Design of Multi Robot Sports Dance Motion Detection System Based on Improved Intelligent Algorithms

In view of the low accuracy of sports dance motion detection and difficulty in adapting to the situation of multiple robots, this paper combines the OpenPose algorithm and YOLOv8 (You Only Look Once version 8) to conduct research. First, the image is marked with regions and categories, and the YOLOv8 algorithm is used for target detection to identify the robot's position and bounding box. Then the OpenPose algorithm is used to detect human body key points, analyze the spatial position and motion trajectory of the key points, and identify the type of gymnastics dance movements. Experimental results show that the detection accuracy of the YOLOv8-OpenPose algorithm reaches 98.13%, which is 15.87% higher than that of the OpenPose algorithm. The results show its significantly high accuracy in multi-robot sports dance motion detection.