EarthOptics helps farmers dig deep into soil for big data insights – TechCrunch


Sustainable and efficient farming has grown from a big tractor problem to a big data problem over the past few decades, and the startup EarthOptic believes the next frontier in precision farming lies deep in the ground. Using high-tech imaging techniques, the company claims to map the physical and chemical composition of fields faster, better, and cheaper than traditional techniques, and has raised $ 10 million to scale its solution.

“Most of the ways we monitor the ground haven’t changed in 50 years,” EarthOptics founder and CEO Lars Dyrud told TechCrunch. “There has been huge progress around precision data and the use of modern data methods in agriculture – but a lot of it has focused on plants and seasonal activity – there has been relatively little investment in soil. “

While you might think it’s obvious to take a more in-depth look at the material plants grow from, the simple fact is, it’s hard to do. Aerial and satellite imagery and IoT-infused sensors for things like humidity and nitrogen have made surface-level data for fields much richer, but beyond the first foot the things get complicated.

Different parts of a field can have very different levels of physical characteristics like soil compaction, which can greatly affect crop performance, and chemical characteristics like dissolved nutrients and the microbiome. The best way to verify these things, however, is to “put a really expensive stick in the ground,” Dyrud said. The laboratory results of these samples affect the decision of which parts of a field should be plowed and fertilized.

It’s always important, so farms do, but having a soil sample every few acres once or twice a year quickly adds up when you have 10,000 acres to track. So many people plow and fertilize everything for lack of data, spending a lot of money (Dyrud estimated that the United States makes about $ 1 billion in unnecessary plowing) in processes that might have no benefit and in fact might be harmful – it can release tons of carbon that has been safely sequestered underground.

EarthOptics aims to improve the data collection process essentially by minimizing the “expensive stick” part. He built an imaging suite that leverages ground penetrating radar and electromagnetic induction to produce a deep soil map that is easier, cheaper, and more accurate than extrapolating hectares of data to from a single sample.

Machine learning is at the heart of the company’s tools pair, GroundOwl and C-Mapper (C as in carbon). The team trained a model that reconciles non-contacting data with traditional samples taken at a much lower rate, learning to accurately predict soil characteristics to a level of accuracy far beyond what was traditionally possible. Imaging equipment can be mounted on regular tractors or trucks and takes readings every few feet. Physical sampling always occurs, but dozens rather than hundreds of times.

With today’s methods, you could split your thousands of acres into 50 acre chunks: this one needs more nitrogen, this one needs to plow, this one needs so-and-so. such treatment. EarthOptics brings that down to the meter scale, and the data can be fed directly into robotic field machines like an intelligent variable-depth tiller.

Drive it along the fields and it only goes as far as it needs to. Of course, not everyone has state-of-the-art equipment, so the data can also be presented in the form of a more ordinary map telling the operator more generally when to plow or perform other tasks.

If this approach takes off, it could mean significant savings for farmers looking to tighten up, or improved productivity per acre per dollar for those looking to expand. And ultimately, the goal is to enable automated and robotic agriculture as well. This transition is only in its early stages as equipment and practices are developed, but they will all need good data.

Dyrud said he hopes to see the EarthOptics sensor suite on robotic tractors, tillers and other farm equipment, but that their product is largely the data and machine learning model they formed with it. tens of thousands of ground truth measurements.

The $ 10.3 million round was led by Leaps by Bayer (the conglomerate’s impact arm), with participation from S2G Ventures, FHB Ventures, VTC Ventures of Middleland Capital and Route 66 Ventures. The plan for the money is to scale the two existing products and get to work on the next one: moisture mapping, obviously a major consideration for any farm.


About Author

Comments are closed.