Superpixel Sentinel-2 Dataset of Türkiye

Superpixel Dataset for Sentinel-2 Turkey Classification A comprehensive toolkit for superpixel-based land cover classification using Sentinel-2 satellite imagery from Turkey. This project implements SLIC (Simple Linear Iterative Clustering) segmentation combined with deep learning classification for automated land cover analysis. 🌟 Features - Automated Superpixel Segmentation: SLIC-based segmentation of satellite imagery- Deep Learning Classification: CNN-based classification of land cover types- Interactive Visualization: Tools for visualizing segments and classification results- Multi-region Support: Includes data from Bursa and Istanbul regions- Comprehensive Training Pipeline: Complete workflow from segmentation to model training- Pre-trained Models: Ready-to-use trained models for immediate deployment 📊 Dataset Overview Geographic Coverage - Bursa Region: 6 images covering Görükle, Karacabey, and Kestel districts (2015 & 2025)- Istanbul Region: 17 images from various locations (2025)- Total: 23 high-resolution Sentinel-2 satellite images Land Cover Classes The dataset includes 7 distinct land cover classes: 1. Tarla (Agricultural Fields) - Class 02. Su Alanı (Water Bodies) - Class 13. Yeşil Alan (Green Areas/Vegetation) - Class 24. Atanamaz (Unclassified) - Class 35. Yol (Roads) - Class 46. Çorak Toprak (Barren Land) - Class 57. Bina (Buildings/Urban) - Class 6