Video Analysis Using OpenCV in Python

Quick answer: Learn how to read, process and analyse video with OpenCV in Python, including frame extraction, motion detection, and object tracking basics.

Reading a Video File

import cv2

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

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

    cv2.imshow('Frame', frame)
    if cv2.waitKey(25) & 0xFF == ord('q'):
        break

cap.release()
cv2.destroyAllWindows()

A video is fundamentally a sequence of individual image frames. cap.read() returns each frame one at a time, along with a boolean indicating whether a frame was successfully read.

Reading from a Webcam

cap = cv2.VideoCapture(0)   # 0 is usually the default webcam

Extracting and Saving Individual Frames

frame_count = 0
while cap.isOpened():
    ret, frame = cap.read()
    if not ret:
        break
    cv2.imwrite(f'frame_{frame_count}.jpg', frame)
    frame_count += 1

Simple Motion Detection

import cv2

cap = cv2.VideoCapture('video.mp4')
ret, prev_frame = cap.read()
prev_gray = cv2.cvtColor(prev_frame, cv2.COLOR_BGR2GRAY)

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

    gray = cv2.cvtColor(frame, cv2.COLOR_BGR2GRAY)
    diff = cv2.absdiff(prev_gray, gray)
    _, thresh = cv2.threshold(diff, 25, 255, cv2.THRESH_BINARY)

    motion_detected = cv2.countNonZero(thresh) > 5000
    prev_gray = gray

This compares each frame to the previous one; a large enough difference between consecutive frames signals motion. Real production systems typically add noise filtering and a background model, but this captures the core idea.

Basic Object Tracking

tracker = cv2.TrackerCSRT_create()

ret, frame = cap.read()
bbox = cv2.selectROI(frame, False)   # manually select the object to track
tracker.init(frame, bbox)

while cap.isOpened():
    ret, frame = cap.read()
    if not ret:
        break
    success, bbox = tracker.update(frame)
    if success:
        x, y, w, h = [int(v) for v in bbox]
        cv2.rectangle(frame, (x, y), (x + w, y + h), (0, 255, 0), 2)

Tracking follows a specific object across frames after it is initially identified, which is generally faster than running full object detection on every single frame.

Practical Applications

• Security and surveillance motion alerts

• Traffic monitoring and vehicle counting

• Sports analytics, tracking players or a ball

• Manufacturing quality control on a production line

Common Interview Questions

How is a video fundamentally represented for processing?

As a sequence of individual image frames, read and processed one at a time, with the frame rate determining how many frames represent one second of video.

What is the difference between object detection and object tracking in video?

Detection identifies an object's location independently in every single frame. Tracking follows a specific, already-identified object across frames, which is generally faster since it does not need to search the entire frame from scratch each time.