Estimating Forage Utilization with Drone-Based Photogrammetric Point Clouds

Jeffrey K. Gillan, Mitchel McClaran, Tyson L. Swetnam, Philip Heilman

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Monitoring of forage utilization typically occurs at sample locations, or key areas, selected for their presumed potential to represent utilization across pastures. However, utilization can vary greatly across landscapes and may not be well represented by traditional ground-based sampling without great effort. Remote sensing from satellite and manned airborne platforms offers spatial coverage at landscape scale, but their poor spatial resolution (satellite) and cost (manned airborne) may limit their use in monitoring forage utilization. High-resolution photogrammetric point clouds obtained from small unmanned aerial systems (sUAS) represent an appealing alternative. We developed a method to estimate utilization by observing the height reduction of herbaceous plants represented by 3-dimensional point clouds. We tested our method in a semiarid savanna in southern Arizona by comparing utilization estimates with ground-based methods after a month-long grazing duration. In six plots, we found strong correlation between imagery and ground-based estimates (r2 = 0.78) and similar average estimate of utilization of across all plots (ground-based = 18%, imagery = 20%). With a few workflow and technological improvements, we think it is feasible to estimate point cloud utilization over the entire pasture (150 ha) and potentially even larger areas. These improvements include optimizing the number of images collected and used, equipping drones with more accurate global navigation satellite systems (e.g., Global Positioning System), and processing images with cloud-based parallel processing. We show proof of concept to provide confident estimates of forage utilization patterns over large pastures and landscapes, at levels of spatial precision that are consistent with ground-based methods. The adoption of drone-based monitoring of utilization of forage on rangelands could follow the paradigm shift already demonstrated by Global Positioning Systems and Geographic information systems technologies, where the initial high computing costs were reduced, use became the norm, and the availability of more precise spatial patterns was applied to prescribe and evaluate management practices.

Original languageFrench
JournalRangeland Ecology and Management
DOIs
StatePublished - Jan 1 2019

Fingerprint

forage
pasture
global positioning systems
pastures
monitoring
imagery
GPS
application coverage
paradigm shift
GNSS
range management
herbaceous plants
methodology
rangeland
rangelands
cost
image processing
geographic information systems
savanna
herb

ASJC Scopus subject areas

  • Ecology
  • Animal Science and Zoology
  • Nature and Landscape Conservation
  • Management, Monitoring, Policy and Law

Cite this

Estimating Forage Utilization with Drone-Based Photogrammetric Point Clouds. / Gillan, Jeffrey K.; McClaran, Mitchel; Swetnam, Tyson L.; Heilman, Philip.

In: Rangeland Ecology and Management, 01.01.2019.

Research output: Contribution to journalArticle

@article{ab2f3ac8a2674be5bbcd4cb127341949,
title = "Estimating Forage Utilization with Drone-Based Photogrammetric Point Clouds",
abstract = "Monitoring of forage utilization typically occurs at sample locations, or key areas, selected for their presumed potential to represent utilization across pastures. However, utilization can vary greatly across landscapes and may not be well represented by traditional ground-based sampling without great effort. Remote sensing from satellite and manned airborne platforms offers spatial coverage at landscape scale, but their poor spatial resolution (satellite) and cost (manned airborne) may limit their use in monitoring forage utilization. High-resolution photogrammetric point clouds obtained from small unmanned aerial systems (sUAS) represent an appealing alternative. We developed a method to estimate utilization by observing the height reduction of herbaceous plants represented by 3-dimensional point clouds. We tested our method in a semiarid savanna in southern Arizona by comparing utilization estimates with ground-based methods after a month-long grazing duration. In six plots, we found strong correlation between imagery and ground-based estimates (r2 = 0.78) and similar average estimate of utilization of across all plots (ground-based = 18{\%}, imagery = 20{\%}). With a few workflow and technological improvements, we think it is feasible to estimate point cloud utilization over the entire pasture (150 ha) and potentially even larger areas. These improvements include optimizing the number of images collected and used, equipping drones with more accurate global navigation satellite systems (e.g., Global Positioning System), and processing images with cloud-based parallel processing. We show proof of concept to provide confident estimates of forage utilization patterns over large pastures and landscapes, at levels of spatial precision that are consistent with ground-based methods. The adoption of drone-based monitoring of utilization of forage on rangelands could follow the paradigm shift already demonstrated by Global Positioning Systems and Geographic information systems technologies, where the initial high computing costs were reduced, use became the norm, and the availability of more precise spatial patterns was applied to prescribe and evaluate management practices.",
author = "Gillan, {Jeffrey K.} and Mitchel McClaran and Swetnam, {Tyson L.} and Philip Heilman",
year = "2019",
month = "1",
day = "1",
doi = "10.1016/j.rama.2019.02.009",
language = "French",
journal = "Rangeland Ecology and Management",
issn = "1550-7424",
publisher = "Society for Range Management",

}

TY - JOUR

T1 - Estimating Forage Utilization with Drone-Based Photogrammetric Point Clouds

AU - Gillan, Jeffrey K.

