Aarhus University Seal / Aarhus Universitets segl

Robust Species Distribution Mapping of Crop Mixtures Using Color Images and Convolutional Neural Networks

Publikation: Bidrag til tidsskrift/Konferencebidrag i tidsskrift /Bidrag til avisTidsskriftartikelForskningpeer review

Standard

Robust Species Distribution Mapping of Crop Mixtures Using Color Images and Convolutional Neural Networks. / Skovsen, Søren Kelstrup; Laursen, Morten Stigaard; Kristensen, Rebekka Kjeldgaard; Rasmussen, Jim; Dyrmann, Mads; Eriksen, Jørgen; Gislum, René; Nyholm Jørgensen, Rasmus; Karstoft, Henrik.

I: Sensors, Bind 21, Nr. 1, 175, 01.2021.

Publikation: Bidrag til tidsskrift/Konferencebidrag i tidsskrift /Bidrag til avisTidsskriftartikelForskningpeer review

Harvard

APA

CBE

MLA

Vancouver

Author

Bibtex

@article{4ccd3a972dde438bad0422186169e395,
title = "Robust Species Distribution Mapping of Crop Mixtures Using Color Images and Convolutional Neural Networks",
abstract = "Crop mixtures are often beneficial in crop rotations to enhance resource utilization and yield stability. While targeted management, dependent on the local species composition, has the potential to increase the crop value, it comes at a higher expense in terms of field surveys. As fine-grained species distribution mapping of within-field variation is typically unfeasible, the potential of targeted management remains an open research area. In this work, we propose a new method for determining the biomass species composition from high resolution color images using a DeepLabv3+ based convolutional neural network. Data collection has been performed at four separate experimental plot trial sites over three growing seasons. The method is thoroughly evaluated by predicting the biomass composition of different grass clover mixtures using only an image of the canopy. With a relative biomass clover content prediction of R2 = 0.91, we present new state-of-the-art results across the largely varying sites. Combining the algorithm with an all terrain vehicle (ATV)-mounted image acquisition system, we demonstrate a feasible method for robust coverage and species distribution mapping of 225 ha of mixed crops at a median capacity of 17 ha per hour at 173 images per hectare.",
keywords = "mixed crop mapping, species composition estimation, targeted fertilization, grass clover mixtures, proximity sensing, precision agriculture, deep learning, Mixed crop mapping, Proximity sensing, Targeted fertilization, Deep learning, Precision agriculture, Species composition estimation, Grass clover mixtures",
author = "Skovsen, {S{\o}ren Kelstrup} and Laursen, {Morten Stigaard} and Kristensen, {Rebekka Kjeldgaard} and Jim Rasmussen and Mads Dyrmann and J{\o}rgen Eriksen and Ren{\'e} Gislum and {Nyholm J{\o}rgensen}, Rasmus and Henrik Karstoft",
year = "2021",
month = jan,
doi = "10.3390/s21010175",
language = "English",
volume = "21",
journal = "Sensors",
issn = "1424-8220",
publisher = "M D P I AG",
number = " 1",

}

RIS

TY - JOUR

T1 - Robust Species Distribution Mapping of Crop Mixtures Using Color Images and Convolutional Neural Networks

AU - Skovsen, Søren Kelstrup

AU - Laursen, Morten Stigaard

AU - Kristensen, Rebekka Kjeldgaard

AU - Rasmussen, Jim

AU - Dyrmann, Mads

AU - Eriksen, Jørgen

AU - Gislum, René

AU - Nyholm Jørgensen, Rasmus

AU - Karstoft, Henrik

PY - 2021/1

Y1 - 2021/1

N2 - Crop mixtures are often beneficial in crop rotations to enhance resource utilization and yield stability. While targeted management, dependent on the local species composition, has the potential to increase the crop value, it comes at a higher expense in terms of field surveys. As fine-grained species distribution mapping of within-field variation is typically unfeasible, the potential of targeted management remains an open research area. In this work, we propose a new method for determining the biomass species composition from high resolution color images using a DeepLabv3+ based convolutional neural network. Data collection has been performed at four separate experimental plot trial sites over three growing seasons. The method is thoroughly evaluated by predicting the biomass composition of different grass clover mixtures using only an image of the canopy. With a relative biomass clover content prediction of R2 = 0.91, we present new state-of-the-art results across the largely varying sites. Combining the algorithm with an all terrain vehicle (ATV)-mounted image acquisition system, we demonstrate a feasible method for robust coverage and species distribution mapping of 225 ha of mixed crops at a median capacity of 17 ha per hour at 173 images per hectare.

AB - Crop mixtures are often beneficial in crop rotations to enhance resource utilization and yield stability. While targeted management, dependent on the local species composition, has the potential to increase the crop value, it comes at a higher expense in terms of field surveys. As fine-grained species distribution mapping of within-field variation is typically unfeasible, the potential of targeted management remains an open research area. In this work, we propose a new method for determining the biomass species composition from high resolution color images using a DeepLabv3+ based convolutional neural network. Data collection has been performed at four separate experimental plot trial sites over three growing seasons. The method is thoroughly evaluated by predicting the biomass composition of different grass clover mixtures using only an image of the canopy. With a relative biomass clover content prediction of R2 = 0.91, we present new state-of-the-art results across the largely varying sites. Combining the algorithm with an all terrain vehicle (ATV)-mounted image acquisition system, we demonstrate a feasible method for robust coverage and species distribution mapping of 225 ha of mixed crops at a median capacity of 17 ha per hour at 173 images per hectare.

KW - mixed crop mapping

KW - species composition estimation

KW - targeted fertilization

KW - grass clover mixtures

KW - proximity sensing

KW - precision agriculture

KW - deep learning

KW - Mixed crop mapping

KW - Proximity sensing

KW - Targeted fertilization

KW - Deep learning

KW - Precision agriculture

KW - Species composition estimation

KW - Grass clover mixtures

U2 - 10.3390/s21010175

DO - 10.3390/s21010175

M3 - Journal article

C2 - 33383904

VL - 21

JO - Sensors

JF - Sensors

SN - 1424-8220

IS - 1

M1 - 175

ER -