情報処理学会 第76回全国大会講演要旨
5E-5
位置情報誤差にロバストなロケーションベース拡張現実による配電設備の巡視作業支援システム
A “robust location-based augmented reality system” for supporting
inspection of power distribution facilities is proposed, implemented,
and tested. By means of the proposed system, the facilities viewed
through a camera mounted on a tablet are identified by the
proposed “robust identification algorithm,” and related information
is overlaid on a live camera view. Since the algorithm only uses GPS
data as well as data from acceleration and geomagnetic sensors, and
complex vision-based methods are unnecessary, the facilities can be
identified robustly without being influenced by measurement errors of
the sensors. It identifies the facilities on the basis of not only
the device’s current location and heading obtained from these
sensors but also the object distance (i.e., the distance between the
user and the object being inspected) obtained by triangulation using
data from acceleration and geomagnetic sensors. Results of field
experiments with a prototype show that the algorithm can identify the
target facility uniquely at a success rate of 90%. As a result, the
system can improve efficiency and reliability of facility-inspection
works because the workers can get accurate information interactively.
inspection of power distribution facilities is proposed, implemented,
and tested. By means of the proposed system, the facilities viewed
through a camera mounted on a tablet are identified by the
proposed “robust identification algorithm,” and related information
is overlaid on a live camera view. Since the algorithm only uses GPS
data as well as data from acceleration and geomagnetic sensors, and
complex vision-based methods are unnecessary, the facilities can be
identified robustly without being influenced by measurement errors of
the sensors. It identifies the facilities on the basis of not only
the device’s current location and heading obtained from these
sensors but also the object distance (i.e., the distance between the
user and the object being inspected) obtained by triangulation using
data from acceleration and geomagnetic sensors. Results of field
experiments with a prototype show that the algorithm can identify the
target facility uniquely at a success rate of 90%. As a result, the
system can improve efficiency and reliability of facility-inspection
works because the workers can get accurate information interactively.
