Back to Home
All projects AI Projects · Deep learning for land use/cover change analysis

Satellite Change Detection Pipeline

Automated satellite image change detection pipeline using U-Net architecture for monitoring urban expansion and deforestation.

Built an end-to-end pipeline for bi-temporal satellite image analysis using deep learning. The system processes Sentinel-2 imagery to detect and classify land use/cover changes with pixel-level accuracy.

Highlights

Manual change detection from satellite imagery is time-consuming and inconsistent across analysts.
Implemented a modified U-Net architecture trained on labeled Sentinel-2 image pairs, with automated preprocessing and post-classification refinement.

Related projects