ThruTracker

ThruTracker performs automated 2-D and 3-D video-based animal tracking to quantify counts and movement trajectories for studies in biomechanics, conservation, and ecology.

Key Features:

• Automated 2-D and 3-D workflows: Implements automated workflows for both 2-D and 3-D video tracking to streamline detection and tracking across frames.
• Thermal camera calibration: Performs intrinsic calibration of thermal cameras and supports 3-D calibration without dedicated calibration objects using scene structures such as wind turbines.
• Large-aggregation processing: Processes large datasets and tracks individuals within large aggregations, enabling analysis over medium spatial scales (~100,000 m³).
• Multi-platform videography support: Operates with thermal videography and accepts video data collected from unmanned aerial vehicles (UAVs).

Scientific Applications:

• Animal counting and movement analysis: Counts animals and documents movement trajectories for bats, birds, and fish that form large aggregations.
• 2-D case study — bat emergence: Applied to count bats emerging from bridges and caves using thermal videography, with performance comparable to human observers.
• 3-D case study — turbine monitoring: Applied to track bats and birds at wind turbines by using the turbine itself as a calibration object for 3-D reconstruction.

Methodology:

Integrates automated computational workflows for 2-D and 3-D tracking, performs intrinsic thermal camera calibration including scene-structure-based 3-D calibration, and processes video from thermal cameras and unmanned aerial vehicles.