Pole-Like Object Grasping Method for an AUV Equipped with a Monocular Camera and a Gripper

Autonomous Underwater Vehicles (AUVs) are in-creasingly used for underwater operations, including intervention tasks such as inspection and sample collection. This study presents a vision-based method for an AUV to grasp a vertically-oriented pole-like object using a monocular camera and a gripper. The proposed method extracts key features from camera images, estimating relative distance and angle of the target, those are then used in a PID-based controller for grasping. The approach has been validated through simulations and tank experiments, demonstrating successful grasping. The method is applicable to docking, perching, and sensor deployment. Future work will enhance image processing and control methods for real-sea operations, improving AUV intervention capabilities in complex underwater environments.