier, Elizabeth Bondi-Kelly, Jackson A. Killian, Catherine Ressijac, Peter Boucher, Andrew Davies, Milind Tambe

TL;DR

This work tackles rhino poaching by leveraging midden locations as a noninvasive proxy for rhino movements using multimodal remote sensing (thermal, RGB, LiDAR). It introduces MultimodAL, an active learning framework that ranks unlabeled images by an informative metric and ensembles across modalities to efficiently identify rare midden samples, dramatically reducing labeling needs. On a Kruger National Park site, passive midden detectors perform well, while MultimodAL matches the best passive performance with 94% fewer labels and discovers middens rapidly under class imbalance. The resulting midden map reveals clustering along animal paths, offering actionable guidance for targeted ranger patrols and conservation planning, with potential to scale to additional rhino habitats.

Abstract

Much of Earth's charismatic megafauna is endangered by human activities, particularly the rhino, which is at risk of extinction due to the poaching crisis in Africa. Monitoring rhinos' movement is crucial to their protection but has unfortunately proven difficult because rhinos are elusive. Therefore, instead of tracking rhinos, we propose the novel approach of mapping communal defecation sites, called middens, which give information about rhinos' spatial behavior valuable to anti-poaching, management, and reintroduction efforts. This paper provides the first-ever mapping of rhino midden locations by building classifiers to detect them using remotely sensed thermal, RGB, and LiDAR imagery in passive and active learning settings. As existing active learning methods perform poorly due to the extreme class imbalance in our dataset, we design MultimodAL, an active learning system employing a ranking technique and multimodality to achieve competitive performance with passive learning models with 94% fewer labels. Our methods could therefore save over 76 hours in labeling time when used on a similarly-sized dataset. Unexpectedly, our midden map reveals that rhino middens are not randomly distributed throughout the landscape; rather, they are clustered. Consequently, rangers should be targeted at areas with high midden densities to strengthen anti-poaching efforts, in line with UN Target 15.7.

Find Rhinos without Finding Rhinos: Active Learning with Multimodal Imagery of South African Rhino Habitats

TL;DR

This work tackles rhino poaching by leveraging midden locations as a noninvasive proxy for rhino movements using multimodal remote sensing (thermal, RGB, LiDAR). It introduces MultimodAL, an active learning framework that ranks unlabeled images by an informative metric and ensembles across modalities to efficiently identify rare midden samples, dramatically reducing labeling needs. On a Kruger National Park site, passive midden detectors perform well, while MultimodAL matches the best passive performance with 94% fewer labels and discovers middens rapidly under class imbalance. The resulting midden map reveals clustering along animal paths, offering actionable guidance for targeted ranger patrols and conservation planning, with potential to scale to additional rhino habitats.

Abstract

Much of Earth's charismatic megafauna is endangered by human activities, particularly the rhino, which is at risk of extinction due to the poaching crisis in Africa. Monitoring rhinos' movement is crucial to their protection but has unfortunately proven difficult because rhinos are elusive. Therefore, instead of tracking rhinos, we propose the novel approach of mapping communal defecation sites, called middens, which give information about rhinos' spatial behavior valuable to anti-poaching, management, and reintroduction efforts. This paper provides the first-ever mapping of rhino midden locations by building classifiers to detect them using remotely sensed thermal, RGB, and LiDAR imagery in passive and active learning settings. As existing active learning methods perform poorly due to the extreme class imbalance in our dataset, we design MultimodAL, an active learning system employing a ranking technique and multimodality to achieve competitive performance with passive learning models with 94% fewer labels. Our methods could therefore save over 76 hours in labeling time when used on a similarly-sized dataset. Unexpectedly, our midden map reveals that rhino middens are not randomly distributed throughout the landscape; rather, they are clustered. Consequently, rangers should be targeted at areas with high midden densities to strengthen anti-poaching efforts, in line with UN Target 15.7.
Paper Structure (14 sections, 4 equations, 9 figures, 2 tables)

This paper contains 14 sections, 4 equations, 9 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Related Work
  3. Setup
  4. Data
  5. Model
  6. Active Learning Methodology
  7. MultimodAL Algorithm
  8. Intuition for Ranking Idea
  9. Results
  10. Detecting Middens with Passive Learning
  11. Efficient Querying with Active Learning
  12. Midden Mapping
  13. Future Work
  14. Conclusion

Figures (9)

  • Figure 1: Left: A white rhino midden next to a road in iMfolozi whatisamidden. Right: A midden in our dataset photographed by a drone. Middens boxed in green.
  • Figure 2: Each pair shows the thermal (left) and RGB (right) images of the same area containing a midden. Green boxes outline middens. Red boxes outline areas that falsely appear to be middens. In (a), the midden is more obvious in the thermal image than in the RGB image, and in (b) the reverse is true.
  • Figure 3: Thermal (left), RGB (middle), and LiDAR (right) orthomosaics comprising the dataset under study.
  • Figure 4: Active learning cycle where the images are ranked by their brightness. (1) Predict on highest-ranked images. (2) Query images predicted to be positive. (3) Assign labels to queried images. (4) Add the newly labeled images to the set of all labeled images. (5) Train the model on a selection of the labeled images. (6) Restart.
  • Figure 5: Probability that an image contains a midden tends to increase with its maximum thermal pixel value.
  • ...and 4 more figures