Note

Hello, welcome to the SunFounder Raspberry Pi & Arduino & ESP32 Enthusiasts Community on Facebook! Dive deeper into Raspberry Pi, Arduino, and ESP32 with fellow enthusiasts.

Why Join?

  • Expert Support: Solve post-sale issues and technical challenges with help from our community and team.

  • Learn & Share: Exchange tips and tutorials to enhance your skills.

  • Exclusive Previews: Get early access to new product announcements and sneak peeks.

  • Special Discounts: Enjoy exclusive discounts on our newest products.

  • Festive Promotions and Giveaways: Take part in giveaways and holiday promotions.

πŸ‘‰ Ready to explore and create with us? Click [here] and join today!

4. Color Detection

Color detection is one of the most fundamental and practical functions in computer vision. In this chapter, we will use step-by-step code and explanations to detect red objects using the HSV color space and draw bounding boxes around them.

This forms the foundation for more advanced object tracking techniques (e.g., CAMShift).

1. Objective and Approach

  • Use Picamera2 to capture real-time camera frames

  • Convert the image from BGR to HSV color space

  • Use cv2.inRange to extract the red regions

  • Use morphological filtering to remove noise

  • Use cv2.findContours to find red object contours

  • Draw bounding boxes around the detected red regions

Color detection preview illustration

2. Run the Code

Important

Before you start, make sure:

  • The pan-tilt is assembled

  • You can access the Raspberry Pi desktop

  • The code package is installed

  • Fusion HAT+ is installed and configured

  • OpenCV is installed

For detailed instructions, see 0. Setup OpenCV.

  1. Open the terminal and enter the following command:

    cd ~/ai-lab-kit/opencv_python
python3 cv_4_color.py

  2. When you run the program, two OpenCV windows will appear on the screen:

    • Red Detection – shows the live camera image with green bounding boxes around detected red objects

    • Red Mask – shows the binary mask image used for red color detection

    The program continuously captures frames from the Raspberry Pi camera and detects red regions in real time. If a red object is detected, a green rectangle and the area value will be displayed on the color image.

    You can exit the program in two ways:

    • Press the q key on the keyboard

    • Close any of the OpenCV windows by clicking the close button (X)

    After exiting, the camera stops streaming and all OpenCV windows are closed.

3. Complete Code

from picamera2 import Picamera2
import cv2
import numpy as np
import time

# -----------------------------
# Camera setup
# -----------------------------
picam2 = Picamera2()
config = picam2.create_preview_configuration(
   main={"size": (640, 480), "format": "XRGB8888"}  # 4-channel format (BGRA-like)
)
picam2.configure(config)
picam2.start()

print("Streaming... press 'q' to quit")

# -----------------------------
# Red color range in HSV
# (Red wraps around 0/180 in HSV, so we use two ranges)
# -----------------------------
LOWER_RED1 = np.array([0,   100, 80], dtype=np.uint8)
UPPER_RED1 = np.array([10,  255, 255], dtype=np.uint8)
LOWER_RED2 = np.array([170, 100, 80], dtype=np.uint8)
UPPER_RED2 = np.array([180, 255, 255], dtype=np.uint8)

# Morphology settings
KERNEL = cv2.getStructuringElement(cv2.MORPH_ELLIPSE, (5, 5))
MIN_AREA = 800  # ignore small blobs

# Window names
WIN_RESULT = "Red Detection"
WIN_MASK = "Red Mask"

# Optional: limit FPS to reduce CPU usage (set to None to disable)
TARGET_FPS = 30
FRAME_INTERVAL = 1.0 / TARGET_FPS if TARGET_FPS else 0

while True:
   loop_start = time.perf_counter()

   # Capture one frame (BGRA-like) and convert to BGR for OpenCV processing
   frame_bgra = picam2.capture_array()
   frame_bgr = cv2.cvtColor(frame_bgra, cv2.COLOR_BGRA2BGR)

   # Convert BGR to HSV
   hsv = cv2.cvtColor(frame_bgr, cv2.COLOR_BGR2HSV)

   # Create red mask using two HSV ranges
   mask1 = cv2.inRange(hsv, LOWER_RED1, UPPER_RED1)
   mask2 = cv2.inRange(hsv, LOWER_RED2, UPPER_RED2)
   mask = cv2.bitwise_or(mask1, mask2)

   # Morphological operations: remove noise + fill holes
   mask = cv2.morphologyEx(mask, cv2.MORPH_OPEN, KERNEL, iterations=1)
   mask = cv2.morphologyEx(mask, cv2.MORPH_CLOSE, KERNEL, iterations=2)

   # Find contours in the mask
   contours, _ = cv2.findContours(mask, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)

