A performance comparison of RGB, NIR, and depth images in immature citrus detection using deep learning algorithms for yield prediction

Daeun Choi, Won Suk Lee, John K. Schueller, Reza Ehsani, Fritz Roka, Justice Diamond

Research output: Contribution to conferencePaper

2 Citations (Scopus)

Abstract

Yield forecasting is important for farm management. In this study, red, green, and blue (RGB), near-infrared (NIR), and depth sensors were implemented in an outdoor machine vision system to determine the number of immature citrus in tree canopies in a citrus grove. The main objective was to compare the performances of three image data types for citrus yield forecasting. The performance comparison was conducted with two machine vision algorithm steps: 1) circular object detection for potential fruit areas and 2) classification of citrus fruit from the background. For circular object detection, circular Hough transform was used in the RGB and NIR images. For the depth images, CHOI's Circle Estimation ('CHOICE') algorithm was developed using depth divergence and vorticity to find circular objects in the depth images. The classification process was conducted using AlexNet, a deep learning algorithm for all three image types. The implementation of a convolutional neural network allowed the machine vision algorithms to remain bias-free process during feature generation and selection. NIR images performed best with 96% true positive rate for both the circular object detection and classification. A machine vision system using this image type will produce a more objective yield prediction with a higher accuracy than other types.

Original languageEnglish (US)
DOIs
StatePublished - Jan 1 2017
Event2017 ASABE Annual International Meeting - Spokane, United States
Duration: Jul 16 2017Jul 19 2017

Other

Other2017 ASABE Annual International Meeting
CountryUnited States
CitySpokane
Period7/16/177/19/17

Fingerprint

computer vision
Learning algorithms
Computer vision
Citrus
learning
immatures
Infrared radiation
prediction
Citrus fruits
Hough transforms
farm management
citrus fruits
Fruits
Vorticity
Farms
neural networks
orchards
canopy
Neural networks
fruits

All Science Journal Classification (ASJC) codes

  • Bioengineering
  • Agronomy and Crop Science

Cite this

Choi, D., Lee, W. S., Schueller, J. K., Ehsani, R., Roka, F., & Diamond, J. (2017). A performance comparison of RGB, NIR, and depth images in immature citrus detection using deep learning algorithms for yield prediction. Paper presented at 2017 ASABE Annual International Meeting, Spokane, United States. https://doi.org/10.13031/aim.201700076
Choi, Daeun ; Lee, Won Suk ; Schueller, John K. ; Ehsani, Reza ; Roka, Fritz ; Diamond, Justice. / A performance comparison of RGB, NIR, and depth images in immature citrus detection using deep learning algorithms for yield prediction. Paper presented at 2017 ASABE Annual International Meeting, Spokane, United States.
@conference{b2ff4fb613ed495a9ef849f5bfe113be,
title = "A performance comparison of RGB, NIR, and depth images in immature citrus detection using deep learning algorithms for yield prediction",
abstract = "Yield forecasting is important for farm management. In this study, red, green, and blue (RGB), near-infrared (NIR), and depth sensors were implemented in an outdoor machine vision system to determine the number of immature citrus in tree canopies in a citrus grove. The main objective was to compare the performances of three image data types for citrus yield forecasting. The performance comparison was conducted with two machine vision algorithm steps: 1) circular object detection for potential fruit areas and 2) classification of citrus fruit from the background. For circular object detection, circular Hough transform was used in the RGB and NIR images. For the depth images, CHOI's Circle Estimation ('CHOICE') algorithm was developed using depth divergence and vorticity to find circular objects in the depth images. The classification process was conducted using AlexNet, a deep learning algorithm for all three image types. The implementation of a convolutional neural network allowed the machine vision algorithms to remain bias-free process during feature generation and selection. NIR images performed best with 96{\%} true positive rate for both the circular object detection and classification. A machine vision system using this image type will produce a more objective yield prediction with a higher accuracy than other types.",
author = "Daeun Choi and Lee, {Won Suk} and Schueller, {John K.} and Reza Ehsani and Fritz Roka and Justice Diamond",
year = "2017",
month = "1",
day = "1",
doi = "10.13031/aim.201700076",
language = "English (US)",
note = "2017 ASABE Annual International Meeting ; Conference date: 16-07-2017 Through 19-07-2017",

