Transect Lab
Autonomous platforms and custom sensors for conservation research
Aerial and surface survey, multispectral and underwater imaging, and AI-driven analytics, in collaboration with ecologists
Survey capability for conservation research, from the air and on the water
Transect Lab builds and operates autonomous survey platforms, custom sensors and machine learning analytics for conservation research. A multirotor UAV carries an in-house multispectral imager over land, wetland and intertidal habitat; a survey boat maps what lies beneath lakes, estuaries and nearshore waters. Both run systematic, repeatable survey lines, the transects the lab is named for, so results can be compared from one survey to the next.
The lab is a methods partner, not a survey contractor. Collaborating scientists bring the research question, field knowledge and ecological interpretation; Transect Lab brings the platforms, processing pipelines and models, and shares the work of publishing the results. Each side brings what the other lacks, so together they can take on questions neither could tackle alone.
From survey to decision
Aerial survey
A Holybro X650 multirotor carries the lab's multispectral imager over terrestrial, wetland and intertidal habitat. It is being prepared for beyond visual line of sight (BVLOS) operation under CASA accreditation.
Surface survey
A quiet, car-portable survey boat built on BlueBoat hulls maps lakes, estuaries and nearshore reef: RTK bathymetry, georeferenced underwater imagery and seabed habitat, launched from the shore by one person.
Custom multispectral imager
Designed and built in-house with swappable narrow bandpass filters, so the bands are chosen to fit the research question rather than fixed by a commercial sensor. Radiometrically calibrated against traceable reflectance targets.
Calibrated, repeatable data
PPK-georeferenced orthomosaics, calibrated reflectance rasters and bathymetry grids, captured to a fixed protocol so repeat surveys are directly comparable and change is measured with stated uncertainty.
Machine learning analytics
Detection, segmentation and change models trained for the question at hand, from invasive weeds and vegetation condition to benthic habitat and sea urchin density, with training data, class definitions and validation documented.
Results built around the question
Target maps, change maps and headline metrics designed with the research or management partner around the decision they need to make, and delivered in the formats their own systems already use.
Publications, data and releases
What the platforms and methods produce, published openly: papers and preprints, example survey data, validation results, and open-source releases.
Publications and preprints
Methods papers on the platforms, the imager and the processing pipeline, and results co-authored with collaborating researchers.
Datasets and maps
Example survey outputs: orthomosaics, calibrated reflectance and index maps, bathymetry grids and underwater imagery.
Validation results
Per-band radiometric accuracy of the imager, positional accuracy of the survey data, and detection model performance.
Open-source releases
Build documentation for the platforms and imager, and the processing pipelines, released with DOIs.
Working with research groups
Transect Lab works through research collaborations. We join projects as a technical partner: shaping the survey design with you, collecting and processing the data, building the analysis, and co-authoring the methods and results. On grant-funded projects the lab can be named as a technical partner, with platform and processing costs written into the budget.
We are interested in projects on weed and vegetation mapping, wetland and saltmarsh habitat, coastal and benthic monitoring (including urchin barrens), and freshwater bathymetry. If your question needs repeatable aerial or underwater survey, get in touch at info@transectlab.au.
Open-source and peer-reviewed
Transect Lab is part of the open-source community. We build on open hardware and open software, from autopilot configurations and sensor hardware to geospatial processing pipelines and trained models, and we contribute back. Our tools are designed to be reproducible, forkable, and useful beyond this lab.
Methods and results are published as preprints and in peer-reviewed journals, and presented at conferences spanning ecology, remote sensing, and computer vision. We believe open methods and freely available findings accelerate conservation science and raise the bar for the entire field.
Daryl Wilding-McBride
Daryl founded Transect Lab in 2026, bringing together three threads he has pursued across his career: a long-held interest in autonomous aerial vehicles, academic training and expertise in using deep learning to find structure in complex, high-dimensional data, and deep experience applying technology to solve hard problems.
His industry career spans decades of technology consulting. His academic work includes a PhD in computational mass spectrometry (detecting peptide features in high-dimensional point cloud data) and postdoctoral research in medical imaging, training deep learning models to monitor tumour response to treatment and improve treatment outcomes in glioblastoma.
He built his first autonomous multirotor UAV from scratch in 2012, and in 2014 co-founded Flying Robot School Inc, a non-profit organisation teaching secondary school students about aerial robotics through hands-on applications. He now builds and operates autonomous survey platforms, a multirotor UAV and a survey boat, for conservation research.