Real-time object detection application

Web application for real-time object detection on video streaming via web browser.

Download model

To download the yolov3.weights, just run:

$ cd models/
$ python dl-weights.py

To run application (without using Docker)

Create and activate an virtual environment, as follows:

$ cd cloned/directory/
$ python3 -m venv env
$ source env/bin/activate

After have installed and activated the environment, install all the dependencies:

$ pip3 install -r requirements.txt

After that, you can run the following command and access the application at 127.0.0.1:5000 on your browser.

$ python application.py

To run application (using Docker)

Make sure you have Docker engine installed on your system before proceeding.

Go to project directory:

$ cd cloned/directory/

In the project directory, build Docker image using Dockerfile of the project:

$ docker build -t web_app:latest .

After successful build of Docker image, you now can create and run a Docker container of the image:

$ docker run --privileged -p 5001:5001 web_app

Now you should be able to access the application at 0.0.0.0:5001 port from your browser

To stop the application, first get CONTAINER_ID of the running Docker container:

$ docker ps

After getting the CONTAINER_ID, use the following command to stop the container:

$ docker stop <CONTAINER_ID>

obs.: This application was tested only on Google Chrome.

Usage

• Camera preview: Enables and disables the webcam preview.
• Run detection model: Enables and disables the detection model.
• Exposure: Buttons which increase or decrease camera exposure stops by 1.
• Contrast: Buttons which increase or decrease camera contrast stops by 4.
• Reset camera: Reset all camera settings based on camera_settings.log file.