ビジュアルオドメトリを用いた２次元自己位置推定検証のための屋外実験系構築と評価

抄録

Many of the public infrastructure facilities currently in use were constructed during the period of high economic growth and are rapidly becoming obsolete. On the other hand, the number of workers who inspect these infrastructure facilities has been decreasing in recent years. In particular, inspection work on underground infrastructure uses hand-held underground radars that are manually pushed to take measurements, which requires a great deal of labor and time. Therefore, the introduction of autonomous mobile robots is required from the perspective of improving inspection work efficiency and saving labor. High self-position estimation accuracy is required for autonomous movement and for identifying the location by comparing the coordinates of underground objects with information from underground radar. In this study, we adopted visual odometry using road surface video images as a self-localization method, and constructed and evaluated an outdoor experimental system to verify self-localization outdoors.