Kaundanya, Adwait Anand2024-12-202024-12-202024-12-19vt_gsexam:42299https://hdl.handle.net/10919/123851Affordable and readily available automation options for plant research remain scarce, however with the availability of such a system, many research tasks can be streamlined. In this project, we demonstrate a prototype of such an open-source, low-cost, heterogeneous robotic system called Mini T-Rex. We combine two over-the-counter robots and leverage the ROS2 framework to control this heterogeneous system. This system provides a unique advantage of sensor-to-plant method to capture multi-view images at any angle and distance within the workspace. We demonstrate how making a digital twin in ROS2 can help to control a heterogeneous system via abstracted hardware control. We also talk about I2GROW Oasis which is a robotic system consisting of a remotely controlled robot with the ability to capture top-view images. In this thesis we describe the hardware and software design of both these robotic systems. To use this robotic system, the plants can be grown on a growth bed or a hydroponic system below the Mini T-Rex robot, and the camera will approach the plant without any contact with the plants due to the precise control of the robotic manipulator. We used the system to capture several large data sets of 3D phenotypic data for Solanum lycopersicum, Lactuca sativa, and Thlaspi. In conclusion, we have developed a 9-degree of freedom, fully open-source heterogeneous robotic system capable of multi-view, camera-to plant image capture for plant 3D model reconstruction called Mini T-Rex. We show how to use gantry like robots for phenotyping and create longitudinal datasets by automating these high precision robotic systems.ETDenIn CopyrightROS2MoveIt2motion planningFarmBotplant phenotyping3D reconstructionroboticsThesis