North Carolina based agtech startup EarthOptics helps the farmers obtain big data insights of soil

Oct, 2021 - By WMR

The company offers to map the physical and chemical characteristics of farms fast, better, and cost-effective than traditional methods with high-tech imaging technology.

Over the last two decades, farming has evolved from a big tractor problem to a big data problem, and EarthOptics acknowledges the next stage of precision farming lies deep in the ground. Physical qualities such as soil compaction, which can have a big impact on crop yields, and chemical properties such as dissolved nutrients and the microbiome can be extremely different for different portions of a field. The results of these samples in the lab have an impact on the decision of which areas of a field should be cultivated and fertilized.

For lack of data, many people simply till and fertilize everything, investing a lot of money in activities that may or may not be useful, and might be harmful – it can release masses of carbon that was previously safely trapped underground. EarthOptics intends to improve the data collection process by reducing the amount of time spent in the lab. It has developed an imaging system that uses soil penetrating radar and electromagnetic induction to create a deep image of the soil that is faster, less expensive, and more exact than projecting acres of information from a data sample. GroundOwl and C-Mapper, the company's two tools, are built around machine learning.

The team developed a model that combines no-contact data with traditional samples taken at a reduced rate, learning to accurately predict soil characteristics at a level of precision far beyond what has previously become possible. The imaging hardware, which can be mounted on regular tractors or trucks, takes readings every few feet. EarthOptics reduces this to meters, and the data can be directly fed into robotic field machinery such as a variable depth smart tiller. For those who don't have cutting-edge technology, the data can also be presented as a more traditional map, instructing the driver on when to till or perform operations.