Tracking people and cars using 3D modeling and CCTV

Area of Science:

• Computer Vision
• Environmental Modeling
• Forensic Science

Background:

• Reconstructing real-world movements from surveillance footage is challenging.
• Integrating CCTV data with 3D environmental models can improve spatial understanding.

Purpose of the Study:

• To develop and validate a method for reconstructing pedestrian and vehicle movements using CCTV footage and 3D environmental models.
• To assess the effectiveness of the proposed method in enhancing situational awareness and tracking capabilities.

Main Methods:

• Synchronizing video streams and displaying them within a 3D model using virtual cameras.
• Representing and animating people and vehicles as geometric primitives (cylinders and boxes) within the 3D model.
• Validating the procedure with GPS ground truth data from test persons and applying it to a forensic case.

Main Results:

• The proposed procedure is feasible for reconstructing movements from video recordings.
• The 3D model significantly improved situational awareness and facilitated multi-stream tracking of individuals.
• The method proved effective in both experimental and forensic applications.

Conclusions:

• The developed method offers a viable solution for movement reconstruction from CCTV data.
• 3D environmental modeling enhances the analysis of surveillance footage, particularly for tracking.
• Recommendations are provided for practical implementation in real-world scenarios.