Cloud Classification Project

This project classifies cloud types from images using a TensorFlow Lite model and publishes the results to an MQTT broker. This project is used in conjuction with the CCSN dataset and an ALLSKY camera.

Table of Contents

• Cloud Classification Project
  • Table of Contents
  • Installation
  • Configuration
  • Usage
    • Run this periodically

Installation

1. Clone the repository:

   git clone https://github.com/Slayingripper/Cloud-Classification.git
   cd Cloud-Classification

2. Install the required Python packages:

   pip install -r requirements.txt
   

Configuration

Edit the config.ini file to set up the project:

MQTT Settings: Configure the MQTT broker.
    server: IP address of the MQTT server.
    port: Port number of the MQTT server.
    topic: MQTT topic to publish the predictions.

Model Settings: Path to the TensorFlow Lite model file.
    path: Path to the .tflite model file.

Image Settings: URL of the image to classify.
    url: URL of the image.    

Usage

Run the miniclass.py script to classify the cloud type from the image and publish the result to the MQTT broker:

    python miniclass.py

Run this periodically

To run the miniclass.py script periodically using cron, follow these steps:

Open the crontab editor:

crontab -e

Add a new cron job to run the script at your desired interval. For example, to run the script every hour, add the following line:

0 * * * * /usr/bin/python3 /path/to/your/project/miniclass.py

Make sure to replace /usr/bin/python3 with the path to your Python interpreter and /path/to/your/project/miniclass.py with the actual path to your script.

Save and close the crontab editor.