Authors: Patrick J. Hennessy, Travis J. Esau, Arnold W. Schumann, Aitazaz A. Farooque, Qamar U. Zaman, Scott N. White
Identifier: CSBE23249
Download file:
Published in: CSBE-SCGAB Technical Conferences » AGM Lethbridge 2023

Download RAW file:
Description: Wild blueberries (Vaccinium angustifolium Ait.) are a perennial crop in northeastern North America. Conventional management relies on costly broadcast applications of agrochemicals such as herbicides. There is an opportunity to reduce agrochemical usage through the implementation of targeted spray applications using machine vision systems. Machine vision techniques such as deep learning convolutional neural networks depend on high-quality images for decision-making. Cameras with rolling shutters produce blurring and warping when capturing images in motion, necessitating the use of more expensive cameras with global shutters. This study examines four cameras for potential use on a machine vision smart sprayer in wild blueberry. Two cameras from FLIR (Blackfly S, Firefly DL) capture images using a global shutter, while cameras from Logitech (c920) and Luxonis (OAK-D) use rolling shutters. Field tests were conducted in commercial wild blueberry fields located in central and northern Nova Scotia, Canada, from late April to early June 2021, 2022, and 2023. The cameras were affixed to the boom of a prototype agricultural sprayer, up to 0.91 m above the ground with their lenses pointed downward. Images were captured while the sprayer moved at speeds up to 2.7 m/s. The images were assessed for blur using a Laplacian kernel. The results of this analysis will be presented at conference. Future work will involve the development of a machine vision smart sprayer to target weeds in wild blueberry fields. By using a machine vision smart sprayer, growers can achieve substantial cost savings by selectively applying agrochemicals based on real-time visual data.

Keywords: Precision agriculture, deep learning, convolutional neural network, Laplacian kernel, mechanized systems
Pages -
Date: 2023-07-23
Technical field:
Conference name: CSBE/SCGAB 2023 Annual Conference, Lethbridge, Alberta, 23-26 July 2023.
Session name:

Other information:
Type: Presentation
Publication type: Text.Abstract
Coverage: North_America
Language 1: en
Language 2:
Rights: Canadian Society for Bioengineering
Other files: