Skip to content

README.md

Latest commit

 

History

History
35 lines (18 loc) · 1.86 KB

README.md

File metadata and controls

35 lines (18 loc) · 1.86 KB

Satellite_Coverage_Classification

This repository contains all the codes and results of our Satellite Coverage Classification Project. In this project, we trained a ResNet Convolutional Neural Network (CNN) model on satellite images to classify different landscapes, such as deserts, water bodies, green areas, and cloudy scenes.

Contents

  • data.zip: This zip file folder contains the dataset used for this project. The dataset includes various types of satellite images, all separated into four distinct folders according to the landscape type they represent.

  • preprocessing_eda.ipynb: This Jupyter notebook contains all the preprocessing steps and the exploratory data analysis (EDA) done on the satellite images.

  • presentation.md: this markdown has the presentation, you can see it with vscode reveal in visual studio

  • resnet_cnn_model.ipynb: This notebook contains first the splitting of the data, the resnet model training, validation and test, the model's results and the performance of the model (confution matrix, classification report). It can be used for making predictions on new satellite images.

  • README.md: this file

Getting Started

  1. Clone the repository

  2. Install the required Python packages:

  3. Open and run the Preprocessing and EDA Jupyter notebook to understand how the data was preprocessed and explored.

  4. Open and run the resnet_cnn_model.ipynb Jupyter notebook to understand how the model was trained and evaluated.

Results

Our model achieved a validation accuracy of and test accuracy of 0.9396 and 0.9096 respectively.

Authors