Adv-Python Unit 2: Complete Concept Guide

2.1 Video Files and Formats in OpenCVCopy Anchor Link

Concept Explanation:

A video is a collection of images (called frames) displayed rapidly one after another to create the illusion of motion.

Important components of a video:

• Frames: Individual images
• Frame Rate (FPS): Number of frames shown per second (common values: 24, 25, 30, 60)
• Resolution: Width × Height of each frame
• Codec: Algorithm used to compress and decompress the video
• Container Format: File type that holds the video data (.mp4, .avi, .mkv, etc.)

OpenCV does not handle audio properly. It mainly works with the video stream.

Common Formats & Codecs:

• .avi → Older but widely supported (codecs: XVID, MJPG)
• .mp4 → Most popular modern format (codecs: H.264 / avc1, mp4v)
• FourCC code is a 4-character code used by OpenCV to specify the codec while writing a video.

2.2 Extracting Frames from VideosCopy Anchor Link

Concept Explanation:

Extracting frames means saving individual images from a video.

This is useful for:

• Creating datasets
• Analyzing specific moments
• Applying image processing on selected frames

We can extract:

• All frames
• Every nth frame
• Frame at a particular time (in seconds)

2.3 Reading Videos from FileCopy Anchor Link

Concept Explanation:

To process a video, we first need to open it using cv2.VideoCapture().

Working pattern (very important for exams):

1. Create a VideoCapture object
2. Check if the video opened successfully
3. Run a loop to read frames one by one
4. Process each frame
5. Display or save the frame
6. Release the resources

cap.read() returns two values:

• ret → Boolean (True if frame is read successfully)
• frame → The actual image

2.4 Capturing Videos from CameraCopy Anchor Link

Concept Explanation:

Instead of a video file, we can capture live video from a webcam or external camera.

• cv2.VideoCapture(0) → Default webcam
• cv2.VideoCapture(1) → External camera

The rest of the process is almost the same as reading from a file.

Only difference: We usually use a very small delay in waitKey() (1 ms) for real-time feel.

2.5 Writing and Displaying VideosCopy Anchor Link

Concept Explanation:

Displaying: Use cv2.imshow() inside the loop.

Writing (Saving) a video requires:

1. Knowing the frame size (width, height)
2. Knowing the FPS
3. Selecting a codec (FourCC)
4. Creating a VideoWriter object
5. Writing each processed frame using out.write(frame)

(2.6, 2.7, 2.8) Drawing Lines, Rectangles, Circles and Adding Text (with Date & Time)Copy Anchor Link

Concept Explanation:

All drawing functions work exactly the same way as on images.

Key Point: Drawing must be done inside the while loop on every frame, otherwise it will appear only on one frame.

Common drawing functions:

• cv2.line()
• cv2.rectangle()
• cv2.circle()
• cv2.putText()

For live date and time, we use Python’s datetime module and update the text every frame.

2.9 Setting Video PropertiesCopy Anchor Link

Concept Explanation:

OpenCV allows us to get and set various properties of a video or camera using:

• cap.get(property_id)
• cap.set(property_id, value)

Important properties:

• CAP_PROP_FRAME_WIDTH
• CAP_PROP_FRAME_HEIGHT
• CAP_PROP_FPS
• CAP_PROP_FRAME_COUNT (total frames)
• CAP_PROP_POS_FRAMES (current frame number)
• CAP_PROP_POS_MSEC (current time in milliseconds)

Note: Not all properties can be changed on all cameras/videos.

2.10 Motion Detection using OpenCVCopy Anchor Link

Concept Explanation:

Motion detection means identifying moving objects in a video.

Basic Principle (Frame Differencing):

1. Take two consecutive frames
2. Find the difference between them
3. Convert difference to grayscale
4. Apply threshold to get a binary image
5. Find contours of white regions
6. Draw bounding boxes around large contours (moving objects)

Advanced methods (for knowledge):

• Background Subtraction (MOG2, KNN)
• Optical Flow

Exams usually expect the simple frame differencing method.

2.11 Object Tracking using OpenCVCopy Anchor Link

Concept Explanation:

Object Tracking means locating the same object across multiple frames after it has been initially selected.

