Pose Detection

1. Experiment Objective

Use MediaPipe Pose to detect the 33 body pose keypoints in the camera image and visualize the human skeleton in real time.

2. Experiment Principle

1. Key Terms

2. Technical Architecture

Terminal 1: start agent + camera + joint models
Terminal 2: start pose detection

3. Experiment Steps

1. Hardware Preparation

2. Start the agent + camera + joint models (Terminal 1)

xxxxxxxxxx
ros2 launch microros_v2_bringup car_base.launch.py

3. Start pose detection (Terminal 2)

xxxxxxxxxx
ros2 launch microros_v2_mediapipe_pkg pose_detector.launch.py

4. Operating Instructions

Press q to exit.

5. How to Stop

Press Ctrl+C in each terminal, one by one, to stop.

4. Observed Results

• 33 keypoints are marked at the corresponding joints (including 11 face points), with colored skeleton connections overlaid in real time
• With smooth_landmarks=true, keypoints show no obvious jitter between frames

5. Code Walkthrough

Source code paths:

• Launch: ~/microros_ws/src/microros_v2_mediapipe_pkg/launch/pose_detector.launch.py
• Node: ~/microros_ws/src/microros_v2_mediapipe_pkg/microros_v2_mediapipe_pkg/PoseDetector.py

1. Core Node Analysis

​x
# Source file: ~/microros_ws/src/microros_v2_mediapipe_pkg/microros_v2_mediapipe_pkg/PoseDetector.py
class PoseDetector(Node):
    def __init__(self):
        super().__init__("pose_detector_node")
        self.declare_parameter("rgb_topic", "/camera/image_raw")
        self.declare_parameter("static_image_mode", False)
        self.declare_parameter("smooth_landmarks", True)
        self.declare_parameter("detection_con", 0.5)
        self.declare_parameter("track_con", 0.5)
​
        self.mp_pose = mp.solutions.pose
        self.pose = self.mp_pose.Pose(
            static_image_mode=static_image_mode,
            smooth_landmarks=smooth_landmarks,
            min_detection_confidence=detection_con,
            min_tracking_confidence=track_con)
​
    def image_callback(self, msg):
        frame = self.bridge.imgmsg_to_cv2(msg, 'bgr8')
        frame_rgb = cv.cvtColor(frame, cv.COLOR_BGR2RGB)
        results = self.pose.process(frame_rgb)
        if results.pose_landmarks:
            self.mp_draw.draw_landmarks(
                frame, results.pose_landmarks, self.mp_pose.POSE_CONNECTIONS)
        cv.imshow(self.window_name, frame)

Key logic:

• mp.solutions.pose.Pose(): 33-keypoint detector
• smooth_landmarks=True (default): enables cross-frame smoothing to reduce keypoint jitter
• static_image_mode=False: enables cross-frame tracking, reducing computation in video stream mode
• POSE_CONNECTIONS: MediaPipe's predefined human skeleton connections, automatically drawing the skeleton lines

2. Key Parameters

Parameter definition files:

• Launch: ~/microros_ws/src/microros_v2_mediapipe_pkg/launch/pose_detector.launch.py
• Node: ~/microros_ws/src/microros_v2_mediapipe_pkg/microros_v2_mediapipe_pkg/PoseDetector.py

3. Node Description

6. Frequently Asked Questions