This project involves classifying satellite images based on the Normalized Difference Vegetation Index (NDVI) using various neural network architectures. The goal is to categorize different land covers using satellite imagery.
The project consists of a Python script that loads satellite image data, preprocesses it, trains different neural network models, and evaluates their performance.
- Satellite image loading and preprocessing for use in machine learning models
- Multiclass classification of satellite images using neural networks
- Two different neural network architectures: a fully-connected network and a 1D Convolutional Neural Network
- valuation of model performance using the accuracy metrics.
-Python: The primary programming language used for implementing the script -Keras: A high-level neural networks API, written in Python and capable of running on top of TensorFlow -TensorFlow: An open-source platform for machine learning -scikit-learn: Machine learning library in Python, used for data preprocessing and splitting the dataset -GDAL: A translator library for raster and vector geospatial data formats used to load satellite images
Accuracy of 95% was acheived