A project is proposed that detects and organizes garbage through artificial vision, which is composed of two parts of the project:
- Python files with trained artificial intelligence
- Server to interpret in react
- Matlab scripts to analyze training
Images
We realized that a very common problem is not knowing which container to put the garbage in, and often even though there are different bins for organic, plastics, paper, others, etc., people end up throwing waste in the wrong place.
Classifies garbage through an AI to correctly sort it, and with hardware it can then separate and deposit the waste in the correct place.
We trained a GoogleNet deep learning model for image recognition to distinguish between different types of garbage (organic, plastics, paper, others, etc.) and designed a prototype system for automated control with catFiles of what will be the proposed intelligent garbage bin.
First, we faced the problem of finding appropriate datasets to train the model because we couldn't find specifically what we needed for the problem. Complications linking technologies such as React with Flask and developing a backend using various programming languages (Python, JS, and Matlab).
-AI
-Catfiles
AI, deep learning, tensorflow, web design
Completely develop a prototype to be mass-produced.
To run the server you need to clone the repository and initialize the file manage.py and start the server.
npm startTo try the Python script, run the following commands:
pip install ./IA/requirements.txt
python ./IA/camera.pyOnce it starts, the camera will pop-up and you should press the spacebar with the garbage item you want to classify. Then another image will pop-up with the classification made by the model