Research of Volleyball Robot Based on YOLOv8 Lightweight and Binocular Vision

The study of volleyball robot capture mechanism is a challenging problem, and how to realize real-time and high-precision capture is the focus of current research. In order to solve these problems, this paper designs a capture system based on an improved YOLO model and binocular vision scheme for the study of volleyball robot capture mechanism. By improving faster as the backbone network and introducing the EMA attention mechanism, the YOLO model is realized to be lightweight, and the parameter computation is reduced by 26.5%, which ensures the real-time performance of the volleyball robot; meanwhile, the accuracy is improved by 1.1 %. On this basis, the improved YOLO is combined with an improved binocular vision scheme based on the principle of solving triangles to obtain the accurate world coordinate system of the volleyball relative to the disk, ensuring that the robot can solve the world coordinates of the volleyball in real time. In practical applications, the distance error within 2 meters is always kept within 1 cm, Through experimental verification, the design scheme in this paper can effectively improve the real-time and accuracy of the robot capture mechanism, and make an important contribution to the development of the robot capture mechanism.