CV

WebGIS Developer

Institute of Geography and Geology, University of Greifswald

• Developing the WebGIS platform of the German-Polish cultural heritage project In terris Gryphi (in-terris-gryphi.de): multilingual (DE/PL/EN) WordPress site with Leaflet map
• Processed and published 4 historical scanned map series (1692-1909, about 420 GB of scans) as web services using ArcGIS Experience Builder on ArcGIS Enterprise. Automated Python pipelines for reprojection, mosaicking, format conversion, footprints, and metadata linking (rioxarray, rasterio, GeoPandas)
• Built a Europe-wide OpenStreetMap basemap (styled in QGIS, converted to ArcGIS Pro, published to ArcGIS Online as a tile layer) and extracted 5,737 heritage objects (castles, churches, manors, viewpoints) for Pomerania with the OSMnx Python package

Remote Sensing & GIS analyst

Institute of Geography and Geology, University of Greifswald

• Co-designed a Sentinel-1/2 data fusion framework for 2022 crop type classification in smallholder agriculture across 3 districts in Kazakhstan, Tajikistan, and Pakistan (770 surveyed fields, 4-6 crop classes per site, incl. maize, wheat, cotton, fodder, rice, sorghum, sugarcane, fallow). Random Forest models on monthly to quarterly time-series statistics of Sentinel-2 bands, vegetation indices (NDVI, NDWI, EVI, SAVI), and Sentinel-1 VV/VH backscatter (120-960 features) with forward feature selection. Achieved 84.7-96.3% overall accuracy in 1000-run repeated 5-fold cross-validation (publication)
• Classified wetland vegetation species for 2020-2022 from UAV multispectral orthomosaics, DEMs, and NDVI with Random Forest (7 classes incl. Typha, Phragmites, Phalaris, Juncus; 92-94% overall accuracy), and classified airborne HyPlant hyperspectral imagery (483 bands) into 12 land cover classes for 3 sites (scikit-learn, rioxarray)
• Quantified statistically and visualized land cover on about 61,000 km² of peatland in 6 Baltic Sea countries from 23 annual BSRLC+ maps (2000-2022), e.g. peat extraction area falling from 741 to 602 km² (rioxarray, dask-geopandas)
• Conducted regular UAV survey missions with the DJI Matrice 300 RTK, P4 Multispectral, and Mini 2, carrying MicaSense Altum-PT multispectral/thermal, Zenmuse P1 RGB, and Zenmuse L1 LiDAR sensors. Flew monthly monitoring at 10-120 m altitude with RTK/SAPOS positioning and surveyed GCPs with an RTK GNSS rover and QField
• Processed drone data in Agisoft Metashape, DJI Terra, and R into orthomosaics and elevation models at 4 mm to 6 cm ground resolution (code). Derived LiDAR products from Zenmuse L1 point clouds with lidR (quasigeoid height correction, noise filtering, CSF ground classification, 0.3 m DSM/DTM)
• Built an EnMAP hyperspectral library using automated Python workflow (GDAL, rasterio, GeoPandas). 63,124 labelled spectra (224 bands) from 10 scenes across Germany for 10 WorldCover land cover classes, with cloud masking, pure-pixel sampling and class balancing

Remote Sensing Intern

Thünen Institute

Large-scale forest inventory based on ALS point cloud

• Developed an automated R workflow (lidR, crownsegmentr) to derive individual tree attributes (coordinates, height, crown area, crown base height, DBH) from airborne LiDAR and applied it to over 4,500 km² of heterogeneous forest in North Rhine-Westphalia, Germany
• Benchmarked raster-based (itcSegment) and point-cloud-based (AMS3D) tree segmentation with 3 developed allometric DBH equations against 4,072 measured trees on 520 inventory plots: Achieved 78% tree detection rate, estimated tree height with R² = 0.86 and DBH with R² up to 0.74

Remote Sensing & GIS technician

Centre for Econics and Ecosystem Management, Eberswalde University

• Prepared on-screen ground truth data for crop type classification across Kazakhstan, Tajikistan, and Pakistan: digitized 1,690 field boundaries from Sentinel-2 and Google Earth imagery. Labelled historic crop types for 2,964 field-years (2019-2021) plus 781 stratified land cover reference points from Sentinel-2 NDVI and Sentinel-1 VV/VH time- series profiles (QGIS, GEE Timeseries Explorer
• Produced an EGU 2023 poster (QGIS), the overview map (publication), and plotted monthly NDVI and Sentinel-1 backscatter profiles per crop type in R (dplyr, ggplot2). Compared 4 global cropland products (GLC_FCS30, GLC10, MODIS, ESA) and reviewed about 70 crop mapping studies

M.Sc

Eberswalde University for Sustainable Development

Forest Information Technology

• 1.0 grade (highest German grade, with distinction)
• Awarded DAAD STIBET scholarship for outstanding students

B.Sc

Hanoi University of Science, Vietnam National University

Environmental Science

Crop type classification in smallholder agriculture of Central and South Asia using Sentinel-1/2 data fusion

Remote Sensing Applications: Society and Environment

C. Raab, V. D. Nguyen, B. Barrett, A. Zakirova, M. Zuberi, H. Alff, M. Spies

An annual land cover dataset for the Baltic Sea Region with crop types and peat bogs at 30 m from 2000 to 2022

Scientific Data

V.-D. Pham, F. de Waard, F. Thiel, B. Bobertz, C. Hellmann, D.-V. Nguyen, et al.

Crop type mapping in Central and South Asia using Sentinel-1 and Sentinel-2 remote sensing data (poster)

EGU General Assembly 2023, Vienna

C. Raab, V. D. Nguyen

Annual land cover visualization, Baltic Sea region (React)

GitHub | Live

Drone photogrammetry workflow in Metashape (Python)

GitHub

CV Builder (React)

GitHub | Live

Library application

GitHub | Live