Editor’s Note: Igor Ivanov is chief commerical officer of Gamaya, a Swiss agtech startup developing proprietary hyperspectral imaging technology, using drones, to provide farmers insights about their fields. He offers a survey of the leading remote sensing companies. 

There are many agtech startups that use remote sensing technology for different precision farming applications. Some try to provide a field-level analysis using drones or airplanes, while others develop large-scale macro analytics using globally available satellite data.

To some extent it is possible to connect airplane or drone-based data with field-level information, and drone-based data with satellite data. Indeed, some of these companies try to combine multiple sources and build intelligence around it.

Though it is entirely possible to group these companies by the crop segment that they serve or their data processing technologies and capabilities, we have decided to focus on the different challenges associated with various data sources.

Here are the most prominent remote sensing companies, sorted by the main source of remote sensing data each uses: satellites, drones, airplanes and ground-based sensors, along with the challenges that come with each data source. (We have excluded companies that focus only on indoor farming from this analysis.)

Satellites offer Big Data but Little Control

Though satellites are an indispensable source of data worldwide, they offer relatively low spatial resolution, which makes it difficult to perform analysis on the plant level. Satellites don’t revisit the same locales often enough to offer daily data, which makes it hard to use the data for short growing seasons, like soybean season in Brazil, though there are some exceptions and Planet Labs claims to have a daily coverage.

Imagery from satellites is also highly susceptible to cloud cover — you can end up waiting for one to two months with no good data due to extensive clouds. Plus, most satellites have only RGB (a basic color spectrum based on the primary colors of red, green, and blue) and/or multispectral cameras, which offer a relatively low amount of information compared to other types.

Most notable players which largely rely on the satellite data:
Drones and Planes Mean Control and Flexibility But Costs Add Up

Drones and small planes are sources of imagery that offer flexibility and a relatively low barrier to entry, often putting control in the hands of the farmer, but that flexibility is not absolute. Challenges associated with using drones or airplanes include difficulty around scaling for thousands of hectares, particularly for large industrial growers and large areas of land. Drones also need to have a network of drone operators.

Furthermore, they are subject to weather interruptions, which add up to a high level of operational complexity. Drone are also subject to regulations that can significantly limit the usage of drones for commercial operations. It can be difficult to justify the high price of services using a drone or airplane-based data, as data acquisition using drones is quite expensive. On average it can vary from $1 to $5 per hectare per data acquisition.

Most prominent players that rely on the drone or airplane-based data:
Multiple Data Sources Means More Data, But Can Be Murky

Integrating data from multiple sources is becoming easier and easier with the development of more sophisticated artificial intelligence and machine learning, but challenges persist. Data is very irregular, fragmented, and noisy. Remote sensing data is exposed to environmental effects of weather, clouds, etc. and this requires a sophisticated calibration of data.

Agriculture is a very local, regional and crop-specific industry that should be addressed using custom approaches. Data collection is not standardized, so integrating and managing multiple sources of imagery data is tricky. There are no well-established data standards and different companies use a combination of various technologies to collect the data. Connecting data with internal physiological and biological properties of crops on a plant level is also very difficult, as this requires enormous ground-truthing (empirical evidence from the ground) and agronomic modeling.

Most prominent players that rely on a combination of drone/ airplane-based imagery, space-borne data and ground-based sensors to develop intelligence around it: