Crop Yield Prediction using Image and Microbiome features
Date
2022-12-08
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
ORCID
0000-0002-9172-1262
Type
Thesis
Degree Level
Masters
Abstract
Predicting crop yield earlier in the developmental cycle could aid breeders in accelerating crop discovery,
and farmers, regions, and nations better anticipate crop yields. Numerous approaches have been proposed to predict these outcomes. Many researchers have used UAV image-derived plant phenotypes to predict yield. However, this approach only covers the plant’s above-ground portion. Other research has focused on the below-ground portion by using microbiome assays of these constituents. This thesis presents a model that utilizes features from both the above-ground and below-ground activity using feature selection and linear
regression. This model could predict the amount of crop anticipated under stable conditions (yield) and the amount of crop anticipated under variable conditions (yield stability) with an R2 of between 91% and 81% in just the third week after planting.
Description
Keywords
Feature Selection, Yield Prediction, Microbiome Features, Image Features
Citation
Degree
Master of Science (M.Sc.)
Department
Computer Science
Program
Computer Science