Graduate Studies, UNL

 

Dissertations and Doctoral Documents from University of Nebraska-Lincoln, 2023–

First Advisor

Yufeng Ge

Degree Name

Doctor of Philosophy (Ph.D.)

Committee Members

James Schnable, Patricio Grassini, Yeyin Shi

Department

Agricultural & Biological Systems Engineering

Date of this Version

2025

Document Type

Dissertation

Citation

A dissertation presented to the faculty of the Graduate College of the University of Nebraska in partial fulfillment of requirements for the degree Doctor of Philosophy (Ph.D.)

Major: Agricultural and Biological Systems Engineering

Under the supervision of Professor

Lincoln, Nebraska, December 2025

Comments

Copyright 2025, the author. Used by permission

Abstract

Maize is a vital crop for global food security, yet quantitative characterization of its three-dimensional (3D) morphology remains a major challenge due to the geometric complexity of curved leaf structures and occlusions during data collection. This work presents an integrated framework for the digital reconstruction, parametric modeling, and functional analysis of maize leaf morphology and canopy architecture, advancing the precision and interpretability of phenotyping and modeling in smart agriculture.

First, we propose a descriptive and parametric model that represents maize leaves through three fundamental components: midrib, cross-section, and blade contour. Each is described by geometric curves and controlled by biologically meaningful parameters. This parametric representation enables scalable, configurable, and high-fidelity reconstruction of leaf morphology, facilitating downstream applications such as geometric analysis, light interception simulation, and synthetic dataset generation for phenotyping pipelines.

Second, we introduce a divide-and-conquer strategy for point cloud completion and surface reconstruction, addressing common challenges of occlusion and incompleteness in 3D plant data. The method reconstructs leaf skeletons, cross-sections, and width profiles individually before assembling them into complete 3D surfaces via gliding operations. This approach, informed by intrinsic morphological priors, achieves superior flexibility, accuracy, and interpretability compared to conventional deep learning methods, while maintaining robustness across synthetic and real-world datasets.

Finally, we explore how leaf morphology influences canopy architecture and light interception using correlation analysis and Bayesian optimization. Results demonstrate that midrib curvature, leaf width distribution, and undulation jointly regulate leaf area and canopy light distribution. The optimization framework identifies an ideal canopy structure characterized by upright upper leaves, moderately erect middle leaves, and relatively flat lower leaves, which can balance light capture and self-shading for enhanced photosynthetic efficiency.

These studies establish a unified pipeline from geometric modeling to functional optimization of maize morphology, bridging the gap between structural realism and physiological relevance. The proposed framework provides both theoretical and computational foundations for digital-twin modeling, precision phenotyping, and morphology-informed breeding strategies in maize and other crops.

Advisor: Yufeng Ge

Download

Share

COinS