Human-Robot Interactive Operating System for Underwater Manipulators Based on Hand Gesture Recognition.

Manipulators are the primary tool for robotic operations, and human-robot interaction for manipulators is an effective way in complex environments or dealing with complex operational tasks. This paper focuses on developing a human-robot interactive operating system for underwater ECA ARM 5E-Micro manipulator based on hand gesture recognition. The system used an RGB-D camera to acquire the operator's hand gestures and a hand image segmentation method was proposed based on color images combined with depth information. A convex hull algorithm and image moments were used to extract hand contour features, and then hand gesture was recognized by using Hu invariant moment features and KNN classification algorithm. The experiment results showed that the human-robot interactive operating system for underwater manipulators based on hand gesture recognition has a good hand gesture recognition rate and interaction effectiveness. This study provides a solution for implementing human-robot interaction for underwater robots.