Raspberry-Object-Detection

Upcomings

• Guideline to run this project.
• Abstract Code WalkThrough
• Basic Learning Resources
  • Raspberry Pi
  • OpenCV
• Fun Project for Contributors. Reference Materials
  • A blog from OpenCV
  • A video from sentdex (pythonprogramming.net)

Guidelines

• This repository contains python script for the object detection on Raspberry Pi in real time using OpenCV.
• It uses a already trained MobileNet Architecture stored as Caffe Model.
• This model uses Single Shot Detection(SSD) algorithm for prediction.
• Look for the architecture detail here
• This code stores the input images in a queue and output the predictions along with box in queue.
• This code runs on real time videostream to output the prediction as well as bounding box across it.
• Above codes can be directly run on raspberry pi having packages imutils and cv2
• To Install, run

  pip install imutils
  pip install opencv-python
  

• USAGE, To run the videostream, clone the repo and in working directory, run

  python real_time_object_detection.py --prototxt MobileNetSSD_deploy.prototxt.txt --model MobileNetSSD_deploy.caffemodel
  

  Note : Press ‘q’ to stop the videostream.

Abstract Code WalkThrough :

Overviewing real_time_object_detection.py

• Importing packages

• Creating the argument parser for the python script

• Initializing the list of class labels MobileNet SSD was trained to detect, then generate a set of bounding box colors for each class.

• Load the model and initialize the video stream.

• For each frame, obtain the detections and predictions.

• For each detection, if prediction is above confidence (self-defined), draw the bounding box over it.

• Shows the output frame with break condition key “q”. On Break condition, it destroys all the windows.

Learning Resources

For Learning basics of Raspberry PI

For Learning basics of OpenCV

Fun Project for Contributors

Face Detection using Haar Cascades

Refer to these tutorials for details.

Refer to this cool video of sentdex