3D scene reconstruction from body-worn camera video using 3DF Zephyr.

Area of Science:

• Forensic Science
• Photogrammetry
• 3D Reconstruction

Background:

• Body-worn cameras (BWCs) are increasingly adopted by law enforcement globally.
• Current BWC use focuses on qualitative documentation, not quantitative forensic measurement.
• Empirical validation of BWC video for 3D forensic reconstruction is lacking.

Purpose of the Study:

• To empirically evaluate the accuracy of 3D reconstruction from BWC video.
• To assess the feasibility of using BWC footage for precise feature measurement in forensic applications.
• To compare the performance of different BWC models and resolutions for 3D reconstruction.

Main Methods:

• Three Axon BWC models (AB2, AF2, AB3) recorded videos at 720P and 1080P resolutions under controlled conditions.
• 3D reconstructions were generated using 3DF Zephyr photogrammetry software.
• Reconstruction accuracy was validated against Faro Focus S350 laser scanner ground truth at multiple distances.

Main Results:

• Axon Body 2 (AB2) and Axon Body 3 (AB3) models showed no significant differences from ground truth measurements across all tested distances and resolutions.
• Axon Flex 2 (AF2) exhibited significant measurement errors, particularly at longer distances and 720P resolution.
• No significant differences in accuracy were found between 720P and 1080P resolutions for the tested models.

Conclusions:

• Accurate 3D reconstruction from BWC video is feasible for forensic applications when using specific camera models (AB2, AB3) and deliberate documentation protocols.
• The study highlights the potential of BWCs for quantitative forensic analysis beyond qualitative documentation.
• Camera model selection is critical for achieving reliable 3D reconstruction accuracy from BWC footage.