Soil Data Insights for Farmers | EarthOptics

The Evolution of Precision Agriculture: From Tractors to Big Data

Over recent decades, sustainable and efficient farming has transitioned from being primarily a challenge of mechanical power to one centered around data management. Startup EarthOptics posits that the subsequent advancement in precision agriculture resides within the soil itself. The company utilizes sophisticated imaging technologies to map the physical and chemical makeup of agricultural fields with greater speed, accuracy, and cost-effectiveness than conventional methods, and has secured $10 million in funding to broaden the reach of its solution.

Addressing a Long-Standing Gap in Soil Analysis

“Traditional soil monitoring techniques have remained largely unchanged for half a century,” explains Lars Dyrud, founder and CEO of EarthOptics, in a statement to TechCrunch. “Significant progress has been made in precision data and the application of modern data methodologies within agriculture, however, much of this has concentrated on plant health and in-season monitoring. Investment in comprehensive soil analysis has been comparatively limited.”

Despite the apparent necessity of understanding the foundation from which crops grow, detailed soil analysis has historically been a complex undertaking. While aerial and satellite imagery, alongside IoT sensors measuring factors like moisture and nitrogen levels, have enriched surface-level field data, assessing conditions beyond the top foot of soil presents considerable difficulties.

The Impact of Soil Characteristics on Crop Yield

Variations in soil characteristics across a field – such as compaction levels – can substantially influence crop yields. Similarly, chemical properties like nutrient content and the composition of the soil microbiome play a crucial role. Currently, the most reliable method for evaluating these factors involves extensive physical sampling, a process Dyrud describes as “putting a really expensive stick in the ground.” The resulting lab analyses then inform decisions regarding tilling and fertilization strategies.

This practice, while essential, is costly, particularly for large-scale farms. Analyzing soil samples from every few acres, potentially multiple times annually, quickly accumulates significant expenses when managing thousands of acres. Consequently, many farmers resort to blanket tilling and fertilization due to a lack of granular data, potentially wasting resources – Dyrud estimates around $1 billion annually in the U.S. is spent on unnecessary tilling – and even causing environmental harm by releasing sequestered carbon.

EarthOptics' Innovative Approach to Soil Mapping

EarthOptics aims to revolutionize data collection by minimizing the need for extensive physical sampling. The company has developed an imaging system leveraging ground penetrating radar and electromagnetic induction to create detailed soil maps more efficiently, affordably, and accurately than extrapolating data from sparse samples.

Machine learning is central to EarthOptics’ two primary tools: GroundOwl and C-Mapper (representing carbon). The team has trained a model that correlates non-invasive data with traditional samples collected at lower densities, enabling accurate prediction of soil characteristics with unprecedented precision. The imaging hardware can be readily mounted on standard tractors or trucks, capturing readings at intervals of just a few feet. While physical sampling remains a component, its frequency is reduced from hundreds to dozens of times.

From Broad Zones to Meter-Scale Precision

Conventional methods often divide large fields into broad zones – for example, a 50-acre section requiring additional nitrogen, another needing tilling, and so on. EarthOptics refines this granularity to the meter scale, providing data that can be directly integrated into automated field machinery, such as variable-depth smart tillers.

These machines can then adjust their operation depth based on real-time soil conditions. For farms without advanced equipment, the data can be presented as a conventional map, guiding operators on when and where to perform specific tasks.

Potential Benefits and the Future of Automated Farming

Successful implementation of this technology could lead to substantial cost savings for farmers or increased productivity per acre and dollar for those seeking expansion. Ultimately, the goal is to facilitate the advancement of automated and robotic farming practices. While this transition is still in its early stages, requiring refinement of both equipment and methodologies, a foundational element will be the availability of high-quality data.

Dyrud envisions the EarthOptics sensor suite integrated into robotic tractors, tillers, and other agricultural machinery, but emphasizes that the core value proposition lies in the data itself and the sophisticated machine learning model trained on tens of thousands of ground truth measurements.

Funding and Future Development

The $10.3 million Series A funding round was spearheaded by Leaps by Bayer, Bayer’s impact investment division, with contributions from S2G Ventures, FHB Ventures, Middleland Capital’s VTC Ventures, and Route 66 Ventures. These funds will be allocated to scaling the existing product offerings and developing a new solution: moisture mapping, a critical consideration for all agricultural operations.