조밀 시차맵을 이용한 구조 정보 기반 차량 검출
- 조밀 시차맵을 이용한 구조 정보 기반 차량 검출
- Date Issued
- This study proposes an effective algorithm for detecting vehicles using dense disparity map. Generally ego-vehicle is in a lineup with driving or parked cars. Horizontal structure is a strong cue for hypothesizing vehicle presence. Height information of vehicle, obtained from a dense disparity map, can be used on vehicle validation. Dense disparity map, being removed ground plane components, is converted to 3D point cloud, and they are accumulated on an accumulation map. To extract candidates of vehicle, line segments are computed by using Probabilistic Hough Transform on the accumulation grid map. Vehicle model fitting and validation are performed using structural property of vehicles. We demonstrate the performance of the algorithm on Karlsruhe Institute of Technology (KIT) dataset and two kinds of LG driving dataset.
- Article Type
- Files in This Item:
- There are no files associated with this item.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.