- You are here:
-
-
All publications
- Acoustic and Hyperspectral Imaging Sensing For Nondestructive Insect Infestation Detection in Apples
All publications from your selection
Acoustic and Hyperspectral Imaging Sensing For Nondestructive Insect Infestation Detection in Apples
Authors: Akinbode Adedeji, Nader Ekramirad, Alfadhl khaled, Chad Parrish, Kevin Donohue, Raul Villanueva
Identifier: CSBE22157
Download file:
Published in: CSBE-SCGAB Technical Conferences » AGM Charlottetown 2022
Download RAW file:
DOI:
Description: Apple is a very important fruit. About $5 billion worth is produced annually in the US. One-third of this is exported to countries with zero tolerance for organisms of quarantine concerns. Codling moth (CM) is the most devastating pest of apples and its detection through non-destructive means is advantageous for the industry – more effective in reducing loss, and rapid. This study evaluated two non-destructive sensing methods, acoustic (passive [0.4-8 Hz] and impulse [35-100 kHz]) emission (AE) and hyperspectral imaging (HSI) (1000 – 1700 nm) to build models that delineate between data obtained from healthy and infested apples. HSI leveraged the capability of infrared energy that penetrates some depth into the apples to provide reflectance that contains fingerprint of the constituent while the acoustic methods measured either larvae activity or attenuated signals from within damaged apples. Two AE methods were deployed: vibro-acoustic (VAE) method that collects emission from CM activities and acoustic-impulse (UE) for attenuated sound through the thickness of apples damaged through the calyx beyond the HSI depth. Machine learning classifiers were used to calibrate, validate, and reduce features in the data. The first two steps were repeated after feature selection. The results obtained show that AE methods are capable of classifying CM-infested apples up to 99% and 97%, respectively. HSI accuracy is 91.6% and is capable of detecting surface CM eggs over AE method. Both sensing methods provide high classification accuracy. Future work will focus on fusing both sensors data for joint deployment to harness the merits of both methods.
Keywords: Apples, Acoustic, Nondestructive Testing, Hyperspectral Imaging, Machine Learning
Résumé:
Mots-clés:
Citation:
Volume:
Issue:
Pages -
Contributor:
Date: 2022-07-24
Technical field:
Conference name: CSBE/SCGAB 2022 Annual Conference, Charlottetow, PEI, 24-27 July 2022.
Session name: Agricultural Engineering3
Other information:
Type: Text.Abstract
Format:
Publication type: Presentation
Source:
Relation:
Coverage: United States of America
Language 1: en
Language 2:
Rights: Canadian Society for Bioengineering
Notes:
Other files:
DOI:
Description: Apple is a very important fruit. About $5 billion worth is produced annually in the US. One-third of this is exported to countries with zero tolerance for organisms of quarantine concerns. Codling moth (CM) is the most devastating pest of apples and its detection through non-destructive means is advantageous for the industry – more effective in reducing loss, and rapid. This study evaluated two non-destructive sensing methods, acoustic (passive [0.4-8 Hz] and impulse [35-100 kHz]) emission (AE) and hyperspectral imaging (HSI) (1000 – 1700 nm) to build models that delineate between data obtained from healthy and infested apples. HSI leveraged the capability of infrared energy that penetrates some depth into the apples to provide reflectance that contains fingerprint of the constituent while the acoustic methods measured either larvae activity or attenuated signals from within damaged apples. Two AE methods were deployed: vibro-acoustic (VAE) method that collects emission from CM activities and acoustic-impulse (UE) for attenuated sound through the thickness of apples damaged through the calyx beyond the HSI depth. Machine learning classifiers were used to calibrate, validate, and reduce features in the data. The first two steps were repeated after feature selection. The results obtained show that AE methods are capable of classifying CM-infested apples up to 99% and 97%, respectively. HSI accuracy is 91.6% and is capable of detecting surface CM eggs over AE method. Both sensing methods provide high classification accuracy. Future work will focus on fusing both sensors data for joint deployment to harness the merits of both methods.
Keywords: Apples, Acoustic, Nondestructive Testing, Hyperspectral Imaging, Machine Learning
Résumé:
Mots-clés:
Citation:
Volume:
Issue:
Pages -
Contributor:
Date: 2022-07-24
Technical field:
Conference name: CSBE/SCGAB 2022 Annual Conference, Charlottetow, PEI, 24-27 July 2022.
Session name: Agricultural Engineering3
Other information:
Type: Text.Abstract
Format:
Publication type: Presentation
Source:
Relation:
Coverage: United States of America
Language 1: en
Language 2:
Rights: Canadian Society for Bioengineering
Notes:
Other files:
©
2023
CSBE-SCGAB