Under the direction of Professor Saugata Datta, a creative new endeavor is taking root at the University of Texas at San Antonio (UTSA). He’s applying big data and artificial intelligence to make soil health testing more impactful and accessible. This innovative AI-powered microscope system is revolutionizing soil analysis and improvement. It speeds up, lowers the cost of, and opens up access to farmers and land managers worldwide. Alec Graves from UTSA’s College of Sciences demonstrates innovative research at the Goldschmidt Conference. His work illustrates the opportunity we have to start shifting our country’s agricultural practices.
The research team developed the system using a comprehensive dataset of several thousand images of fungi collected from soils across South Central Texas. Their aim is to have a polished, deployable device ready for in-the-field testing within two years. This AI technology was tuned specifically for the everyday microscope magnifications that most labs and schools use – 100x, 400x. Exposure at these magnifications, unfortunately, may be all too common in the school lab. This accessibility could democratize soil health testing, making it easier for less experienced individuals to obtain critical information about soil biology.
Advancements in Soil Analysis
Unfortunately, the limitations of how biological soil analysis is currently done or reported are common. Conventional methods require costly lab infrastructure to determine molecular makeup. They leave it up to experts whose eyes on the ground can visually identify organisms using a laboratory microscope.
“Current forms of biological soil analysis are limited, requiring either expensive laboratory equipment to measure molecular composition or an expert to identify organisms by sight using laboratory microscopes.” – Alec Graves
The new, AI-powered microscope system meets these challenges head-on. The interdisciplinary research team is working to combine machine learning algorithms with optical microscopy. With this innovative approach, we’re hoping to develop a low-cost solution that makes soil testing easier and requires less labor and technical expertise. This cutting-edge approach holds the potential to give us a much fuller picture of soil biology.
“Using machine learning algorithms and an optical microscope, we’re creating a low-cost solution for soil testing that reduces the labor and expertise required, while providing a more complete picture of soil biology.” – Alec Graves
The Technology Behind the System
The cutting-edge system employs machine learning and artificial intelligence to study thousands of videos of soil core samples. We take this footage and separate it into single frames. Then, we use a specific neural network to accurately identify and quantify fungi abundantly found in the soil. This approach improves and increases the efficiency and accuracy of the analysis. Those specific enhancements are critical for farmers and ranchers to make the important on-farm decisions.
The accessibility of this technology is noteworthy. You’ll find that it’s just as simple to find microscopes that are compatible with the system in educational environments. This access enables farmers and land managers to test drive this tool without a large upfront investment. The research team expects to publish other technical, in-depth findings about their new machine learning algorithm in a peer-reviewed journal later this year. Together, this work will yield precious lessons learned to the greater scientific community.
Future Implications for Agriculture
The consequences of this research go far beyond better soil health testing. With the AI-powered microscope system, farmers and land managers now will have easy access to detailed, reliable data on soil biology. Such access helps facilitate more data-driven and climate-smart agricultural practices. Soil health is key to ensuring crop yield and sustainability. Unless we harness this technology to increase the rate of global food production.
He’s passionate about the need for sustainable practices in agriculture. With this new and unusual approach, he hopes to further scientific understanding. Further, he wants to advance practical applications that will serve the end-users across numerous agricultural markets.