AU - McClaran, Mitchel

AU - Swetnam, Tyson L.

AU - Heilman, Philip

PY - 2019/1/1

Y1 - 2019/1/1

N2 - Monitoring of forage utilization typically occurs at sample locations, or key areas, selected for their presumed potential to represent utilization across pastures. However, utilization can vary greatly across landscapes and may not be well represented by traditional ground-based sampling without great effort. Remote sensing from satellite and manned airborne platforms offers spatial coverage at landscape scale, but their poor spatial resolution (satellite) and cost (manned airborne) may limit their use in monitoring forage utilization. High-resolution photogrammetric point clouds obtained from small unmanned aerial systems (sUAS) represent an appealing alternative. We developed a method to estimate utilization by observing the height reduction of herbaceous plants represented by 3-dimensional point clouds. We tested our method in a semiarid savanna in southern Arizona by comparing utilization estimates with ground-based methods after a month-long grazing duration. In six plots, we found strong correlation between imagery and ground-based estimates (r2 = 0.78) and similar average estimate of utilization of across all plots (ground-based = 18%, imagery = 20%). With a few workflow and technological improvements, we think it is feasible to estimate point cloud utilization over the entire pasture (150 ha) and potentially even larger areas. These improvements include optimizing the number of images collected and used, equipping drones with more accurate global navigation satellite systems (e.g., Global Positioning System), and processing images with cloud-based parallel processing. We show proof of concept to provide confident estimates of forage utilization patterns over large pastures and landscapes, at levels of spatial precision that are consistent with ground-based methods. The adoption of drone-based monitoring of utilization of forage on rangelands could follow the paradigm shift already demonstrated by Global Positioning Systems and Geographic information systems technologies, where the initial high computing costs were reduced, use became the norm, and the availability of more precise spatial patterns was applied to prescribe and evaluate management practices.

AB - Monitoring of forage utilization typically occurs at sample locations, or key areas, selected for their presumed potential to represent utilization across pastures. However, utilization can vary greatly across landscapes and may not be well represented by traditional ground-based sampling without great effort. Remote sensing from satellite and manned airborne platforms offers spatial coverage at landscape scale, but their poor spatial resolution (satellite) and cost (manned airborne) may limit their use in monitoring forage utilization. High-resolution photogrammetric point clouds obtained from small unmanned aerial systems (sUAS) represent an appealing alternative. We developed a method to estimate utilization by observing the height reduction of herbaceous plants represented by 3-dimensional point clouds. We tested our method in a semiarid savanna in southern Arizona by comparing utilization estimates with ground-based methods after a month-long grazing duration. In six plots, we found strong correlation between imagery and ground-based estimates (r2 = 0.78) and similar average estimate of utilization of across all plots (ground-based = 18%, imagery = 20%). With a few workflow and technological improvements, we think it is feasible to estimate point cloud utilization over the entire pasture (150 ha) and potentially even larger areas. These improvements include optimizing the number of images collected and used, equipping drones with more accurate global navigation satellite systems (e.g., Global Positioning System), and processing images with cloud-based parallel processing. We show proof of concept to provide confident estimates of forage utilization patterns over large pastures and landscapes, at levels of spatial precision that are consistent with ground-based methods. The adoption of drone-based monitoring of utilization of forage on rangelands could follow the paradigm shift already demonstrated by Global Positioning Systems and Geographic information systems technologies, where the initial high computing costs were reduced, use became the norm, and the availability of more precise spatial patterns was applied to prescribe and evaluate management practices.

UR - http://www.scopus.com/inward/record.url?scp=85063724669&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85063724669&partnerID=8YFLogxK

U2 - 10.1016/j.rama.2019.02.009

DO - 10.1016/j.rama.2019.02.009

M3 - Article

AN - SCOPUS:85063724669

JO - Rangeland Ecology and Management

JF - Rangeland Ecology and Management

SN - 1550-7424

ER -