Endangered Alert: A Field-Validated Self-Training Scheme for Detecting and Protecting Threatened Wildlife on Roads and Roadsides

Abstract

Traffic accidents, including animal-vehicle collisions (AVCs), endanger both humans and wildlife. This letter presents an innovative self-training methodology aimed at detecting rare animals, such as cassowaries in Australia, whose survival is threatened by road accidents. The proposed method addresses critical real-world challenges, including the acquisition and labelling of sensor data for rare animal species in resource-limited environments. It achieves this by leveraging cloud and edge computing, and automatic data labelling to improve the detection performance of the field-deployed model iteratively. Our approach introduces Label-Augmentation Non-Maximum Suppression (LA-NMS), which incorporates a vision-language model (VLM) to enable automated data labelling. During a five-month deployment, we confirmed the robustness and effectiveness of the method, achieving improved object detection accuracy and increased prediction confidence.