摘要 |
<P>PROBLEM TO BE SOLVED: To provide a vehicle detection device capable of appropriately grouping position data of a detected object to detect a vehicle such as a preceding vehicle. <P>SOLUTION: A vehicle detection device 1 includes: lamp candidate extraction means 12 for extracting a pixel region g that possibly corresponds to a tail lamp TL of a vehicle, as a lamp candidate g<SB POS="POST">L</SB>, among pixel regions g that each include pixels extracted from an image and integrated by integration processing means 11; and grouping means 14 for grouping position data detected by position detection means 9 to generate groups G, re-grouping groups G<SB POS="POST">L</SB>including the lamp candidate g<SB POS="POST">L</SB>, and then re-grouping all the groups G. During the re-grouping processing, a threshold Δdp at the time of re-grouping the groups G<SB POS="POST">L</SB>including the lamp candidate g<SB POS="POST">L</SB>is assumed to be a threshold that is easier to be re-grouped than a threshold Δdp<SP POS="POST">*</SP>at the time of subsequently re-grouping all the groups G. <P>COPYRIGHT: (C)2012,JPO&INPIT |