Document Type

Dissertation - Open Access

Award Date

2024

Degree Name

Doctor of Philosophy (PhD)

Department / School

Agronomy, Horticulture, and Plant Science

First Advisor

Anne Fennell

Abstract

Plant phenotypic features are essential for determining the structure, composition, and function of plants. These traits are controlled by both environmental variables and genetic makeup. Conventional manual methods of phenotyping, although time-consuming and not very accurate, have laid the foundation for more modern approaches such as high throughput phenotyping (HTP). HTP utilizes imaging, sensor networks, robotics, and data analytics to automate the process of collecting and analyzing data. This study utilizes hyperspectral and digital imaging techniques to quantify physiological, biochemical, and morphological characteristics in grapevines. Hyperspectral imaging was leveraged to collect spectral data for evaluating physiological factors, whereas digital imaging was used to monitor leaf morphology and biochemical features. Large-scale dataset analysis and phenotypic trend prediction were achieved by the application of machine learning and deep learning techniques. The study employed an optimized statistical methodology to quickly analyze a large number of phenotypes for genetic mapping. This followed with the identification and understanding of molecular pathways associated with reported traits. Overall, the rationale of study was to improve viticulture techniques by integrating big data and AI to increase grapevine vigor, and performance. This work aids the progress of viticulture by offering valuable understanding of how genotype, phenotype, and environmental factors interact, thereby fostering innovation and adaptability in response to challenges presented by climate change and other factors.

Publisher

South Dakota State University

Download

Available for download on Sunday, December 15, 2024

Included in

Horticulture Commons

Share

COinS
 

Rights Statement

In Copyright