作者: Tao Hu , Sinan Mutlu , Oswald Lanz
DOI: 10.1007/978-3-642-41184-7_70
关键词:
摘要: We propose a novel people detection method using Locus-based Probabilistic Occupancy Map (LPOM). Given the calibration data and motion edges extracted from all views, is able to compute probabilistic occupancy map for targets in scene. integrate algorithm into Bayesian-based tracker do experiments with challenging video sequences. Experimental results demonstrate robustness high-precision of when tracking multiple presence clutters occlusions.