In recent years several implementations of position and movement realization via visual means have been developed. With the cost of RGB-D cameras decreasing and the availability of powerful hardware for processing images the area of visual odometry and VSLAM has become a heavily researched area. This project attempts to find an improved and cost effective environment for image processing to allow for collision avoidance and visual odometry. We attempt to create a multi-core embedded processing environment that will allow for an extremely fast and accurate estimation of position. This paper focuses on the key issue of improvement of visual odometry.
|出版ステータス||Published - 2013 1 1|
|イベント||2013 52nd Annual Conference of the Society of Instrument and Control Engineers of Japan, SICE 2013 - Nagoya, Japan|
継続期間: 2013 9 14 → 2013 9 17
|Other||2013 52nd Annual Conference of the Society of Instrument and Control Engineers of Japan, SICE 2013|
|Period||13/9/14 → 13/9/17|
ASJC Scopus subject areas
- コンピュータ サイエンスの応用