OpenClaw Color Line Patrol

1. Course Content

Course Overview

In modern warehouse logistics AGV (Automated Guided Vehicle) and inspection robots, visual line following (Visual Line Following) is an autonomous tracking solution with extremely low cost and extremely flexible deployment. The robot only needs colored tape (such as red, green, blue, yellow tracks) laid on the ground, and it can travel smoothly along the established route.

In computer vision, there are two core keys to achieving highly robust line following:

1. Why not use RGB, but must convert to HSV color space? Ordinary RGB color mode tightly mixes "color" with "lighting brightness". The same red tape, when illuminated strongly, will appear pinkish white, and in shadows will appear dark grayish black, with vastly different RGB values, easily causing the car to get lost and "derail". However, HSV space (H Hue, S Saturation, V Value) completely separates "color's own attribute (Hue)" from "ambient brightness (Value)"! Even when lights are turned on/off or passing through sunlight-shadow boundaries, the red hue value almost remains unchanged, allowing the robot to stably and precisely extract the color line contour.

2. High-frequency closed-loop and independent subprocess decoupling architecture: Line following requires the chassis to calculate deviation and correct front wheel steering angle at extremely high frequencies above 30Hz. The system cleverly uses an independent subprocess to drive the algorithm, ensuring smooth line-following driving while not blocking the MCP large model's monitoring of other commands.

Learning Objectives

1. Skilled in using colorSelect graphical tool to complete on-site environment HSV threshold pre-calibration.

2. Master the method of starting red, green, blue, yellow four-color automatic line following through OpenClaw dialogue using FollowLine tool.

3. Understand HSV color space anti-interference mechanism for target extraction under complex ambient light.

4. Master multi-process management and master-slave asynchronous communication design philosophy.

2. On-site Color Calibration (Pre-preparation)

Before running the line patrol program, it is recommended to complete the color tape HSV calibration based on the on-site ambient lighting. Although the device has default calibration parameters from the factory, if your classroom/lab lighting is warm-toned, or the ground material reflections cause inaccurate recognition, please use the dedicated graphical tool for quick calibration.

• Color HSV threshold configuration file path:

~/yahboomcar_ros2_ws/dofbot_pro_ws/src/ackermann_mcp/config/followlineHSV.text

2.1 On-site Environment Graphical Color Calibration (colorSelect)

Enter the following command in the terminal to start the graphical calibration tool:

xxxxxxxxxx
ros2 launch dofbot_pro_bringup colorSelect.launch.py

• After the program starts, a Color Selector interaction window will pop up on the screen:

• Drag and select an area of pure color that needs calibration in the window using the mouse (note: don't mix floor tile background colors into the selection):

• After selection, press the Space key on the keyboard, and the system will automatically calculate the maximum and minimum HSV range of the colors in the selection and persistently write to followlineHSV.text;

• After confirming the terminal prompt for successful saving, press Ctrl + C to exit the calibration program:

3. Preparation

After completing color calibration and exiting the calibration tool, start core nodes such as odometry, TF transformation, robotic arm assistance, camera node, and MCP service:

xxxxxxxxxx
ros2 launch dofbot_pro_bringup mcp_service.launch.py

• If you need voice replies, you can use the command with the launch_openclaw_bridge parameter, otherwise you can simply use the basic command above:

xxxxxxxxxx
ros2 launch dofbot_pro_bringup mcp_service.launch.py   launch_openclaw_bridge:=true 

• For voice dialogue interaction, you can run the following command (choose either the WebChat web interface or voice wake-up interaction):

xxxxxxxxxx
ros2 run ackermann_mcp asr_detect

After shouting the wake-up word Hi,yahboom, the indicator light turns on, and you can then speak the content you want to discuss.

4. Color Line Patrol Case Practice

Core calling tool:

Step 1: Issue Line Patrol Command

Enter the line patrol command in the WebChat interface, for example:

"Follow the red line" or "Start green line patrol"

Step 2: Hand-Eye Downward View and Real-time Debug Window

• The robotic arm quickly adjusts to the dedicated downward-looking posture, so the camera faces the road directly in front of the car;

• The line patrol monitoring window can optionally be enabled or disabled depending on your board's performance; the left side shows the original RGB color image, and the right side shows the black and white binarized image after HSV threshold filtering:

Image observation technique:
In the right binarization window, only the selected target color tape is displayed as pure white, all other background areas are pure black with no noise interference. The Ackermann chassis dynamically controls front wheel left-right steering based on the white area's geometric center deviation, guiding the car to travel stably as if on rails.

5. Core Source Code Analysis

• Source code file path:

xxxxxxxxxx
~/yahboomcar_ros2_ws/dofbot_pro_ws/src/ackermann_mcp/ackermann_mcp/action_controller.py

FollowLine() — Semantic Mapping and Independent Subprocess High-frequency Control

​x
def FollowLine(self, color: str) -> bool:
    '''Main entry for starting color line patrol task'''
    import subprocess
    # 1. Move robotic arm to preset observation posture, ensuring camera line of sight is vertically downward to car front ground
    self._pubSix_Arm(self.waste_detect_pose)

    self.follow_line_future = Future()  
    color = color.strip("'\"").lower()  
    
    # 2. Map natural language color string understood by large model to the underlying algorithm's agreed digital enum ID
    color_map = {'red': 1.0, 'green': 2.0, 'blue': 3.0, 'yellow': 4.0}
    if color not in color_map:
        error_msg = f"不支持的颜色: '{color}'. 可选颜色: red, green, blue, yellow"
        self.get_logger().error(error_msg)
        return [False, error_msg]
    target_color = color_map[color]

    # 3. Core design: Start line patrol algorithm node as independent subprocess (Subprocess)!
    self.follow_line_process = subprocess.Popen([
        'ros2', 'run', 'ackermann_mcp', 'follow_line',
        '--ros-args', '-p', f'color:={target_color}'
    ])
    
    # 4. Asynchronously listen for line patrol task end signal (triggered by chassis detecting end transverse line or manual interruption)
    while not self.follow_line_future.done():
        time.sleep(0.1)
    self.follow_line_clear_future = None
    
    # 5. After task ends, thoroughly clean up subprocess tree, and restore robotic arm initial posture
    kill_process_tree(self.follow_line_process.pid)
    self.InitArmPose()
    return [True, "Follow line finished"]

Line patrol source code:

xxxxxxxxxx
~\yahboomcar_ros2_ws\dofbot_pro_ws\src\ackermann_mcp\ackermann_mcp\external_action\follow_line.py
#usb camera
~\yahboomcar_ros2_ws\dofbot_pro_ws\src\ackermann_mcp\ackermann_mcp\vision\usb_backend.py

Find display_windows in the depth camera line patrol source code to decide whether to enable visualization:

xxxxxxxxxx
self.declare_parameter('colcor', 2.0)
self.colcor_type = int(self.get_parameter('colcor').get_parameter_value().double_value)
self.declare_parameter('display_windows', False)
self.display_windows = self.get_parameter('display_windows').get_parameter_value().bool_value
self.declare_parameter('max_angular', 0.35)
self.max_angular = self.get_parameter('max_angular').get_parameter_value().double_value

Find display_windows in the USB camera line patrol source code to choose whether to enable visualization.