Multi Class Satelite Image Classification

---

This is a project to create machine learning models capable of classifying satellite images belonging to 4 classes: cloudy, desert, green_area, and water

Models used

---

• EfficientNetB0: Transfer learning based model, retrained on provided dataset
• 2 sequential CNN models containing 3 convolution layers each

Building the project

---

Step 1

Download the file from either the website as a zip or using the following command in the command line interface

git clone https://github.com/agamsanghera/Multi-Class-Satellite-Image-Classification.git

Step 2

Unzip the zip file if downloaded from the website, if cloned repository, open the folder in any IDE or Jupyter with the following command in the shell in the directory where the project was downloaded

jupyter notebook

and run the notebooks as directed