Search papers, labs, and topics across Lattice.
This paper addresses the problem of real-time crowd counting on railway platforms from a moving train-mounted camera by proposing a physics-constrained tracking framework. They integrate a YOLOv11m detector and EfficientNet-B0 appearance encoding within DeepSORT, and introduce a novel physics-constrained Kalman model (Phys-3D) to enforce physically plausible 3D motion based on pinhole geometry. The proposed method achieves a 2.97% counting error on their newly introduced MOT-RailwayPlatformCrowdHead Dataset (MOT-RPCH), demonstrating robustness to motion and occlusions.
Achieve accurate crowd counting from a moving train camera by explicitly modeling physically plausible 3D motion, outperforming methods that ignore these constraints.
Accurate, real-time crowd counting on railway platforms is essential for safety and capacity management. We propose to use a single camera mounted in a train, scanning the platform while arriving. While hardware constraints are simple, counting remains challenging due to dense occlusions, camera motion, and perspective distortions during train arrivals. Most existing tracking-by-detection approaches assume static cameras or ignore physical consistency in motion modeling, leading to unreliable counting under dynamic conditions. We propose a physics-constrained tracking framework that unifies detection, appearance, and 3D motion reasoning in a real-time pipeline. Our approach integrates a transfer-learned YOLOv11m detector with EfficientNet-B0 appearance encoding within DeepSORT, while introducing a physics-constrained Kalman model (Phys-3D) that enforces physically plausible 3D motion dynamics through pinhole geometry. To address counting brittleness under occlusions, we implement a virtual counting band with persistence. On our platform benchmark, MOT-RailwayPlatformCrowdHead Dataset(MOT-RPCH), our method reduces counting error to 2.97%, demonstrating robust performance despite motion and occlusions. Our results show that incorporating first-principles geometry and motion priors enables reliable crowd counting in safety-critical transportation scenarios, facilitating effective train scheduling and platform safety management.
