Flight Considerations and Hyperspectral Image Classifications for Dryland Vegetation Management from a Fixed-wing UAS
Abstract
Unmanned Aerial Systems (UAS)-based hyperspectral remote sensing capabilities developed by the Idaho National Lab and Boise Center Aerospace Lab were tested via demonstration flights that explored the influence of altitude on geometric error, image mosaicking, and dryland vegetation classification. The motivation for this study was to better understand the challenges associated with UAS-based hyperspectral data for distinguishing native grasses such as Sandberg bluegrass (Poa secunda) from invasives such as burr buttercup (Ranunculus testiculatus) in a shrubland environment. The test flights successfully acquired usable flightline data capable of supporting classifiable composite images. Unsupervised classification results support vegetation management objectives that rely on mapping shrub cover and distribution patterns. However, supervised classifications performed poorly despite spectral separability in the image-derived endmember pixels. In many cases, the supervised classifications accentuated noise or features in the mosaic that were artifacts of color balancing and feathering in areas of flightline overlap. Future UAS flight missions that optimize flight planning; minimize illumination differences between flightlines; and leverage ground reference data and time series analysis should be able to effectively distinguish native grasses such as Sandberg bluegrass from burr buttercup.
- Authors:
-
- Boise State Univ., ID (United States)
- Idaho National Lab. (INL), Idaho Falls, ID (United States)
- Publication Date:
- Research Org.:
- Idaho National Lab. (INL), Idaho Falls, ID (United States)
- Sponsoring Org.:
- USDOE Office of Nuclear Energy (NE)
- OSTI Identifier:
- 1469647
- Report Number(s):
- INL/JOU-16-38849-Rev000
Journal ID: ISSN 2164-7682
- Grant/Contract Number:
- AC07-05ID14517
- Resource Type:
- Accepted Manuscript
- Journal Name:
- Environmental Management and Sustainable Development
- Additional Journal Information:
- Journal Volume: 5; Journal Issue: 2; Journal ID: ISSN 2164-7682
- Publisher:
- Macrothink Institute
- Country of Publication:
- United States
- Language:
- English
- Subject:
- 58 GEOSCIENCES; Unmanned; Fixed-wing; drones; Imaging spectroscopy; Vegetation; Management; Sagebrush; Monitoring
Citation Formats
Mitchell, Jessica J., Glenn, Nancy F., Anderson, Matthew O., and Hruska, Ryan. Flight Considerations and Hyperspectral Image Classifications for Dryland Vegetation Management from a Fixed-wing UAS. United States: N. p., 2016.
Web. doi:10.5296/emsd.v5i2.9343.
Mitchell, Jessica J., Glenn, Nancy F., Anderson, Matthew O., & Hruska, Ryan. Flight Considerations and Hyperspectral Image Classifications for Dryland Vegetation Management from a Fixed-wing UAS. United States. https://doi.org/10.5296/emsd.v5i2.9343
Mitchell, Jessica J., Glenn, Nancy F., Anderson, Matthew O., and Hruska, Ryan. Sun .
"Flight Considerations and Hyperspectral Image Classifications for Dryland Vegetation Management from a Fixed-wing UAS". United States. https://doi.org/10.5296/emsd.v5i2.9343. https://www.osti.gov/servlets/purl/1469647.
@article{osti_1469647,
title = {Flight Considerations and Hyperspectral Image Classifications for Dryland Vegetation Management from a Fixed-wing UAS},
author = {Mitchell, Jessica J. and Glenn, Nancy F. and Anderson, Matthew O. and Hruska, Ryan},
abstractNote = {Unmanned Aerial Systems (UAS)-based hyperspectral remote sensing capabilities developed by the Idaho National Lab and Boise Center Aerospace Lab were tested via demonstration flights that explored the influence of altitude on geometric error, image mosaicking, and dryland vegetation classification. The motivation for this study was to better understand the challenges associated with UAS-based hyperspectral data for distinguishing native grasses such as Sandberg bluegrass (Poa secunda) from invasives such as burr buttercup (Ranunculus testiculatus) in a shrubland environment. The test flights successfully acquired usable flightline data capable of supporting classifiable composite images. Unsupervised classification results support vegetation management objectives that rely on mapping shrub cover and distribution patterns. However, supervised classifications performed poorly despite spectral separability in the image-derived endmember pixels. In many cases, the supervised classifications accentuated noise or features in the mosaic that were artifacts of color balancing and feathering in areas of flightline overlap. Future UAS flight missions that optimize flight planning; minimize illumination differences between flightlines; and leverage ground reference data and time series analysis should be able to effectively distinguish native grasses such as Sandberg bluegrass from burr buttercup.},
doi = {10.5296/emsd.v5i2.9343},
journal = {Environmental Management and Sustainable Development},
number = 2,
volume = 5,
place = {United States},
year = {2016},
month = {6}
}
Works referencing / citing this record:
Use of visible spectrum sUAS photography for land cover classification at nest sites of a declining bird species ( Falco sparverius)
journal, February 2019
- Kamm, Matthew; Reed, J. Michael
- Remote Sensing in Ecology and Conservation, Vol. 5, Issue 3
Regional Scale Dryland Vegetation Classification with an Integrated Lidar-Hyperspectral Approach
journal, September 2019
- Dashti, Hamid; Poley, Andrew; F. Glenn, Nancy
- Remote Sensing, Vol. 11, Issue 18