Building a DIY Robot Goal Keeper using Computer Vision

Key Concepts

• Computer Vision (CV): The field of AI that enables computers to interpret and process visual data from the world.
• OpenCV & CVZone: OpenCV is the core library for image processing; CVZone is a Python wrapper that simplifies complex OpenCV tasks like color detection and contour finding.
• HSV Color Space: A color model (Hue, Saturation, Value) used for robust color detection, which is more effective than RGB under varying lighting conditions.
• Contours: The boundaries of objects detected in an image, used here to locate the ball's coordinates.
• Serial Communication: The protocol used to send data from the Python environment to the Arduino microcontroller.
• Servo Motor: An actuator used to physically move the goalkeeper arm based on signals received from the computer.

---

1. Hardware and Software Setup

• Hardware: A camera (positioned to view the table), an Arduino, and a servo motor connected to pin 9.
• Software Environment: PyCharm is used as the IDE. The project requires CVZone and PySerial libraries.
• Camera Calibration: The camera must be positioned to capture the entire trajectory of the ball. A green mat is placed under the playing area to provide high contrast, making it easier to isolate the ball's color.

2. Color Detection Methodology

• Process: The ColorFinder class from CVZone is used.
• Debugging: A track bar is enabled to adjust HSV values in real-time until only the ball is highlighted in the mask.
• Implementation: Once the optimal HSV range is found, the track bar is disabled, and the values are hardcoded into the script for consistent detection.

3. Ball Tracking and Logic

• Contour Detection: The system uses findContours to identify the ball's position. The function returns the bounding box and center coordinates.
• Trajectory Estimation: A horizontal line is drawn across the frame at a specific vertical position (defined by a variable from 0 to 1).
• Decision Making:
  • If the ball crosses the line on the left side, the system sends a command to move the servo to the "left" position.
  • If it crosses on the right, it sends a command to move to the "right."
  • A "cool-down" period of 2 seconds is implemented to prevent erratic movement and allow the servo to reset to the center (90°) position.

4. Arduino Integration

• Communication: The Python script sends specific integers to the Arduino via serial:
  • 0: Center (90°)
  • 1: Right (0°)
  • 2: Left (180°)
• Troubleshooting: The author notes that serial communication errors (e.g., "Permission Denied") often occur if the Arduino Serial Monitor is left open while the Python script is running.

5. AI-Assisted Development

• Workflow: The developer uses an AI coding assistant (Cursor) to generate boilerplate code.
• Methodology: By providing the AI with the CVZone library examples and the specific requirements (e.g., "draw a tracking line," "reset after 2 seconds"), the AI writes the logic for tracking and serial communication, which the developer then refines.

6. Notable Quotes

• "The main idea is that the ball should be visible at the very start and we need to estimate which direction is it going."
• "If you have a much faster servo, then it will be pretty much impossible to have a goal."

---

Synthesis and Conclusion

The project demonstrates a practical application of computer vision and robotics. By combining HSV-based color filtering for object detection with contour analysis for spatial tracking, the system can predict the trajectory of a ball. The integration of PySerial allows the Python-based vision system to control physical hardware (the servo) in real-time.

Main Takeaways:

1. Modularity: Using libraries like CVZone significantly reduces the complexity of OpenCV implementation.
2. Debugging: Real-time visual feedback (track bars and masks) is essential for tuning computer vision parameters.
3. Future Scalability: The project can be made autonomous and portable by migrating the code from a PC to a Raspberry Pi or Jetson Nano, powered by a portable battery.