yolo-detection-api

Description

Runs a Flask API using YOLO to detect objects in a frame. Use Docker to deploy where possible.

Notice these lines of code in api.py

USER_DATA = {
    os.environ['DETECT_API_USERNAME']: os.environ['DETECT_API_PASSWORD']
}

You need to set DETECT_API_USERNAME and DETECT_API_PASSWORD that matches your streaming client. You can add as many "users" as you would like.

Additionally, rename example.env to .env after filling out the contents.

Images - https://hub.docker.com/r/doorman/yoloapi/
sudo docker run --name redis-yolo -d redis
sudo docker run -d --link redis-yolo:redis -e REDIS_URL=redis://redis:6379/0 --volume "/home/pi/doorman/yolo-detection-api:/src/app" -p 5001:5001 doorman/yoloapi:rpi (replace rpi with the version you need from the available Docker images)

Download the model folder from here and put it the current directory (where this file lives).
sudo docker-compose up --build -d or for GPU sudo docker-compose -f gpu-compose.yml up --build -d
Navigate to http://localhost:5001

Download the model folder from here and put it the current directory (where this file lives).
Create a virtual environment with Python 3.6
Run pip install -r requirements.txt
Install darkflow

Run redis-server or set the REDIS_URL environmental variable with the Redis connection string.
(Optional) Set the THROTTLE_SECONDS environment variable to limit request calls to this server. The default will be 5 if not set. Use an integer.
Run python api.py
Navigate to http://localhost:5001