楕円体近似モデルを用いた複眼立体ビジュアルサーボによる対象物追従

抄録

Nowadays, measuring of farming fish is essential in checking fish’s growing rate in size. We have developed Automatically Fish Size Measuring System using Dual-eyes camera with the purpose of reducing time-consumed, labor and fish’s damage in manual measurement. By applying Dual-eyes visual servoing system based on this system to an underwater vehicle, we aim to construct a system of capturing underwater lives such as fishes after recognizing and tracing it automatically. Dual-eyes visual servoing is a method of controlling robot’s motion by utilizing 3D visual information of the environment as feedback signals, which is obtained by cameras. In this study, different models of fish were used in the experiments of tracking and capturing operations. Finally, the proposed system was verified to confirm the performance of tracking and capturing gold fishes alive in a pool.