Analyzing Human's Motion Pattern Using Sensor Fusion in Complex Spatial Environments

摘要: We propose hybrid-sensing system for human tracking. This uses laser scanners and image sensors is applicable to wide crowded area such as hallway of university. Concretely, tracking using at base are used identification when lose persons by occlusion, entering room or going up stairs. developed the method this system. Our following: 1. Best-shot images (human which show feature clearly) obtained help position direction data scanners. 2. Human conducted calculating correlation between color histograms best-shot images. It becomes possible conduct even in scenes estimating In experiment station, some effectiveness became clear.