Tracking Land Conversion with the Eye in the Sky
summary
Project uses Landsat 8 images (2013–2015) exported from Google Earth Engine, processed with rasterio/numpy/pandas and an Enhanced Vegetation Index input. Monthly samples (12 per year) produce 67 engineered features, morphology smoothing, and clustering with k-means; silhouette scores guide choosing ~11–14 clusters. Results identify center‑pivot cropland, distinguish river vegetation and land conversion in Nebraska and Florida, but an "unsure" class remains and the pipeline is not yet productionized.
AI-written summary · about
chapters (6)
- 0:00Introduction
- 1:47Land Conversion
- 10:23Identifying Cropland
- 11:07Identifying NonCropland
- 12:18Improving the Model
- 17:42Ground Truth
related talks
36mWatching Millions of Trees with Python: For Impact and Profit
36mGeoMonitoring with Python: Innovative Solutions for Latin America’s Challenges
21mActivity Map from space: supporting mine clearance with Python
30mCreating 3D Maps using Python
34mGeoscience Big Data Computing with Python
29mGeospatial Analysis using Python
40mData-centric exploration using intake, dask, hvplot, and more
02h 59mExploring Eco topics with Python