Webcam Motion Detector Using MOG2, KNN

Webcam Motion Detector is a complete Python application that uses a webcam and classical computer-vision techniques to detect moving objects in real time. The project combines OpenCV, Tkinter, background subtraction, contour analysis, morphological processing, and bounding-box visualization into a practical, ready-to-run desktop application.

The application captures live webcam video, builds an adaptive background model using MOG2 or KNN, extracts foreground regions, removes small noise, identifies motion contours, filters them based on object area, and displays bounding boxes around detected motion directly on the live video.

Key Features

• Real-time webcam motion detection
• MOG2 and KNN background-subtraction methods
• Adjustable minimum motion-object area
• Adjustable detection threshold
• Gaussian blur control for noise reduction
• Morphological image processing
• Foreground-mask visualization
• Real-time motion-object counting
• Motion bounding-box visualization
• FPS monitoring
• Camera index selection
• Start Camera control
• Stop Camera control
• Reset Detector for relearning the current background
• Clean and responsive Tkinter GUI
• Fully offline processing
• Windows-friendly webcam initialization
• No cloud API or internet connection required

How It Works

The project follows a straightforward computer-vision pipeline:

Webcam → Frame Processing → Background Subtraction → Foreground Mask → Morphological Cleanup → Contour Detection → Area Filtering → Bounding Boxes

The background-subtraction model continuously learns the scene and identifies regions that differ from the learned background. Morphological processing helps reduce noise and connect nearby foreground regions, while contour analysis determines the locations and sizes of detected motion areas.

Reset Detector

The Reset Detector function creates a fresh background-subtraction model and allows the application to relearn the current scene. This is useful when the camera is moved, lighting conditions change significantly, the scene is rearranged, or the detector begins producing excessive false motion.

Technologies Used

• Python
• Tkinter
• OpenCV
• Pillow
• MOG2
• KNN
• Contour-based image processing
• Morphological image processing

Suitable For

This source-code project is suitable for:

• B.Tech / BE computer-vision projects
• Python mini projects
• Computer-vision demonstrations
• Image-processing projects
• Final-year project prototypes
• Students learning OpenCV
• Academic demonstrations
• Researchers developing motion-analysis applications
• Developers looking for a foundation for surveillance or monitoring applications

Possible Extensions

The project can be further extended with motion-event snapshots, video recording, event logging, motion zones, multiple-camera support, alerts, CSV logging, motion heatmaps, and additional monitoring capabilities.

What You Receive

• Complete Python source code
• Tkinter graphical user interface
• OpenCV motion-detection implementation
• MOG2/KNN background subtraction
• Configuration controls
• Project documentation
• Requirements file
• Ready-to-run project structure

A practical, lightweight computer-vision project demonstrating real-time motion detection without requiring deep-learning models or custom datasets.