Drones can enable precision and predictive farming

By February 25, 2016
Un drone au dessus d'un champ.

Although drones still have to face some technical challenges, they offer great opportunities for farmers to better manage their inputs as well as predict their yield.

Amazon is telling us that the UAVs are about to change the delivery sector. Google is not only entering delivery as well, but the Mountain View based firm is trying to bring internet coverage to remote areas through drones as well. However drones are also full of promises for agriculture.

With on-board sensors, infrared and color cameras but also laser scanners, drones offer great opportunities for growers to collect data about their plants and trees. For instance, a drone can spot an area in a field suffering from stress and through a wifi or cellular connection, notify the farmer. A task that the latter had to complete by visually checking the field before.

Yet today’s technology still has to evolve: drones are heavy and thus can only go on short missions. Their agility and autonomy need to progress. Especially in the US where half of the country’s cropland belongs to farms with a least 1,100 acres. Besides, even if prices are starting to drop, on-board technologies such as cameras and lasers remain expensive.

Vijay Kumar

Vijay Kumar presenting his drones during a TEDx confenrece. © Kumar Robotics

Autonomous and agile UAVs

Professor Vijay Kumar,  the Dean of the School of Engineering & Applied Science at the University of Pennsylvania, and his team hence decided to focus their attention on three aspects: making drones more autonomous, more agile, and capable of collecting data that is valuable to the growers. The primary objective is to go from data to information, i.e., provide actionable intelligence to growers.

The UAVs that the Vijay Kumar Lab has built span a wide range of sizes and capabilities. Some are small and light, consuming about 3 watts of power, and measuring 10 cm diameter. Inspired by bees’ movement, they can execute curve trajectories - circle around a tree for instance, and automatically go away when a human gently pushes them. Larger UAVs are being used for precision agriculture and infrastructure mapping, ranging between 40-100 cm diameter, and carrying sensor payloads that make the robot smart. For example, onboard cameras, inertial sensors, and GPS allow the robots to be aware of the surrounding environment and their own locations. Since smartphones already have many of the sensors used on these flying robots, a recent UAV from Kumar Lab is in fact a smartphone mounted on an aerial robot frame, enabling autonomous exploration of the environment using the phone’s computer and sensor data.

Data analytics to optimize the production chain

Back to the agricultural application, they are capable of mapping an orchard: counting the number of fruits per tree which allows an estimation of the total yield of a farm, and as a result optimizes the production chain. Moreover, the drone can build a 3D reconstruction model of plants creating an overview of the plant’s performance and helping detect nutrient deficiency or disease.

Kumar Lab, which is running on UPenn fundings and federal grants, is teaming up with growers across the country as well as Nouka Technology, a San Francisco based precision agriculture startup to test their prototypes. Collaborators include the Gallo Family Vineyard in California, Lipman Produce in Florida, Washington Tree Fruit Research Commission in Washington state, as well as University of Florida with whom Kumar Lab shares a new USDA grant focusing on use of UAV swarms for precision agriculture.

With the contribution of Jnaneshwar Das

Legal mentions © L’Atelier BNP Paribas