Improvement of Tracking Accuracy Using Travel Pattern for Agent-Based Human Tracking
Kozo Tanigawa, Masaru Shiozuka, Keníchi Takahashi · 2024
Security systems are widely employed in various facilities, such as companies and schools, to prevent crime. However, as the monitored area increases, it becomes challenging for operators to track numerous targets effectively due to limited resources. To address this issue, we propose an agent-based human tracking system. When utilizing image processing for person detection, the inherent inaccuracies in image processing make it challenging to accurately track the target. This often results in detection errors and tracking failures. Therefore, we propose the calculation of the existence probability. The existence probability can be calculated by combining the traveling pattern of the targets with the result of image processing. Simulation results indicate that the calculation of the existence probability can achieve effective and accurate tracking.