Skip to content

Geospatial & Remote Sensing Engineering Notes

Production-grade references for the geospatial Python engineering stack, coordinate reference system transformations, state estimation algorithms, and hyperspectral physics — the foundational layers of any large-scale Earth observation system.

Data Engineering & Python Stack

  • Geospatial Python Stack

    Canonical import patterns, optimised I/O routines, and production-tested code references for GeoPandas, Rasterio, Rioxarray, Xarray, and GDAL.

  • Coordinate Reference Systems & Projections

    Precision CRS transformations, EPSG codes, geodetic vs. projected systems, and float32 precision boundaries in large-scale raster workflows.

  • Production Geospatial Patterns

    Battle-tested implementation patterns for large-scale raster and vector workflows: chunked I/O, memory-safe processing loops, spatial joins at scale, and reproducible pipeline conventions.

  • Interactive Visualization with hvPlot

    High-level declarative visualizations for Xarray, Pandas, and GeoPandas datasets — browser-native interactivity with HoloViews and Bokeh backends for multi-dimensional geospatial data exploration.

State Estimation & Sensor Fusion

  • Kalman Filtering & State Estimation

    Mathematical derivation of the predict-update cycle, extended Kalman filters, and application to LiDAR-IMU sensor fusion for vehicle localization systems.

Machine Learning for Earth Observation

  • Geospatial PEFT

    Parameter-Efficient Fine-Tuning (LoRA) applied to geospatial vision transformers (Prithvi, Clay) for hydrological and cryospheric classification.

Sensor Physics & Imaging

  • Hyperspectral Sensor Physics

    Continuous narrow-band spectral arrays vs. discrete multispectral sensors — AVIRIS, PACE OCI, and the radiometric physics behind atmospheric correction and ocean colour retrieval.