This repository contains code for a deep learning project that performs semantic segmentation on satellite imagery using U-Net architecture. The project aims to accurately segment different classes in satellite images.
The project involves:
- Preprocessing and analyzing the satellite imagery data.
- Implementing U-Net architecture for semantic segmentation.
- Training and evaluating the model on various aerial images using various metrics.
- Visualizing segmentation results and model performance.
main.py: Python script containing the main model training and evaluation code.project_overview.ipynb: Jupyter notebook that provides an overview of the project, data exploration, and visualizations.requirements.txt: File listing the required libraries for the project.Src/: Directory containing source code files, including model definitions and utility functions.plots/: Directory for storing generated plots and visualizations.
The dataset is available on Kaggle: Semantic Segmentation of Aerial Imagery.
The dataset consists of satellite images obtained from Mohammed Bin Rashid Space Center in Dubai by MBRSC satellites and annotated with pixel-wise semantic segmentation in 6 classes (Building, Land, Road, Vegetation, Water and Unlabeled) that require semantic segmentation. Each image is labeled for training purposes to enable accurate segmentation by the model.
To run the code, ensure you have Python 3 installed along with the required libraries listed in requirements.txt. You can install the dependencies using pip:
pip install -r requirements.txt