Me and my education
I am a Research Scientist at the Department of Primary Industries and Regional Development (DPIRD) in Western Australia, where I lead UAV data capture and analysis for the Australian Plant Phenomics Network's DPIRD Node. My work focuses on turning aerial imagery - RGB, multispectral, and hyperspectral - into sub-centimetre-accurate data products that support agricultural research across the state. Day to day, that means operating sophisticated sensor systems like the Headwall Co-aligned VNIR+SWIR and PhaseOne, building the pipelines that process what comes off them, and developing Python-based geospatial tools that field researchers actually use.
I hold a PhD in Geomatics Engineering from the University of Melbourne, where my thesis explored remote sensing and computer vision techniques for urban building footprint extraction. Before DPIRD, I worked across academia and industry - leading satellite and machine learning projects at ListenField in Thailand, researching tree inventory and flood-claim validation systems at Thammasat University, and managing spatial databases for child health monitoring in Nepal. I also tutored remote sensing and spatial data analytics subjects at the University of Melbourne.
I am a CASA-licensed remote pilot and have published in journals including IEEE GRSL, Expert Systems with Applications, and Remote Sensing. You can find my open-sourced code on GitHub and my papers on Google Scholar.
Education
2021-2025
PhD
University of Melbourne
My doctoral research developed deep learning methods for extracting building footprints from off-nadir aerial and satellite imagery - a task complicated by oblique viewing angles, façade occlusion, and misaligned labels in open datasets. I proposed new architectures combining CNNs, Vision Transformers, and edge-aware boundary regularisation, alongside techniques for multi-scale feature aggregation, domain adaptation, and knowledge transfer. The work contributed to a framework for nationwide building footprint data products in Australia, with applications in urban planning, disaster response, and sustainable city development.
Thesis: Remote Sensing Images and Computer Vision Techniques for Urban Building Footprint Extraction
Supervisors: A/Prof. Jagannath Aryal and Prof. Abbas Rajabifard
Funded by Melbourne Research Scholarship
Nominated for The Chancellor's Prize for Excellence in PhD Thesis
2017-2019
Masters Degree
SIIT, Thammasat University
My research-based Master's focused on applied AI for agriculture, combining UAV photogrammetry with deep learning to detect and count individual plants in commercial farms. The thesis demonstrated one of the early end-to-end pipelines for banana plant counting from high-resolution drone imagery, later published in PLOS ONE and protected under a Thai patent. The work also explored machine learning for short-term weather forecasting in agricultural contexts.
Thesis: Useful AI Applications in Agriculture - Aggregation of Machine Learning Techniques for Weather Forecasting and Banana Plant Counting
Funded by SIIT Faculty Scholarship
2011-2015
Bachelor
Kathmandu University
A four-year engineering degree covering the foundations of surveying, photogrammetry, GIS, remote sensing, geodesy, cartography, and spatial database systems. The programme built the technical base - coordinate systems, sensor principles, and geospatial software - that has underpinned everything since. During my studies, I served as Joint Secretary of the Nepal Geomatics Engineering Society (2015–2017).
Regional Merit Scholarship recipient (Govt. of Nepal & Kathmandu University)