S.O.I.L. AI

By Michael Wallinger & Microfy Systems

VIDEO

Concept & Prototype description

SoilAI is an accessible, user-friendly cyberphysical system designed for automated microbiome analysis of soil samples using image processing and artificial intelligence. The system consists of a digital microscope, desktop app, AI module, and dashboard, enabling non-lab users to easily analyze soil samples with minimal preparation. The digital microscope scans soil samples on multiple planes, capturing high-resolution images that are then uploaded to the cloud for processing. AI algorithms classify microorganisms, such as bacteria, fungi, nematodes, and other soil organisms, into functional groups, determining whether they are beneficial or pathogenic. The system provides users with a clear, understandable report, including insights relevant to the type of crops being grown, allowing for informed soil management decisions.

Connection to the scenario
Future scenario: Patterns that Persist

The scenario we have chosen is “Patterns that persist”. This scenario draws a utopia in which biodiversity becomes the new benchmark for a healthy food system after a new law by the European Union. But it also draws a polarized future between early adopters and people who feel left behind. While the scenario and a large part of the discourse pays a lot of attention on increasing the population of pollinator insects, the invisible but much larger biodiversity of soil microorganisms sometimes seems to be a bit out of focus. Our team has selected this scenario because it is seriously concerned about the overall food system’s health and sustainability and how food production & consumption will compensate each other over the years to come, considering Earth’s growing population combined with a future lack of resources. Our concept focuses on soil microorganism communities and aims to support the transition to healthy i.e. regenerative soil ecosystems. As an accessible decision support system, it helps to assess agro-ecological and other biodiversity-friendly practices based on indicators of soil biodiversity.

Technology

SoilAI is an integrated system combining hardware and software to autonomously analyze soil microbiomes. The hardware includes an automated microscope with robotics for scanning soil samples at different magnifications, while the software consists of an app, a dashboard, and AI models. The microscope scans the sample, capturing images for processing in the cloud, where AI algorithms identify and classify soil organisms, such as bacteria, fungi, nematodes, and others. The app controls the microscope and communicates with the cloud, while the dashboard displays results. Future advancements include automating the sample loader and objective switching. AI elements include traditional image processing algorithms for edge detection (Sobel, Laplacian, Canny) and deep learning models (FastRCNN/MaskRCNN, EfficientNET) for detecting and classifying species. Custom datasets will be created to train the models for accurate identification and classification, focusing on various soil organisms at different magnifications.