| 摘要 |
In a target recognition apparatus, a candidate detection section detects a target candidate, provided that a target exists in a basic detection area. A candidate addition section adds, regarding each target candidate detected in a folding area, a target candidate determined provided that the target candidate detected in the folding area is a virtual image, and a corresponding real image exists in an additional detection area. A tracking section determines, regarding each detected and added target candidate, presence/absence of a history connection with the target candidate detected in a past measurement cycle. A combination determination section determines, regarding each target candidate, presence/absence of a combination with an image target, based on whether or not an image target associated with the target candidate exists. A likelihood calculation section sets and updates a likelihood of a virtual image of the image target by using a determination result of the combination determination section. |
