运动检测数据集

该数据集专注于健身动作人体运动分析，基于视频数据，使用MediaPipe 姿态估计提取关键关节地标，并结合YOLOv6 目标检测提升精度，逐帧计算俯卧撑、开合跳、引体向上、深蹲、俄罗斯扭转五种标准健身动作的关键关节角度，形成结构化运动角度时序数据，用于健身姿态分析与智能训练系统研发

This dataset focuses on human motion analysis for fitness movements. Based on video data, it uses MediaPipe Pose Estimation to extract key joint landmarks, and combines YOLOv6 object detection to improve detection accuracy. It calculates the key joint angles of five standard fitness movements, namely push-ups, jumping jacks, pull-ups, squats and Russian twists, frame by frame to generate structured temporal motion angle data, which is intended for the research and development of fitness posture analysis and intelligent training systems.

搜集汇总

背景与挑战

背景概述

该数据集专注于健身动作的人体运动分析，基于视频数据，通过MediaPipe姿态估计和YOLOv6目标检测技术，提取并计算五种标准健身动作的关键关节角度，形成结构化时序数据。它适用于健身动作矫正、运动表现追踪和姿态分类，可用于智能健身助手和实时姿态反馈系统的开发。