}

Choi, D, Lee, WS, Schueller, JK, Ehsani, R, Roka, F & Diamond, J 2017, 'A performance comparison of RGB, NIR, and depth images in immature citrus detection using deep learning algorithms for yield prediction', Paper presented at 2017 ASABE Annual International Meeting, Spokane, United States, 7/16/17 - 7/19/17. https://doi.org/10.13031/aim.201700076

A performance comparison of RGB, NIR, and depth images in immature citrus detection using deep learning algorithms for yield prediction. / Choi, Daeun; Lee, Won Suk; Schueller, John K.; Ehsani, Reza; Roka, Fritz; Diamond, Justice.

2017. Paper presented at 2017 ASABE Annual International Meeting, Spokane, United States.

Research output: Contribution to conferencePaper

TY - CONF

T1 - A performance comparison of RGB, NIR, and depth images in immature citrus detection using deep learning algorithms for yield prediction

AU - Choi, Daeun

AU - Lee, Won Suk

AU - Schueller, John K.

AU - Ehsani, Reza

AU - Roka, Fritz

AU - Diamond, Justice

PY - 2017/1/1

Y1 - 2017/1/1

N2 - Yield forecasting is important for farm management. In this study, red, green, and blue (RGB), near-infrared (NIR), and depth sensors were implemented in an outdoor machine vision system to determine the number of immature citrus in tree canopies in a citrus grove. The main objective was to compare the performances of three image data types for citrus yield forecasting. The performance comparison was conducted with two machine vision algorithm steps: 1) circular object detection for potential fruit areas and 2) classification of citrus fruit from the background. For circular object detection, circular Hough transform was used in the RGB and NIR images. For the depth images, CHOI's Circle Estimation ('CHOICE') algorithm was developed using depth divergence and vorticity to find circular objects in the depth images. The classification process was conducted using AlexNet, a deep learning algorithm for all three image types. The implementation of a convolutional neural network allowed the machine vision algorithms to remain bias-free process during feature generation and selection. NIR images performed best with 96% true positive rate for both the circular object detection and classification. A machine vision system using this image type will produce a more objective yield prediction with a higher accuracy than other types.

AB - Yield forecasting is important for farm management. In this study, red, green, and blue (RGB), near-infrared (NIR), and depth sensors were implemented in an outdoor machine vision system to determine the number of immature citrus in tree canopies in a citrus grove. The main objective was to compare the performances of three image data types for citrus yield forecasting. The performance comparison was conducted with two machine vision algorithm steps: 1) circular object detection for potential fruit areas and 2) classification of citrus fruit from the background. For circular object detection, circular Hough transform was used in the RGB and NIR images. For the depth images, CHOI's Circle Estimation ('CHOICE') algorithm was developed using depth divergence and vorticity to find circular objects in the depth images. The classification process was conducted using AlexNet, a deep learning algorithm for all three image types. The implementation of a convolutional neural network allowed the machine vision algorithms to remain bias-free process during feature generation and selection. NIR images performed best with 96% true positive rate for both the circular object detection and classification. A machine vision system using this image type will produce a more objective yield prediction with a higher accuracy than other types.

UR - http://www.scopus.com/inward/record.url?scp=85035356469&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85035356469&partnerID=8YFLogxK

U2 - 10.13031/aim.201700076

DO - 10.13031/aim.201700076

M3 - Paper

AN - SCOPUS:85035356469

ER -

Choi D, Lee WS, Schueller JK, Ehsani R, Roka F, Diamond J. A performance comparison of RGB, NIR, and depth images in immature citrus detection using deep learning algorithms for yield prediction. 2017. Paper presented at 2017 ASABE Annual International Meeting, Spokane, United States. https://doi.org/10.13031/aim.201700076