1. Why Choose OpenCV and Python?¶

OpenCV is an open-source computer vision library that enables image and video processing. Python’s concise syntax is beginner-friendly, and OpenCV provides a dedicated Python interface (opencv-python), making installation and usage straightforward. With Python+OpenCV, you can easily implement real-time camera capture, image display, and basic image processing.

2. Installation Environment Setup¶

1. Install Python:
   Ensure Python (version 3.6 or higher) is installed on your computer. Download it from the Python official website.

2. Install OpenCV:
   Open the command line (Windows: cmd; Mac/Linux: Terminal) and run:

   pip install opencv-python

If installation fails (due to missing dependencies), install numpy first:

   pip install numpy

3. Basic Process for Real-Time Camera Capture and Display¶

To display camera feed in real-time, follow these steps:
1. Open the Camera: Create a “camera capture object” to connect to the hardware.
2. Read Frames Continuously: Continuously fetch image data (each frame) from the camera.
3. Display the Image: Show each frame in a window.
4. Release Resources: Close the window and release the camera device.

4. Complete Code Example with Step-by-Step Explanation¶

Here’s the full code for real-time camera display, with line-by-line breakdown:

# 1. Import the OpenCV library (abbreviated as cv2 in Python)
import cv2

# 2. Create a camera object; parameter 0 refers to the default camera
cap = cv2.VideoCapture(0)

# 3. Loop to capture images (True for infinite loop until 'q' is pressed)
while True:
    # 4. Read camera frame: ret (boolean, success status), frame (image data)
    ret, frame = cap.read()

    # Check if camera is accessible (ret=False means failure)
    if not ret:
        print("Failed to capture frame! Check device connection or permissions.")
        break  # Exit loop to avoid infinite loops

    # 5. Display the image: window name "Camera", show current frame
    cv2.imshow("Camera", frame)

    # 6. Wait for key input: 1ms delay; exit if 'q' is pressed
    key = cv2.waitKey(1)
    if key & 0xFF == ord('q'):  # Press 'q' to exit
        break

# 7. Release resources: close camera and all windows
cap.release()
cv2.destroyAllWindows()

5. Key Code Explanations¶

• cv2.VideoCapture(0):
  VideoCapture is a class for capturing video/camera input. Parameter 0 specifies the “default camera” (usually the built-in webcam). For external cameras, try 1 or 2.

• cap.read():
  Reads one frame from the camera, returning two values:

• ret: True if successful, False if the camera is disconnected or malfunctioning.

• frame: Image data (a 3D array: width × height × RGB/BGR channels).

• cv2.imshow("Camera", frame):
  Displays the image frame in a window named “Camera”. Window names can be customized (e.g., “My Camera”), but keep names concise.

• cv2.waitKey(1):
  Waits 1 millisecond to allow window updates. A 0 parameter causes the window to freeze; 1 ensures non-blocking operation and enables key-based control (e.g., pressing q to exit).

• cap.release() & cv2.destroyAllWindows():
  Release camera resources and close all display windows to prevent memory leaks.

6. Common Issues and Solutions¶

1. Installation Failures:
   - If “opencv-python not found”, update pip first:

     pip install --upgrade pip

• For Ubuntu/Linux missing dependencies:

     sudo apt-get install python3-dev python3-pip

2. Camera Won’t Open:
   - Check if the camera is used by another program (e.g., system camera apps).
   - Try changing the parameter (e.g., cap = cv2.VideoCapture(1)).
   - Permissions: Add sudo for Linux/Mac (e.g., sudo python3 demo.py).

3. Window Flashes and Disappears:
   - Forgetting cv2.waitKey(1) after imshow causes rapid window refresh. Always include the wait time.

7. Extension Exercises (Optional)¶

Modify the code to implement:
1. Grayscale Display: Convert color to black-and-white:

   gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
   cv2.imshow("Gray Camera", gray)

2. Flip Image: Horizontally flip the image:

   flipped = cv2.flip(frame, 1)
   cv2.imshow("Flipped Camera", flipped)

3. Save Image: Press s to save the current frame:

   if key & 0xFF == ord('s'):
       cv2.imwrite("saved_image.jpg", frame)

8. Conclusion¶

With Python+OpenCV, you can now capture and display real-time camera feed. The core logic is: Open camera → loop read → display frame → release resources. After mastering this, explore advanced tasks like face recognition or object detection!