This project implements a live object detection. The windows environment setup and simulation demo are given below.Here are the steps we followed,

  1. Activate the created conda environment to work on that, conda activate object_det_env , Once activated, it will add(object_det_env) to the prefix.
  2. Install Tensorflow, conda install tensorflow, Note: If you are using GPU. Type 'tensorflow-gpu' instead of 'tensorflow'.
  3. Install Keras, conda install keras
  4. Install Matplotlib Library, conda install matplotlib
  5. Install Scikit Learn Library, conda install scikit_learn
  6. Install Opencv using Conda forge, conda install -c conda-forge opencv
  7. Install Pillow library for OpenCV,  conda install -c anaconda pillow
  8. Install OpenCV using Python-PIP. It can install all the dependencies for opencv, pip install opencv-python
  9. Check all the libraries are installed and importing properly.
$ python
$ import tensorflow
$ import keras
$ import matplotlib
$ import opencv
# import scikit-learn
  • Dataset: MS CoCo object detection dataset which can identify up to 60 objects. You can  download dataset from
  • SSD Mobilenet_v2_coco pre-trained model was used to improve the performance of object detection.
  • Tensorflow object detection API library was used for transfer learning. You can download the library from
  • The training was done on Google Colaboratory.
  • Raspberry Pi 3 B+ was used to run this model.

PC Requirements:
 1. Intel i5 6th gen or latest.
 2. At least 6GB ram.
 3. Minimum 500Gb hard disk (SSD is an advantage.)

Request source code for academic purpose, fill REQUEST FORM or contact +91 7904568456 by WhatsApp or, fee applicable.

SIMULATION VIDEO DEMO-Setting Up the Windows Environment