Overview

ScanSpectrum is a line of low-cost, portable spectrometers. We aim for both affordability and high build quality, combining careful design with inexpensive yet powerful components.

Our inspiration for ScanSpectrum is the scarcity of publicly available agricultural and medical data. The pain points we address are: (1) the high prices of desktop lab instruments for compositional assessment, and (2) the logistical costs of ferrying samples between the field and the lab. We believe collecting larger, more spatially distributed datasets with “good enough” sensors is much more effective than having only a few expensive sensors at limited sites. When datasets are large enough, machine learning provides a pathway to calibrating inexpensive sensors against “gold standard” equipment.

While ScanSpectrum is still in early-stage development, it is a very promising prototype. We are piloting devices with partners in a few countries, and been able to successfully detect particular plant stresses and illnesses. We also promote a hybrid approach that switches between dry and wet chemistry, depending on what is being assessed. Our technology can be reshaped into a variety of different form factors, such as handheld scanners for in-field manual assessment, or ruggedized modules for automated assembly lines.

Our current in-field sensor is a gun-shaped device carried in a compact, fitted case complete with documentation, accessories, and spare parts. It interfaces with an Android phone via Bluetooth, and reports both the spectrum and NDVI in a matter of seconds, while also saving this georeferenced and data to the cloud in openly accessible formats. Sample preparation procedures differ depending on targets, but can be automated with custom design.

We are interested in partnering with organizations aiming to solve specific industrial use cases with spectroscopy, and manufacturing partners or philanthropists to scale up production.