Full TGIF Record # 225014
Item 1 of 1
Web URL(s):http://www.swss.ws/wp-content/uploads/docs/2004%20Proceedings-SWSS.pdf#page=443
    Last checked: 07/17/2013
    Requires: PDF Reader
    Notes: Item is within a single large file
Publication Type:
i
Report
Author(s):Hutto, K. C.; Shaw, D. R.; Byrd, J. D. Jr.; Taylor, J. M.; Gray, C. J.
Author Affiliation:Department of Plant and Soil Sciences, Mississippi State University, Mississippi State, MS
Title:Hyperspectral radiometry to identify turfgrass stress and weed species
Section:Posters
Other records with the "Posters" Section
Meeting Info.:Memphis, Tennessee: January 26-28, 2004
Source:Proceedings: Southern Weed Science Society: 57th Annual Meeting. Vol. 57, 2004, p. 345.
# of Pages:1
Publishing Information:Champaign, Illinois: Southern Weed Science Society
Keywords:TIC Keywords: Evaluative methods; Hyperspectral radiometry; Species identification; Stress; Weed identification
Abstract/Contents:"Research was conducted in summer 2002 and 2003 at the Plant Science Research Center, Starkville, MS to identify spectral reflectance characteristics of certain warm season turfgrass weeds, as well as to determine the accuracy of hyperspectral radiometry to distinguish between various turfgrass stresses. The weed species analyzed were large crabgrass [Digitaria sanguinalis (L.) Scop.], dallisgrass (Paspalum dilatatum Poir.), Virginia buttonweed (Diodia virginiana L.), and eclipta (Eclipta prostrata L.). The simulated stresses were traffic (three levels), herbicide application (2.0 lb ai/A MSMA or 3 pt/A Trimec Classic), and moisture stress. Major warm season turfgrasses were also evaluated, which consisted of 'Tifway' bermudagrass (Cynodon dactylon Pers. x transvaalensis Burrt-Davy), 'Meyer' zoysiagrass (Zoysia japonica Steud.), centipedegrass [Eremochloa ophiuroides (Munro) Hack.], and St. Augustinegrass [Stenotaphrum secundatum (Walt.) Kuntze]. Data were analyzed using receiver operating characteristics-best spectral band combination with a feature classification of maximum likelihood to identify best spectral bands used to separate all classes. These bands were subjected to stepwise linear regression to identify the most significant bands attributing to the classification of species and stresses. Discriminant analysis was used to obtain classification accuracies. The 2003 data was validated against the 2002 data to evaluate the accuracy of classification methods over years. An overall classification accuracy of 99% using the 15 most weighted bands (ranging from 350 to 439 nm, and 2327 nm) was achieved for all weed species. Using discriminant analysis, an overall accuracy of 87% was achieved using 13 of 15 bands (excluding 376 and 378 nm). Only Virginia buttonweed was classified accurately (96%) when validating the method over years. All other weeds were classified below 57%. An overall accuracy of 63% was achieved using bands in the near infrared region (NIR) (ranging from 746 to 841 nm) for all turfgrass stresses with moisture stress being the most accurately classified (94%). Discriminant analysis eliminated bands 816 and 824 nm, classifying moisture stress correctly 88%. The validation analysis classified moisture stress correctly 43%. Turfgrass species were all correctly classified using bands in the visible (ranging from 386 to 405 nm) and NIR (ranging from 725 to 764 nm). All turfgrass species, with the exception of centipedegrass, were correctly classified 93% or greater validating over years (excluding 389 and 728 nm). An overall accuracy of 80% was achieved for all species analyzed using bands in the visible region (ranging from 350 to 389 nm). Discriminant analysis provided an overall accuracy of 50% (excluding bands 356, 357, 358, 361, and 369 nm). Validation analysis provided an overall accuracy of 57% for all species. Bermudagrass (71%), centipedegrass (77%), and dallisgrass (70%) were the most accurately classified species over years."
Language:English
References:0
Note:"The changing world of weed science"
This item is an abstract only!
ASA/CSSA/SSSA Citation (Crop Science-Like - may be incomplete):
Hutto, K. C., D. R. Shaw, J. D. Jr. Byrd, J. M. Taylor, and C. J. Gray. 2004. Hyperspectral radiometry to identify turfgrass stress and weed species. South. Weed Sci. Soc. Proc. 57:p. 345.
Fastlink to access this record outside TGIF: https://tic.msu.edu/tgif/flink?recno=225014
If there are problems with this record, send us feedback about record 225014.
Choices for finding the above item:
Web URL(s):
http://www.swss.ws/wp-content/uploads/docs/2004%20Proceedings-SWSS.pdf#page=443
    Last checked: 07/17/2013
    Requires: PDF Reader
    Notes: Item is within a single large file
Find Item @ MSU
MSU catalog number: b2207931
Find from within TIC:
   Digitally in TIC by file name: swssp2004
Request through your local library's inter-library loan service (bring or send a copy of this TGIF record)