Object Detection Using OpenCV

Get dependencies

Download the pre-trained YOLO weight file (237 MB): https://pjreddie.com/media/files/yolov3.weights

Install Open-CV

pip install opencv-python

Run object detection code

cd opencv
python cv-detection.py
python pic-detection.py

Object Detection Using Tensorflow

Get Tensorflow Object detection API

Clone models from Tensorflow Model Garden:

git clone https://github.com/tensorflow/models

Install dependencies

Follow the instructions to install Anaconda if it's not installed on your device yet:

https://docs.anaconda.com/anaconda/install/

TensorFlow or TensorFlow-GPU if you have a GPU on your device

pip install tensorflow

pip install tensorflow-gpu

Open-CV:

pip install opencv-python

Some other dependencies:

pip install jupyter lxml matplotlib pillow Cython tf_slim contextlib2 

Download and extract Protobuf (mechanism for serializing structured data) to models/research folder:

https://github.com/protocolbuffers/protobuf/releases

Compile protocol buffer description files

Generate Python code from a .proto file

cd models/research
protoc --python_out=. object_detection/protos/*.proto
ls object_detection/protos/*.proto

Telegram bot integration

If you want to sent notifications to Telegram, you have to create a bot and provide its credentials in the corresponding function in the code. To create a Telegram bot you need to use another Telegram bot, BotFather. Just talk to him (press start): https://telegram.me/botfather

Run object detection code

cp tensorflow/*-detection.py models/research
cd models/research
python object-detection.py
python image-detection.py