MEMS-Based Photonics Systems for Agricultural Analysis

General Info

MEMS-based photonics systems represent a low-cost device for the rapid measurement of spectra, suitable for the accumulation of information related to agricultural products, either by directly scanning them or through soil scanning. These systems utilize miniature Vis-NIR spectrometers integrated into mobile applications, offering a user-friendly and efficient way to assess soil and food quality. 

Key Components

To retrieve the information contained in a single spectral signature, three key components are required: 

Mobile Application: 

It is a unified interface used for operating MEMS-based photonics devices and assisting the operator to capture the reflectance of the food or soil asset. Through intuitive graphical user interface, the user can follow an easy step by step procedure and capture the required spectral signature through the point and shoot method. As a minimum operational requirement, each application should enable the user to offload the captured data, while the THEROS application automatically transmits the obtained spectral signatures to the THEROS harmonization platform. This data transmission complies with security protocols that ensure that no data spoofing or loss will take place and facilitates data archiving and the production of analysis-ready data. Furthermore, the captured reflectance is visualized, while the extracted information is provided as a report containing specific information per use case (i.e. organic farming compliance marker, product purity score to address product adulteration, and key soil properties quantification). 

Food & Soil Spectral Libraries: 

To convert each measurement obtained via the mobile application, a spectral library is required (Food or Soil accordingly). The library consists of a large number of different spectral signatures coupled with reference measurements (i.e. chemical analyses, or organic/conventional farming labels). These datasets a) will be enhanced with new measurements captured through the THEROS pilot demonstrations and b) will be used as calibration and validation sets for developing AI models for extracting the targeted information as modelling estimations.

AI Tools for Data Transformation: 

The AI component is used to transform a spectral signature to meaningful indicators of either soil or food assessment. To this end, state-of-the-art techniques are used to provide results with low uncertainty, end enhance organic farming certification, combat food adulteration and provide farmers with soil quality reports.

Interfaces

  • Spectral data collection via Vis-NIR scanning devices.

  • ML & DL algorithms for data analysis.

  • Mobile application UI for data visualization and report generation.

Services

Product Purity Index:

  • Estimates product purity by analyzing nutrient content deviations.

Organic Food Labelling Detection:

  • Develops deep learning classifiers to differentiate between organic and non-organic products.

Soil Quality Assessment:

  • Provides quantified information on soil properties such as organic carbon, clay content, acidity/alkalinity, calcium carbonates, and salinity.

Usage and applications

Designed for simplicity, the system employs a “point-and-shoot” method, enabling farmers, certification bodies, and retailers/wholesalers to easily assess soil and food quality. The mobile application could operate without an internet connection, storing data for later transmission. This technological advancement supports sustainable farming practices, compliance with organic standards, and enhances food safety monitoring.