   # Draw bounding boxes for valid red regions
   for cnt in contours:
      area = cv2.contourArea(cnt)
      if area < MIN_AREA:
            continue

      x, y, w, h = cv2.boundingRect(cnt)
      cv2.rectangle(frame_bgr, (x, y), (x + w, y + h), (0, 255, 0), 2)
      cv2.putText(
            frame_bgr,
            f"red area={int(area)}",
            (x, max(0, y - 6)),
            cv2.FONT_HERSHEY_SIMPLEX,
            0.5,
            (0, 255, 0),
            1,
            cv2.LINE_AA
      )

   # Show both windows
   cv2.imshow(WIN_RESULT, frame_bgr)
   cv2.imshow(WIN_MASK, mask)

   # Process GUI events + keyboard input
   key = cv2.waitKey(1) & 0xFF
   if key == ord("q"):
      break

   # Exit if the user closes any window (click X)
   if (cv2.getWindowProperty(WIN_RESULT, cv2.WND_PROP_VISIBLE) < 1 or
      cv2.getWindowProperty(WIN_MASK, cv2.WND_PROP_VISIBLE) < 1):
      break

# Cleanup
picam2.stop()
cv2.destroyAllWindows()

4. Code Explanation

  1. Initialize Picamera2 and start streaming:

    picam2 = Picamera2()
config = picam2.create_preview_configuration(
    main={"size": (640, 480), "format": "XRGB8888"}
)
picam2.configure(config)
picam2.start()

    This configures the camera at 640Γ—480 and starts the preview stream. XRGB8888 is a 4-channel format, so the captured frames are BGRA-like.

  2. Convert the captured frame to a format OpenCV commonly uses:

    frame_bgra = picam2.capture_array()
frame_bgr = cv2.cvtColor(frame_bgra, cv2.COLOR_BGRA2BGR)

    Picamera2 returns a 4-channel image here, so we convert it to standard 3-channel BGR for processing.

  3. Use HSV color space for robust color detection:

    hsv = cv2.cvtColor(frame_bgr, cv2.COLOR_BGR2HSV)

    HSV separates color (Hue) from brightness, which makes color detection more stable under different lighting.

  4. Define two HSV ranges for red:

    mask1 = cv2.inRange(hsv, LOWER_RED1, UPPER_RED1)
mask2 = cv2.inRange(hsv, LOWER_RED2, UPPER_RED2)
mask = cv2.bitwise_or(mask1, mask2)

    Red β€œwraps around” the Hue scale in OpenCV HSV (near 0 and near 180), so two ranges are combined to cover all reds.

  5. Clean the mask with morphology (reduce noise and fill holes):

    mask = cv2.morphologyEx(mask, cv2.MORPH_OPEN, KERNEL, iterations=1)
mask = cv2.morphologyEx(mask, cv2.MORPH_CLOSE, KERNEL, iterations=2)

    • OPEN removes small noisy dots.

    • CLOSE fills small holes inside the detected red regions.

  6. Find red regions and filter small blobs:

    contours, _ = cv2.findContours(mask, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)

for cnt in contours:
    area = cv2.contourArea(cnt)
    if area < MIN_AREA:
        continue

    Contours are detected from the binary mask. MIN_AREA ignores small red regions to reduce false detections.

  7. Draw bounding boxes and labels on the result image:

    x, y, w, h = cv2.boundingRect(cnt)
cv2.rectangle(frame_bgr, (x, y), (x + w, y + h), (0, 255, 0), 2)
cv2.putText(frame_bgr, f"red area={int(area)}", ...)

    This shows where OpenCV found red objects, and prints the detected blob area for reference.

  8. Display both the result and the mask:

    cv2.imshow(WIN_RESULT, frame_bgr)
cv2.imshow(WIN_MASK, mask)

    The result window shows the camera view with boxes, and the mask window shows the red-only binary image.

  9. Exit conditions (keyboard + window close):

    key = cv2.waitKey(1) & 0xFF
if key == ord("q"):
    break

if (cv2.getWindowProperty(WIN_RESULT, cv2.WND_PROP_VISIBLE) < 1 or
    cv2.getWindowProperty(WIN_MASK, cv2.WND_PROP_VISIBLE) < 1):
    break

    Press q to quit, or close either window to exit safely.

  10. Cleanup:

    picam2.stop()
cv2.destroyAllWindows()

    Always stop the camera and close OpenCV windows to release resources.

5. Parameter Tuning Tips

  • LOWER_RED1 / UPPER_RED1: adjust this range to detect other colors. For example, green β‰ˆ [35, 50, 50] to [85, 255, 255].

  • KERNEL: larger kernels give stronger filtering but may remove small objects.

  • MIN_AREA: increasing this value filters out small noisy contours; decreasing it makes detection more sensitive.

Note

You can start by only displaying the mask and tuning the thresholds until the target region looks clear, then proceed with the rest of the pipeline.

6. Extensions and Practice

  • Modify the HSV threshold to detect other colors (e.g., blue or green).

  • Experiment with different morphological parameters in more complex backgrounds.