I am a PhD in Biological Sciences (Ecology), a forest ecosystem researcher, and a developer of geospatial tools in Python.
My work focuses on remote sensing of forests: UAV and satellite imagery, LiDAR and point clouds, forest-stand characteristics, and carbon-stock assessment. I use statistical methods and machine learning, including PyTorch and Mask R-CNN, to analyze spatial data, segment tree crowns, and combine remote-sensing products with field measurements.
I develop open-source projects at the intersection of geoinformatics, remote sensing, data analysis, and scientific programming.
- forest ecology and remote sensing;
- UAV data and satellite imagery;
- LiDAR and point-cloud processing;
- statistical analysis of spatial and ecological data;
- tree-crown segmentation and Mask R-CNN;
- GIS automation and spatial analysis;
- research software development in Python.
| Project | Purpose |
|---|---|
| priorityclip-geo | Assigns polygon overlaps by numeric priority. Accepts Polygon/MultiPolygon vector layers such as GeoPackage or GeoJSON, and writes a non-overlapping GeoPackage layer plus a CSV area audit. |
| lazproof-geo | Verifies that a LAS/LAZ output is the exact, order-preserving spatial subset selected from its source, including header, CRS, schema, and point-record checks with a JSON report. |
| track2corridor-geo | Builds an auditable centerline and corridor from noisy ground mobile-mapping trajectory points, with inspectable GeoPackage layers and JSON diagnostics. |
| concavewrap-geo | Builds auditable concave hulls for Polygon/MultiPolygon layers while verifying complete-area coverage, with GeoPackage output and JSON diagnostics. |
More geospatial utilities are being prepared for independent public releases.