Difference from Detection:

• Detection → Finding objects in every frame independently
• Tracking → Following an object once it is detected/selected

OpenCV provides several trackers (in opencv-contrib):

• CSRT → Most accurate
• KCF → Good balance of speed and accuracy
• MOSSE → Very fast
• MIL, TLD, MEDIANFLOW, etc.

Basic steps:

1. Select the object (ROI) in the first frame
2. Initialize the tracker
3. In every next frame, update the tracker
4. Draw the updated bounding box

Practical Code Examples (Concept → Code)Copy Anchor Link

1. Reading a Video FileCopy Anchor Link

import cv2

cap = cv2.VideoCapture('video.mp4')

if not cap.isOpened():
    print("Error opening video")
else:
    while True:
        ret, frame = cap.read()
        if not ret:
            break
        cv2.imshow('Video', frame)
        if cv2.waitKey(25) & 0xFF == ord('q'):
            break

cap.release()
cv2.destroyAllWindows()

2. Capturing from WebcamCopy Anchor Link

import cv2

cap = cv2.VideoCapture(0)

while True:
    ret, frame = cap.read()
    if not ret:
        break
    cv2.imshow('Webcam', frame)
    if cv2.waitKey(1) & 0xFF == ord('q'):
        break

cap.release()
cv2.destroyAllWindows()

3. Writing (Saving) a VideoCopy Anchor Link

import cv2

cap = cv2.VideoCapture(0)
width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
fps = 20

fourcc = cv2.VideoWriter_fourcc(*'XVID')
out = cv2.VideoWriter('output.avi', fourcc, fps, (width, height))

while True:
    ret, frame = cap.read()
    if not ret:
        break
    out.write(frame)
    cv2.imshow('Recording', frame)
    if cv2.waitKey(1) & 0xFF == ord('q'):
        break

cap.release()
out.release()
cv2.destroyAllWindows()

4. Drawing + Date & Time on VideoCopy Anchor Link

import cv2
import datetime

cap = cv2.VideoCapture(0)

while True:
    ret, frame = cap.read()
    if not ret:
        break

    # Drawing
    cv2.rectangle(frame, (100, 100), (300, 250), (0, 255, 0), 2)
    cv2.circle(frame, (400, 200), 40, (0, 0, 255), -1)
    cv2.line(frame, (50, 50), (250, 50), (255, 0, 0), 3)

    # Date and Time
    now = datetime.datetime.now().strftime("%Y-%m-%d %H:%M:%S")
    cv2.putText(frame, now, (10, 30), cv2.FONT_HERSHEY_SIMPLEX, 0.8, (0, 255, 255), 2)

    cv2.imshow('Annotated', frame)
    if cv2.waitKey(1) & 0xFF == ord('q'):
        break

cap.release()
cv2.destroyAllWindows()

5. Basic Motion DetectionCopy Anchor Link

import cv2

cap = cv2.VideoCapture(0)
ret, frame1 = cap.read()
ret, frame2 = cap.read()

while True:
    diff = cv2.absdiff(frame1, frame2)
    gray = cv2.cvtColor(diff, cv2.COLOR_BGR2GRAY)
    blur = cv2.GaussianBlur(gray, (5,5), 0)
    _, thresh = cv2.threshold(blur, 20, 255, cv2.THRESH_BINARY)
    dilated = cv2.dilate(thresh, None, iterations=3)

    contours, _ = cv2.findContours(dilated, cv2.RETR_TREE, cv2.CHAIN_APPROX_SIMPLE)

    for c in contours:
        if cv2.contourArea(c) < 900:
            continue
        x, y, w, h = cv2.boundingRect(c)
        cv2.rectangle(frame1, (x,y), (x+w, y+h), (0,255,0), 2)
        cv2.putText(frame1, "Motion", (10,20), cv2.FONT_HERSHEY_SIMPLEX, 0.7, (0,0,255), 2)

    cv2.imshow("Motion Detection", frame1)
    frame1 = frame2
    ret, frame2 = cap.read()

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

cap.release()
cv2.destroyAllWindows()

Exam Tips for Unit IICopy Anchor Link

• Always write the complete flow: open → check → loop → read → process → display → release
• Remember difference between waitKey(1) (camera) and waitKey(25) (video file)
• Motion detection = Frame Differencing + Threshold + Contours
• Tracking needs initialization with ROI first
• FourCC is compulsory when writing videos
• ret value is very important (theory + coding questions)