Development of the technologies behind precision agriculture is driven by the need to increase crop yields to feed the Earth’s multiplying population, while lowering costs and protecting the environment — which is not a small task. The concept of precision agriculture started at the field level using near-infrared imagery and GPS to help farmers make better decisions about applying fertilizers, pesticides and irrigation. Over time, the range of tools appropriate for collecting data to support precision agriculture has greatly expanded to include satellites, airplanes, drones, and a variety of ground and mobile sensors. Even more remarkable is the development of sophisticated analytical programs that enable users to partially automate feature extraction and change detection, analyze and compare data, and obtain an accurate understanding of real-world conditions. Depending on the collection and analysis methods selected, this situational intelligence can be used at a macro-level to predict global food production or at a local level to adjust the treatment of individual fields to maximize productivity.
Widespread adoption of remote sensing in the agriculture industry has been hampered by high imagery costs and difficulty proving a positive return on investment (ROI). However, new sources of data are now being offered as Earth observation (EO) satellite companies and drone service providers experiment to find more cost-effective and efficient methods of collecting agriculture and forestry data.
New Satellite Constellations Planned
The trend in EO satellites is to provide frequent coverage with constellations of multiple satellites. The European Commission owns two Sentinel-2 satellites that collect imagery in 13 spectral bands with a revisit time of five days at the Equator. By coordinating with NASA’s Landsat 8 operations, the revisit time is reduced and more data products are produced to support global agriculture and environmental monitoring efforts. The scheduled launch of Landsat 9 in 2020 will add to the data collection resources.
California-based Planet is a major data provider in the small satellite category and is already operating over 175 Dove and SkySat satellites, with more launches planned in 2018. In addition, Planet owns archive EO imagery dating back to 2009, which is useful for performing change detection and improving machine-learning algorithms. Planet signed a multi-year data distribution deal with Farmers Edge, a global agricultural insights provider, in October 2017.
There are several other projects underway that will also employ constellations of satellites. British company, Earth-I, currently utilizes four KOMPSAT satellites (three optical and one SAR) and three DMC3/TripleSat very high-resolution optical satellites to offer advanced EO analytics for agriculture applications. In 2019, Earth-I’s Vivid-I constellation featuring full-color video capabilities and still imagery is scheduled to launch. The following year, Canadian company, Urthecast, plans to launch the EarthDaily constellation consisting of eight 5-meter resolution satellites to provide global coverage every 24 hours, thus providing the agriculture market another source of data.
Drones Offer New Possibilities
The drone industry is taking a different and complementary approach by leveraging the small size, flexibility and convenience of unmanned aerial vehicles (UAV) and custom sensors. The rapid development of drone technology has focused on shrinking the size of the payloads, adapting sensors to specific applications, extending flight times, and improving collision avoidance and beyond visual line of sight (BVLOS) capabilities. Hardware options include fixed-wing or multi-rotor platforms equipped with combinations of sensors collecting video, multispectral, thermal, LiDAR and other types of data at the plant level.
Delair is one example of an end-to-end UAV solution company that designs fixed-wing systems and partners with leading sensor providers to create high-performance packages. Starting as a hardware manufacturer in 2011 in Toulouse, France, Delair’s founders quickly recognized the need for more accurate, in-depth analytical tools. Today, Delair offers a variety of agriculture algorithms to extract insights from the UAV data. Leveraging the plant-resolution measures and crop-specific traits acquired by the flying camera allows users to make data-driven business decisions.
“Business intelligence and actionable data is the goal, not just collecting data,” says Lénaïc Grignard, agriculture and forestry product manager at Delair. “Information supports new applications, such as using missing plant detection and plant counting data to make replanting decisions, with seamless connection to a precision planter. Also, seed research is being advanced by phenotyping plants (characterizing height, population, vigor, chlorophyll, etc.), and even simple bird view images help with drainage design and more.”
Grignard continues, “In forestry, we provide value to our customers by applying machine learning and image classification algorithms to colorized point clouds, collected simultaneously with LiDAR and an RGB camera, and delivering area-specific reports on a variety of topics, such as logging.”
Delair’s DT18Ag fixed-wing UAV is suitable for the agriculture market because of its two-hour flying time and capacity to collect up to 2,800 acres of video and still images on one flight. This 2-kg UAV is certified for VLOS and BVLOS flying.
“As the multispectral sensors on UAVs get smaller and lighter, and the flying range of UAVs extends, collecting large areas becomes increasingly cost effective,” Grignard says. “We believe that even growers of low-value, large-scale crops will experience a positive ROI since we can fly in a few days what would take weeks to monitor on foot. Early detection of problems can prevent widespread crop failure and yield reduction and save growers a lot of money.”
Effective platform and payload combinations are being created specifically to target the agriculture and forestry market. At the 2018 ILMF conference in Denver, Delair announced its DT26X LiDAR UAV, the first fixed-wing platform to package the new RIEGL miniVUX-1DL LiDAR sensor with an RGB camera.
“The combined payload of a lightweight sensor and integrated camera allows the acquisition of LiDAR and photogrammetry data in a single flight,” explains Grignard. “This drastically reduces cost and immediately provides an extremely detailed digital model of the inspected assets.”
Intelligence Adds Value to Precision Agriculture
Whether your interest is in big data on a global scale or field data at a local level, precision agriculture has come a long way. Although adoption of this technology can still seem daunting, numerous obstacles have been overcome and an increasing number of growers are recognizing the value of remotely sensed data in their daily operations.
“Increased productivity with multiple fields acquired in one flight and the ability to perform onboard quality checks are key to gaining acceptance of the technology,” says Grignard. “Also, timeliness of delivering insights is very important in agriculture because crops don’t wait for you to be ready to do weed control. Quicker information turnaround is now possible thanks to quasi fully automated cloud processing services.”
In addition to collecting and summarizing data, precision agriculture delivers actionable information. Drone-to-tractor data integration through services like the John Deere Operations Center uses agriculture intelligence from partner providers such as Delair, Agribotix, and PrecisionHawk to directly manage reseeding, fertilizer and pesticide applications.
“New sensors will enrich the possibility of analysis, and thus, the range and efficiency of actions in the field,” says Grignard. “Combined with other tools in the ecosystem, such as precise weed control robots, the resulting intelligence will contribute to more efficient, sustainable and smarter crop systems.”