About This Project

Soil looks like one brown background, but water, minerals, organic matter, roots, fungi, bacteria, and decaying plant material all change how it reflects light. I will build a low-cost benchtop soil spectra scanner and use it to measure controlled soil samples across moisture, and texture. The output will be an open hardware recipe, calibration protocol, pilot dataset, and analysis notebooks that help hyperspectral biology researchers test faint biological signals against real soil backgrounds.

Ask the Scientists

What is the context of this research?

Biological signals have usually been measured up close, through microscopes, lab assays, or direct sampling. Hyperspectral biology expands that view by measuring how molecules and organisms reflect or absorb many wavelengths of light, revealing features ordinary RGB cameras miss. But hyperspectral cameras often cost $10,000 to $100,000, limiting access for small labs and early experiments.

Recent work showed that engineered bacteria with hyperspectral reporter molecules could be detected outdoors from up to 90 meters away in a 4,000 m² image, but only by separating the signal from backgrounds like dirt and sand. Soil is a useful first target: water, minerals, organic matter, roots, fungi, microbes, plant residues, texture, and surface roughness all affect reflectance.

My hypothesis is that a low-cost, controlled-light benchtop scanner can produce repeatable soil reflectance curves that separate major background changes, especially moisture, texture, and organic matter gradients.

What is the significance of this project?

Hyperspectral biology needs more than better reporters. It needs cheap instruments, calibration protocols, and public datasets that let researchers test whether a biological signal can be separated from real-world backgrounds.

This project is useful because it lowers the barrier to entry. A researcher, community lab, classroom, or independent group should be able to reproduce the scanner, collect soil spectra, and compare their data against an open pilot dataset. The grant program says it is interested in early, cross-disciplinary, tool-oriented, field-building projects where a small grant makes a real difference. This is exactly that kind of project.

The project also avoids the riskiest parts of hyperspectral biosensing. There will be no engineered organism release, no drone deployment, and no field biosensor claim. The first step is simpler: measure the soil background well.

What are the goals of the project?

I will build a low-cost benchtop scanner to measure soil reflectance spectra under controlled lighting. The prototype will use an off-the-shelf mini-spectrometer, fixed sample geometry, white/dark references, and a light-controlled chamber. Repeatability will be tested by rescanning soil cups across sessions.

I will create a pilot dataset from about six source soils scanned at controlled moisture levels. A subset will be scanned across compost-amendment levels by dry mass to test detection of organic-matter-like background changes, without claiming exact soil organic matter for every replicate. About six soils will receive external testing where feasible, including organic matter, carbon/nitrogen, and texture.

Analysis will convert raw readings into reflectance curves, quantify repeatability, and test models for separating moisture, texture, and compost gradients. The release will include spectra, metadata, code, notebooks, bill of materials, enclosure files, and a build guide.