AI Recycling

This is the main project repo for the Automated AI-Recycling OSU Project.

Overview

This current release of the AI-Recycling project involves a video recognition model capable of detecting and counting objects in a video. The model is based on the YOLOv5 architecture and is trained on a custom dataset of recyclable objects. The model is capable of detecting and counting objects in a video, and the results are output to the screen. The model is also capable of running in real-time on a video stream.

Deployment Information

For deployment, the client will check with the MLFlow server for new weights for the custom dataset, download those, and convert them to an openVINO model for inference on integrated graphics. Then, it will run inference on a video or video stream and report object counts.

Prerequisites

Before starting, ensure you have the following installed on your system:

• Git
• 3.8 <= Python <= 3.11 (PyTorch does not work with newer versions of Python yet)
• Requirements as specificed in requirements.txt (pip install -r requirements.txt)

Installation

To get the project up and running on your local machine, follow these steps:

1. Clone the Repository

First, clone the repository to your local machine using Git:

bash git clone https://github.com/osu-ai-recycling/ai-recycling cd ai-recycling pip install -r requirements.txt

2. Configuration

Before running inference, you should check and adjust the model, source, and parameters according to your needs:

Model and Parameters: Inspect test_server.py to verify the model settings, source paths, and parameters such as debug_save, confidences, etc. Customized Parameters: Review detect.py for customized parameters, including paths where the results should be saved.

3. Running Inference

To run inference with the model, execute the following command:

bash python test_server.py video_path

This command will process the input data using the YOLOv5 model and output the results according to the configurations set in testserver.py and detect.py. This process will run the model on the input video and count objects detected in the video as well as output the results to the screen using OpenCV. The parameter videopath is the path to the input video.

Examples: bash python test_server.py recyclingvideo.mp4 python test_server.py rtsp://127.0.0.1:1234/stream