Integration of local knowledge for climate smart farming with spectrometer

Important results from the project

The project aims to create a solution to easily, quickly and economically analyze soils with a handheld spectrometer and external sensors directly in the field and/or after drying/sieving in a home environment. The translation from spectral signal to soil property is done via an online database of reference samples and models defined via machine learning. Sampling, analysis and prediction of properties are controlled via a mobile app.

Expected long term effects

The project has led to a complete system solution for soil analysis with a handheld spectrometer controlled by a mobile app connected to an online database with services for modelling, calculation and presentation of spectral soil analyses. Sensors, light sources, electronics and the software of the spectrometer have been integrated and provide stable and accurate measurements. The system is technically ready for demonstrations and has generated interest among investors and potential partners.

Approach and implementation

** Denna text är maskinöversatt ** From roughly 40,000 soil samples with associated spectra across Europe, a basic database with associated models has been built up. For evaluation of the developed spectrometer, approximately 200 soil samples have been tested and analyzed. The European models have an explanation rate of up to 90% for carbon and nitrogen content. Sampling directly in the field requires data on humidity and does not yet give as good results.