2026年4月9日 深度拆解AI跑步助手：技术原理与代码实战

 传统方案：拍摄 → 存储 → 人工逐帧查看
def traditional_posture_analysis(video_path):
    frames = load_video(video_path)       加载整段视频
    print("正在逐帧播放，请教练观察...")
    for frame in frames:
        display(frame)                    教练肉眼观察
        user_input = input("是否有问题？(y/n)")
        if user_input == "y":
            manual_note("第{}帧疑似跑姿问题".format(frame.id))
    print("分析完成，生成笔记")
     缺点：耗时、主观性强、无法量化、非实时

pip install mediapipe opencv-python numpy

import cv2
import mediapipe as mp
import numpy as np

 ========== 初始化 ==========
mp_pose = mp.solutions.pose
pose = mp_pose.Pose(
    static_image_mode=False,       视频流模式
    model_complexity=1,            模型复杂度 0/1/2
    enable_segmentation=False,
    min_detection_confidence=0.5,
    min_tracking_confidence=0.5
)
mp_drawing = mp.solutions.drawing_utils

 ========== 定义关键点索引 ==========
 MediaPipe Pose 33个关键点，跑步主要关注：
LEFT_SHOULDER = 11
RIGHT_SHOULDER = 12
LEFT_HIP = 23
RIGHT_HIP = 24
LEFT_KNEE = 25
RIGHT_KNEE = 26
LEFT_ANKLE = 27
RIGHT_ANKLE = 28

def calculate_angle(a, b, c):
    """计算三点之间的夹角（度）
    a, b, c 为 (x, y) 坐标，b 为顶点
    """
    a = np.array(a)
    b = np.array(b)
    c = np.array(c)
    ba = a - b
    bc = c - b
    cosine_angle = np.dot(ba, bc) / (np.linalg.norm(ba)  np.linalg.norm(bc))
    angle = np.arccos(np.clip(cosine_angle, -1.0, 1.0))
    return np.degrees(angle)

def analyze_running_pose(landmarks, frame_shape):
    """分析跑步姿态，返回关键指标和反馈建议"""
    h, w = frame_shape[:2]
    
     提取坐标（像素级别）
    left_hip = [landmarks[LEFT_HIP].x  w, landmarks[LEFT_HIP].y  h]
    left_knee = [landmarks[LEFT_KNEE].x  w, landmarks[LEFT_KNEE].y  h]
    left_ankle = [landmarks[LEFT_ANKLE].x  w, landmarks[LEFT_ANKLE].y  h]
    right_hip = [landmarks[RIGHT_HIP].x  w, landmarks[RIGHT_HIP].y  h]
    right_knee = [landmarks[RIGHT_KNEE].x  w, landmarks[RIGHT_KNEE].y  h]
    right_ankle = [landmarks[RIGHT_ANKLE].x  w, landmarks[RIGHT_ANKLE].y  h]
    
     计算膝关节角度（跑步时理想范围：160°~175°）
    left_knee_angle = calculate_angle(left_hip, left_knee, left_ankle)
    right_knee_angle = calculate_angle(right_hip, right_knee, right_ankle)
    
     计算躯干前倾角（理想：5°~10°）
    left_shoulder = [landmarks[LEFT_SHOULDER].x  w, landmarks[LEFT_SHOULDER].y  h]
    left_hip_angle = calculate_angle(left_shoulder, left_hip, left_knee)
    
     生成反馈建议
    feedback = []
    if left_knee_angle < 155 or right_knee_angle < 155:
        feedback.append("⚠️ 膝关节弯曲过大，着地瞬间应保持微屈")
    elif left_knee_angle > 180 or right_knee_angle > 180:
        feedback.append("⚠️ 膝关节过度伸直，增加冲击风险")
    
    if left_hip_angle < 5:
        feedback.append("⚠️ 躯干过度前倾")
    elif left_hip_angle > 15:
        feedback.append("⚠️ 躯干前倾不足，影响步幅")
    
    return {
        'left_knee_angle': round(left_knee_angle, 1),
        'right_knee_angle': round(right_knee_angle, 1),
        'trunk_angle': round(left_hip_angle, 1),
        'feedback': feedback if feedback else ["✅ 跑姿良好，继续保持"]
    }

 ========== 视频处理主循环 ==========
cap = cv2.VideoCapture(0)   打开摄像头

print("AI跑步助手已启动，按'q'退出")

while cap.isOpened():
    success, frame = cap.read()
    if not success:
        break
    
     转为RGB（MediaPipe要求）
    frame_rgb = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
    results = pose.process(frame_rgb)
    
     绘制姿态骨架
    if results.pose_landmarks:
         在原图上绘制骨架和关键点
        mp_drawing.draw_landmarks(
            frame, results.pose_landmarks, mp_pose.POSE_CONNECTIONS,
            mp_drawing.DrawingSpec(color=(0, 255, 0), thickness=2, circle_radius=2),
            mp_drawing.DrawingSpec(color=(0, 0, 255), thickness=2)
        )
        
         分析跑步姿态
        analysis = analyze_running_pose(results.pose_landmarks.landmark, frame.shape)
        
         在画面上叠加数据
        y_offset = 30
        cv2.putText(frame, f"Left Knee: {analysis['left_knee_angle']}°", 
                    (10, y_offset), cv2.FONT_HERSHEY_SIMPLEX, 0.6, (255, 255, 255), 2)
        cv2.putText(frame, f"Right Knee: {analysis['right_knee_angle']}°", 
                    (10, y_offset+30), cv2.FONT_HERSHEY_SIMPLEX, 0.6, (255, 255, 255), 2)
        cv2.putText(frame, f"Trunk: {analysis['trunk_angle']}°", 
                    (10, y_offset+60), cv2.FONT_HERSHEY_SIMPLEX, 0.6, (255, 255, 255), 2)
        
         反馈信息（红色高亮）
        for i, msg in enumerate(analysis['feedback']):
            cv2.putText(frame, msg, (10, frame.shape[0] - 80 + i30),
                        cv2.FONT_HERSHEY_SIMPLEX, 0.55, (0, 0, 255), 2)
    
    cv2.imshow('AI Running Assistant', frame)
    if cv2.waitKey(5) & 0xFF == ord('q'):
        break

cap.release()
cv2.destroyAllWindows()

摄像头帧 → RGB转换 → MediaPipe Pose推理 → 33个关键点坐标
                                              ↓
                              膝关节角度计算（三点夹角公式）
                                              ↓
                              与黄金阈值比较（160°~175°）
                                              ↓
                              生成视觉反馈 + 语音提示