Momina Khurshid

Processed satellite, drone data and health indices analysis for crop anomalies
report (health, salinity, water logging, crop lodging, etc.)
Planned and executed automated KMZ-based flights using Phantom 4 Pro,
Mavic 3 MS, and Air3S, achieving down to ~0.3 mm GSD.
Extracted UAV telemetry from DJI .dat logs (ultrasonic height, camera angles)
for precise geo-alignment without stitching.
Designed UAV-based workflows for forest monitoring, biomass assessment,
and restoration evaluation.
Conducted UAV and satellite-based analysis for tree counting, health
classification, and invasive species mapping, utilizing geospatial alignment,
semi-supervised labeling, and NDVI metrics for targeted ecological
management.
Generated orthomosaics and DSMs using Metashape/Pix4D and automated
large-scale processing using GDAL and ArcPy.
Processed Sentinel-1/2 and PlanetScope imagery and built pipelines on AWS
EC2 & Jupyter for scalable analysis.
Created labeled datasets using SAM, GroundingDINO, and Label Studio for
tree, crop, and disease detection tasks.
Trained YOLOv8, DeTR, and Detectron2 models with RayTune optimization
and metric tracking (P/R/F1).
Worked on robust evaluation workflows to identify best and worst detections
in evaluation data and use semi-supervised learning method to improve and
label data.
Developed and deployed a cloud-based dual image alignment pipeline on
AWS Lambda using feature-based matching (GIM-model SIFT + LightGlue)
for high-accuracy UAV imagery registration.

I hold an MS in Geospatial Data Science from NED University and a Master’s and Bachelor’s degree in Geology from the University of Karachi, where I secured 2nd position in both programs. I completed my Higher Secondary School education in Pre-Engineering from Sheikh Zayed College, Karachi.