Robotics and automation technology is being adopted to boost agricultural productivity and improve environmental sustainability. Precision agriculture through robotics can reduce the risk of environmental pollution including chemical runoff, promoting eco-friendly farming practices. One of the challenges farmers face today is determining the optimal application rate of nitrogen fertilizer, often resorting to over-fertilization to guarantee adequate farm production. While nitrogen is crucial for plant growth, excessive amounts can have negative effects on both productivity and the environment.
To address this challenge, the Center for Precision & Automated Agricultural Systems (CPAAS) at the Washington State University is developing a machine-vision solution for precision nitrogen application in apple orchards. Their goal is to assess the nitrogen state of individual trees in commercial orchards using the system and apply precise amounts of nitrogen, at the right time and locations.
Precision Nitrogen Application
All forms of nitrogen that are added to the soil eventually convert to inorganic nitrate. Nitrate is the primary form of nitrogen that plants absorb and utilize for their mineral nutrition. However, due to the negative charge of most soils, nitrate is not easily retained and can move freely within the soil. While crucial for enhancing crop yield and quality, excess free nitrate in soil poses an environmental risk.
Studies show that applying more than 50 kg of nitrogen per hectare (44 lbs per acre) as fertilizer carries risk of groundwater contamination without any corresponding production benefits. This underscores a pressing need for an improved management system to reduce excess fertilizer supplied to trees. The CPAAS is using Warthog to develop an autonomous system to optimize nitrogen application level to individual trees, improving fruit yield and quality while simultaneously minimizing environmental impact.
Mapping Tree Growth & Optimizing Nitrogen Requirements
Developing a machine-vision system to monitor tree growth and assess nitrogen requirements is a complex task, especially when it involves mapping individual trees at various growth stages and integrating the collected data into a cohesive system. The team chose a Warthog as a base platform to build their autonomous system. The robot was the ideal platform to integrate multiple sensors and to fulfill the workload needs of the project, including carrying heavy manipulators and spray tanks.
The team utilized multiple NVIDIA Jetson devices on the system as well as a GPS module for accurate positioning and navigation. The integrated UR5e manipulator arm on the system was connected with solenoid valves and a nozzle for actuating the spray system. Sensors included in the setup were Stereolabs ZED2 and Intel Realsense cameras to gather information from fruit tree canopies to accurately apply nitrogen to the root zone of the trees. The front-facing ZED2 camera was used to identify rows in the orchard and generate left and right row lines, enabling automatic centering of the Warthog. The side-facing Realsense camera was used to examine the canopy and its features. This information was used to determine the optimal spraying of nitrogen to individual trees.
Without the Warthog, the team would have to build their own platform or develop their own autonomy for use in a conventional tractor. These alternative approaches would be time-consuming and risky as compared to purchasing an off-the-shelf autonomous mobile platform.
Warthog Optimizes Orchard Operations
The team was successful in developing a system that is able to navigate between rows of orchards, identify individual trees, and accurately apply a specific amount of nitrogen to the targeted trees. The team’s ongoing efforts involve mapping each tree’s information and combining information from multiple field passes into a decision support system to create a prescription map for nitrogen at an individual tree level.
Apart from nitrogen management, Warthog is also being utilized for multiple projects that are focused on developing an automated robotics solution for orchard operations including harvesting, pruning, pollination, and flower and fruitlet thinning.
“We chose the Warthog robot due to its higher payload. It is a robust, well-supported and documented system. The robot has been around for a long time and some of our collaborators use it, so we knew what we were getting.” – Achyut Paudal, PhD Student, Center for Precision & Automated Agricultural Systems, Washington State University
The team plans to continue to make the platform more autonomous and apply it to multiple robotic projects for orchards.
The team members involved in this project consist of Manoj Karkee – Professor; Achyut Paudel, Dawood Ahmed, Ranjan Sapkota, Priyanka Updahyaya, Atif Bilal Asad, and Syed Usama – PhD Students; Safal Kshetri and Salik Khanal – Post Doctoral Scholars.
Collaborators involved in the project: Joseph R. Davidson, Cindy Grimm, and Ashley Thompson – Professors at Oregon State University, and Jostan Brown – Masters’s Student at Oregon State University.
If you would like to learn more about the Center for Precision & Automated Agricultural Systems (CPAAS), you can visit their website here.
If you would like to learn more about Warthog, you can visit our website here.
Check out some of our other recent Warthog spotlights:
