Russell J.

Aerial Vision Science is Russell — a researcher with a background spanning research science, AI, and electronic engineering. I decided to make a drone survey business that provides high level analysis and insight on top of drone survey data.

My focus is applying sophisticated computer vision analysis to wildlife and plant species census work. My research background includes studying several medical imaging modalities — MRI, fMRI, and EEG, as well as working with direct neuronal recordings. These are all examples of time-varying images of the brain with many direct analogues to remote sensing. Both fields utilise precise object detection and segmentation in high-resolution imagery, and methodologies that hold up to scientific scrutiny. More recently I completed the Computer Vision unit at La Trobe University as part of a Graduate Certificate in AI, a short course in Drones for Environmental Science at Duke University (Coursera), and a short course in Multispectral Imaging at Firetail Environmentals.

I've spent decades solving large-scale data problems on Linux infrastructure, which means processing large survey datasets is routine rather than exceptional.

My services are well matched to scientists, researchers, land managers, and anyone with an appreciation for transparent, replicable methods. If your project requires a methodology that can be audited, replicated, and re-run years later with fresh data — that's exactly what I build.

The analysis framework is cloud-hosted, lightweight, and built on Docker with GPU acceleration for model training. Data annotation uses Label Studio; spatial results are validated through a MapLibre web interface. The entire stack can be handed over as a virtual machine for clients who want long-term analytical independence.

180

Custom software, GeoAI, Pytorch, torchvision, QGIS, Open Drone Map, LibreMap, Cloud Compare.

8