By Scout Nelson
Artificial intelligence is revolutionizing agriculture, with researchers exploring ways to automate manual tasks like planting, weeding, and harvesting. A project at South Dakota State University (SDSU) is applying AI-driven technology to robotic chili pepper harvesting, paving the way for improved agricultural automation.
A research team at SDSU’s Machine Vision and Optical Sensor Lab is developing AI models to assist robots in identifying and harvesting chili peppers. The project, led by Assistant Professor Pappu Yadav, focused on three chili pepper varieties and was titled “AI-Driven Computer Vision for Detection and Pose Estimation of Chili Peppers for Robotic Harvesting.”
A key contributor to the project was a Future Innovators of America Fellow, who worked on the AI aspects of the study. The fellowship, part of the Jerome J. Lohr College of Engineering, provides undergraduate students with research opportunities and funding to enhance their learning.
AI Model Development
The research involved capturing 1,000 images of peppers at different growth stages and orientations, scanning them into a computer, and creating an AI model to direct robotic harvesting. The model’s performance showed 79% accuracy in detecting ripeness for Hungarian hot wax and poblano peppers and 53% accuracy for black Hungarian peppers. Orientation detection was highly accurate, reaching 89-90% consistency.
Future Improvements and Applications
To extend robotic harvesting beyond greenhouse environments, further algorithm improvements are needed to adjust for outdoor lighting conditions and enhance robot stability.
A research paper based on the project will be presented at the Society of Photographic Instrumentation Engineers Defense + Commercial Sensing Conference in Orlando, Florida, in April 2025.
The project demonstrates how AI can streamline agricultural tasks, enhance precision farming, and reduce manual labor in harvesting specialty crops. Researchers continue to refine the technology, aiming for greater efficiency and accuracy in automated farming solutions.
Photo Credits:South Dakota State University
Categories: South Dakota, Education