| 摘要 |
<P>PROBLEM TO BE SOLVED: To accurately and relatively easily detect a person by extracting leg features regardless of whether the person is moving or remains stationary in a person detection device, a person detection method and a program. <P>SOLUTION: A person detection device comprises: a scanning unit that scans a scanning range in a plane almost perpendicular to the gravity direction to measure the distances from a reference position to objects in the scanning range, thereby outputting range data; an extraction unit that forms segments by sampling the range data at a predetermined cycle, and extracts leg features based on the segments; and a detection unit that detects a person based on motion characteristics and geometric information of the extracted features. The extraction unit includes a stationary feature strong classifier that, based on movement characteristics of a pair of two segments acquired by using the nearest neighborhood algorithm, classifies stationary segments in which each segment of the pair remains stationary, and a movement feature strong classifier that, based on the same, classifies movement segments in which each segment of the pair is moving, and extracts a leg pair from pairs of two segments based on features of the classified segments. <P>COPYRIGHT: (C)2012,JPO&INPIT |
