T. Wang, J. Dickinson, S. Lang, S. Khosla, and J. Wu (Canada)
Laser Scan, CAD, Modeling, Tomato Plant
This paper presents a new approach for modeling commercial tomato plants using a hybrid method integrating laser scans, CAD tools and mathematical models to yield an architecturally accurate and human maniplable tomato plant model reflecting crop growth and management over time. 3D point cloud data captured by a laser scanner is intelligently thinned and transferred to a CAD system as reference points for parametric models. A hierarchical structure is used to decompose the tomato plants into weekly-growth-sections (WGS) to relate age with plant growth geometry and solving occlusion problems. A mathematical plant growth model is deployed to define geometry and topology of the plant components, WGS and the plant as a whole. A model validation system is developed as feedback to increase accuracy.
Important Links:
Go Back
