Overhead Wildlife Locator (OWL): Benchmarking Weakly Supervised Learning for Aerial Wildlife Surveys

2026-06-11

AI Summary

This study addresses the high cost of bounding box annotation in aerial wildlife surveys by proposing OWL, a weakly supervised learning framework that establishes the first systematic benchmark for aerial animal detection using only point-level labels. OWL comprises three models tailored to varying animal density distributions: OWL-C based on a fully convolutional network, OWL-T leveraging a Swin Transformer, and OWL-D employing a frozen DINOv2 ViT-H/16 encoder with a DPT decoder. The work also introduces the first large-scale, open-access aerial reindeer patch dataset. Experiments demonstrate that OWL-D achieves an AP of 0.934 on the Delplanque dataset, outperforming HerdNet, and attains state-of-the-art results on four out of five public datasets. Furthermore, OWL-C achieves an F1 score of 0.965 in Alaskan reindeer census tasks with a counting error of merely +3.1%.

Abstract

Automated aerial wildlife surveys increasingly rely on deep learning, yet standard object detectors require bounding-box annotations, reported to be up to seven times slower and three times more expensive to produce than point-level labels. To address this bottleneck, we introduce the Overhead Wildlife Locator (OWL), a weakly supervised density-estimation framework with three variants: OWL-C, a fully convolutional model for high-throughput screening; OWL-T, a Swin-augmented hybrid for heterogeneous, cluttered scenes; and OWL-D, built on a frozen DINOv3 ViT-H+/16 encoder with a DPT-style fusion decoder. We benchmark all three against POLO, YOLOv11n, and YOLOv11l across five public aerial datasets, from sparse fixed-wing savanna surveys to dense UAV paddock imagery, and against the published HerdNet baseline on its native Delplanque split. OWL-D sets a new state of the art on Delplanque (0.934 AP vs. HerdNet's 0.840) and records the highest AP on four of the five datasets. Performance is regime-dependent: on the extreme-density SheepCounter UAV dataset the hybrid OWL-T leads (0.978 AP) and the convolutional variants attain the lowest counting error, whereas the foundation-based OWL-D degrades, indicating which variant suits which survey type. We further validate operational readiness on the Alaska Department of Fish and Game's 2022 Central Arctic Caribou census: under cross-herd and cross-temporal transfer, OWL-C fine-tuned on the 2017 Porcupine Caribou Herd split attains F1 = 0.965 on a held-out patch test set, with a signed count error of +3.1% aggregated across the released test patches. We release the OWL code, model weights, and the annotated Porcupine Caribou Herd 2017 (PCH) and Central Arctic Herd 2022 (CAH) patches, the first open patch-level datasets for large-scale caribou aerial surveys, at https://github.com/microsoft/MegaDetector-Overhead.

Problem

• weakly supervised learning
• aerial wildlife surveys
• object detection
• density estimation
• annotation cost

Innovation

• weakly supervised learning
• density estimation
• aerial wildlife survey
• foundation model
• object counting