Estimating grape yield on the vine from multiple images
LE3 .A278 2021
2021
Silver, Danny
Acadia University
Master of Science
Masters
Computer Science
Estimating grape yield prior to harvest is important to commercial vineyard operations. Many critical vineyard and winery decisions depend on knowing the volume of grapes to be processed while they are still hanging on the vine. The current method of yield estimation is time consuming, and depending on the experience of the viticulturist, varies in accuracy from 75-90%. This paper lays the groundwork for the development of an application to estimate grape yields based on a multiple task learning (MTL) convolutional neural network (CNN) approach that uses images captured by inexpensive smartphones secured in a simple tripod arrangement. The CNN models developed for this research use MTL transfer from autoencoders to achieve 85% and 82% accuracy from image data captured six (6) and sixteen (16) days prior to harvest, respectively.
The author retains copyright in this thesis. Any substantial copying or any other actions that exceed fair dealing or other exceptions in the Copyright Act require the permission of the author.
https://scholar.acadiau.ca/islandora/object/theses:3637
