Related Experiment Video
Updated: May 28, 2026

Long-term Behavioral Tracking of Freely Swimming Weakly Electric Fish
Published on: March 6, 2014
Underwater Multi-Target Tracking and Behavioral Rhythm Analysis of Chinese Giant Salamander Based on
Nanqing Sun1,2, Xinyao Yang1,2, Mokai Xie1,2
1School of Intelligent Science and Engineering, Hubei Minzu University, Enshi 445000, China.
Abstract:
To address the challenges posed by traditional tracking algorithms in adapting to complex underwater environments characterized by nonlinear motion, drastic morphological changes, mimicry camouflage, and frequent occlusions in wild Chinese giant salamanders, this study proposes a multi-object tracking and behavior analysis method based on the TransTrack-OC-SORT algorithm. The algorithm employs a dual-branch Transformer motion predictor to replace linear Kalman filtering, effectively capturing nonlinear motion patterns such as velocity changes and directional turns exhibited by salamanders. Simultaneously, it introduces the BIOU matching metric, which integrates center distance and aspect ratio penalties with overlap degree, thereby enhancing the robustness of associations in scenarios involving occlusion and mimicry camouflage. The results indicate that the multi-object tracking accuracy (MOTA) based on the TransTrack-OC-SORT algorithm reaches 80.9%, while the identity preservation metric (IDF1) achieves 83.7%. Quantitative behavioral analysis based on continuous trajectory data obtained from this algorithm revealed significant diurnal behavioral rhythms in the Chinese giant salamander. During the daytime, stationary behavior accounted for 95.5% of the total behavioral duration, while swimming for ventilation and foraging behaviors constituted only 4.1% and 0.4%, respectively. At night, the salamander's behavioral patterns underwent fundamental changes; although stationary behavior remained dominant (approximately 72.6%), the proportions of swimming for ventilation and foraging behaviors significantly increased to 15.1% and 12.3% of the nocturnal period, respectively. These findings enhance our understanding of the salamander's ecological habits from the perspectives of visual adaptation, energy allocation, and predation strategies. This study provides a reliable technical tool for non-invasive behavioral monitoring and rhythm research in endangered amphibians.

