Author ORCID Identifier

https://orcid.org/0000-0003-4172-3451

Semester

Spring

Date of Graduation

2025

Document Type

Thesis

Degree Type

MS

College

Statler College of Engineering and Mineral Resources

Department

Mechanical and Aerospace Engineering

Committee Chair

Jason Gross

Committee Co-Chair

Yu Gu

Committee Member

Yu Gu

Committee Member

Cagri Kilic

Abstract

The use of automated systems for agriculture is integral to keeping the food supply secure. Both industry and academia are exploring and applying methods to increase the yield of plants in environments ranging from outdoor fields to greenhouses. Specifically, many automated systems use continuous monitoring of plants to track plant health and yield. The use of computer vision is necessary when it comes to precision operations that use robotics. Today, robots are trained to weed, harvest, and pollinate. To accomplish these tasks autonomously, a lot of data is needed, which is where spatial-temporal observations of the plants are being recorded using cameras. Flowering plants like those of the bramble family produce terminal clusters of flowers that yield fruits. In robotics, accurately matching these flower clusters for precision flower pollination is particularly challenging. Through the use of vision sensors, visual data can be obtained. However, plant growth and external effects like manipulation, wind, and even light conditions can increase the challenge of matching. Additionally, with limited computation on board a mobile robot, the algorithm needs to be feasible for real-time operation in the use case of robotics pollination. This thesis explores the use of the Unscented Transform and descriptors in MATLAB to perform cluster matching based on visual data. Using a robot equipped with an RGB-D camera, the positions of the flowers in the cluster can be obtained using a vision model and transformed through a descriptor function. The simulated results are evaluated and validated using a Monte Carlo simulation. The experimental results are evaluated for matching based on the collected datasets. The simulation and experimental data results show that the proposed algorithm is a robust method for cluster matching and feasible for real-time application in precision robotic pollination.

Download

Share

COinS