Robot That Chases a Colored Ball with OpenCV

Color tracking is the friendliest way into computer vision. OpenCV converts each frame to HSV, keeps only pixels of your ball's color and finds their center; the robot steers to keep that center in the middle.

Mission steps

1. Build the base

   Use the obstacle robot wiring for motors and power, and mount the camera low and facing forward.

2. Install tools

   sudo apt install python3-opencv python3-picamera2 python3-gpiozero.

3. Pick the color

   Use a solid, bright ball (orange or lime green) and adjust the HSV range until only the ball shows white in the mask.

4. Steer to it

   If the ball is left of center, turn left; if centered, drive forward; if it fills much of the frame, stop.

5. Improve

   Add a search mode that slowly spins when the ball is lost, and smooth the steering to avoid jerky moves.

Code

import cv2, numpy as np
from picamera2 import Picamera2
from gpiozero import Robot, Motor, OutputDevice

OutputDevice(25).on()   # TB6612 STBY
bot = Robot(left=Motor(5, 6, enable=12), right=Motor(13, 19, enable=26))
cam = Picamera2()
cam.configure(cam.create_preview_configuration(main={'format': 'RGB888', 'size': (320, 240)}))
cam.start()
LOW, HIGH = np.array([5, 120, 120]), np.array([20, 255, 255])   # orange

while True:
    hsv = cv2.cvtColor(cam.capture_array(), cv2.COLOR_BGR2HSV)
    mask = cv2.inRange(hsv, LOW, HIGH)
    m = cv2.moments(mask)
    if m['m00'] < 50000:              # not enough pixels: search
        bot.right(0.3); continue
    x = m['m10'] / m['m00']
    area = m['m00'] / 255
    err = (x - 160) / 160             # -1 left ... +1 right
    if area > 15000:
        bot.stop()
    elif abs(err) > 0.25:
        bot.right(0.4) if err > 0 else bot.left(0.4)
    else:
        bot.forward(0.5)

Tip: test every script on the bench before installing it in a moving machine.

Questions