04 / Geospatial machine learning
LULC Change Detection.
A geospatial ML system using U-Net segmentation and satellite raster datasets for land-cover classification and temporal change detection.
01 / OVERVIEW
The context.
The platform processes satellite raster data to classify land-cover categories and compare them over time. The workflow includes image preprocessing, segmentation, class-wise overlays and statistical analysis of observed changes.
02 / CONTRIBUTIONS
What I worked on.
- Developed a geospatial machine-learning workflow using U-Net segmentation models and satellite raster datasets.
- Implemented satellite-image preprocessing and land-cover classification.
- Built class-wise overlays and statistical change-analysis workflows using NumPy, OpenCV and raster-processing techniques.
03 / ENGINEERING NOTES
How it comes together.
01The project combines pixel-level segmentation with temporal comparison so that classified land-cover outputs can be examined across multiple time points.
02The emphasis is on the end-to-end geospatial workflow: preparing raster inputs, generating class predictions and analyzing changes.
04 / TECHNOLOGY