Example of use of our bio-recorders for the detection and recording of rare events. A) A biorecorder was used in this study. B) Bio-recorder attached to the back of a striated shearwater. C) The bird with the bio-logger is released and moves freely in a wild environment. The bio-logger automatically records videos of interesting events without monitoring by researchers. D) Example frames and sensor data collected by our bio-logger showing foraging behaviors of striated shearwaters. Credit: Nexus PNAS (2024). DOI: 10.1093/pnasnexus/pgad447
For centuries, naturalists have braved trailless forests, windy cliff tops, and the cramped confines of hides and submarines, hoping to capture rare behaviors that might reveal important aspects of biology and nature. animal ecology. Takuya Maekawa and his colleagues sought to deploy wearable trackers, which have become common in animal biology, to capture rare behaviors for study.
Since animal-mounted video recorders can only capture a few hours of video due to battery limitations, one of the main challenges is deciding when to record. The authors created an integrated AI program capable of performing “unsupervised learning” to automatically find and record rare behaviors, without the supervision of human naturalists.
The results are published in the journal Nexus PNAS.
First, an outlier detection program was trained on unlabeled accelerometer and water depth data from seabirds to automatically determine when unusual behavior occurs. This outlier detection program was used to create simplified outlier detectors (one for accelerometer data and one for water depth data) that fit a low-power microcontroller. power consumption on a logger with limited memory and computing power.
These detectors turn on a video camera on the recorder when rare behavior occurs in real time. The latest AI-enabled bio-recorder includes a video camera, a three-axis acceleration sensor, a GPS unit, a water pressure sensor, a thermometer, a magnetometer and an illuminometer, which were then affixed on a streaked shearwater (Calonectris leucomelas).
The biologger weighs 23 g, less than 5% of the weight of a shearwater. In field trials in 2022, the authors attached the biologgers to 18 birds.
Acceleration-based rare behavior detectors recorded videos of vigorous head jerks near the start of flight that the authors suggested could function to clear nasal salt gland fluids and other external materials in order to increase the efficiency of subsequent flight. Depth-based rare behavior detectors captured 50 minutes of active searching for fish, including preliminary glimpses underwater before diving, behavior rarely captured on camera.
According to the authors, AI-based biologgers can be used across a range of species to capture many types of rarely observed moments, including deep-sea mating rituals, hunting strategies used for prey rare and causes of death. of wild animals.
More information:
Kei Tanigaki et al, Automatic recording of rare wild animal behaviors using video bio-recorders with integrated lightweight outlier detector, Nexus PNAS (2024). DOI: 10.1093/pnasnexus/pgad447
Quote: AI-based bio-loggers capture behavior of rare birds (January 17, 2024) retrieved January 17, 2024 from
This document is subject to copyright. Except for fair use for private study or research purposes, no part may be reproduced without written permission. The content is provided for information only.
