| 摘要 |
<P>PROBLEM TO BE SOLVED: To achieve object identification which does not depend upon the viewpoint direction of a camera while targeting camera images from a plurality of viewpoints without mounting a tag on an object. <P>SOLUTION: An object identification system 1 includes a feature point extraction part 22 which extracts a plurality of feature points of an object (an object for collation and a target object) from image data generated by photographing the object from the plurality of viewpoints; a feature point selection part 23 which selects a feature point whose geometric invariable can be calculated from the plurality of feature points; a geometric invariable calculation part 24 which calculates the geometric invariable from the position of the selected feature point; and an identification part 26 which identifies the target object by collating a geometric invariable as to the object for collation with the geometric invariable as to the target object. <P>COPYRIGHT: (C)2012,JPO&INPIT |
