This repository contains scripts for real-time object detection and interaction using both MATLAB and Python, demonstrating different approaches and tools for computer vision tasks.
The MATLAB script utilizes MATLAB's Computer Vision Toolbox to perform the following tasks:
- Foreground Detection: Utilizes Gaussian mixture models (vision.ForegroundDetector) to detect foreground objects in a video stream.
- Blob Analysis: Uses vision.BlobAnalysis to identify connected components (blobs) based on the detected foreground.
- Visualization: Displays the original video frames, the extracted foreground, and the cleaned foreground (after morphological operations).
- Object Detection: Draws bounding boxes around detected objects and counts them, displaying the results in real-time using imshow and vision.VideoPlayer.
The Python script employs OpenCV for video processing and interaction via Tkinter and email notifications:
- Object Detection with Caffe Model: Uses cv2.dnn module to perform object detection using a pre-trained MobileNet SSD model (MobileNetSSD_deploy.caffemodel).
- User Interaction: Utilizes Tkinter for displaying message boxes to prompt user actions based on detected objects (e.g., sending email notifications with or without object images).
- Email Notifications: Sends email alerts with optional image attachments using smtplib and email modules upon detecting specific objects (e.g., 'bottle').
- Efficient Video Processing: Implements video capture and processing efficiently using VideoStream from imutils.video.
- Both scripts demonstrate real-time object detection capabilities, allowing users to interact based on detected objects.
- MATLAB script emphasizes ease of use with built-in toolbox functions for foreground detection and blob analysis.
- Python script showcases integration of OpenCV for complex computer vision tasks and external interactions such as email notifications.