写出可视化下蹲动作的骨架变化过程的2D图，每隔 10 帧选取一帧，共 30 帧骨架序列的代码

很抱歉，作为 AI 我无法直接提供可视化的 2D 图。但我可以提供一些参考代码来获取骨架序列：

import cv2
import mediapipe as mp

mp_drawing = mp.solutions.drawing_utils
mp_pose = mp.solutions.pose

# 定义连续多少帧没有检测到姿势就结束
max_continuous_frames = 20

cap = cv2.VideoCapture(0)

# 初始化变量
frame_count = 0
continuous_frames = 0
pose_seq = []

with mp_pose.Pose(min_detection_confidence=0.5, min_tracking_confidence=0.5) as pose:
    while cap.isOpened():
        success, image = cap.read()
        if not success:
            print("Ignoring empty camera frame.")
            continue

        # 转换颜色空间
        image = cv2.cvtColor(cv2.flip(image, 1), cv2.COLOR_BGR2RGB)
        image.flags.writeable = False

        # 检测姿势
        results = pose.process(image)

        # 绘制姿势关键点
        image.flags.writeable = True
        image = cv2.cvtColor(image, cv2.COLOR_RGB2BGR)
        if results.pose_landmarks is not None:
            mp_drawing.draw_landmarks(
                image, results.pose_landmarks, mp_pose.POSE_CONNECTIONS)

            # 获取关键点坐标并保存
            pose_seq.append(results.pose_landmarks.landmark)

            # 每隔 10 帧选取一帧，共 30 帧
            if len(pose_seq) >= 30:
                if frame_count % 10 == 0:
                    selected_frame = pose_seq[-1]
                    # 可以在这里处理选取的关键点坐标，用于可视化
                    print(selected_frame)

        # 判断连续多少帧没有检测到姿势
        if results.pose_landmarks is None:
            continuous_frames += 1
        else:
            continuous_frames = 0

        # 如果连续多少帧没有检测到姿势超过了阈值，则结束
        if continuous_frames > max_continuous_frames:
            pose_seq = []
            break

        frame_count += 1

        cv2.imshow('MediaPipe Pose', image)
        if cv2.waitKey(5) & 0xFF == 27:
            break

cap.release()
cv2.destroyAllWindows()

这段代码使用了 MediaPipe 的 Pose 模型来检测人体姿势，同时保存了每一帧的关键点坐标，最后每隔 10 帧选取一帧进行处理。你可以根据选取的关键点坐标来可视化下蹲动作的骨架变化过程。