Full TGIF Record # 309573
Item 1 of 1
Web URL(s):https://scisoc.confex.com/scisoc/2019am/meetingapp.cgi/Paper/121248
    Last checked: 12/04/2019
    Requires: JavaScript
Publication Type:
Content Type:Abstract or Summary only
Author(s):Hahn, Daniel
Author Affiliation:Wageningen University & Research Centre, Wageningen, Netherlands
Title:Object based image analysis of high resolution multi-spectral imagery for classifying and quantifying weeds in turfgrass areas
Section:C05 turfgrass science
Other records with the "C05 turfgrass science" Section

Turfgrass pest management poster: Diseases, insects, weeds (includes student competition)
Other records with the "Turfgrass pest management poster: Diseases, insects, weeds (includes student competition)" Section
Meeting Info.:San Antonio, Texas: November 10-13, 2019
Source:ASA, CSSA and SSSA International Annual Meetings. 2019, p. 121248.
Publishing Information:[Madison, Wisconsin]: [American Society of Agronomy, Crop Science Society of America, and Soil Science Society of America]
# of Pages:1
Keywords:TIC Keywords: Achillea millefolium; Bellis perennis; Festuca; Image analysis; Models; Multispectral analysis; Percent living ground cover; Percent weed cover; Trifolium repens
Abstract/Contents:"Accurate data collection by visual scores requires the use of experienced assessors to reduce variability in data recording. As an alternative approach, we conducted a study to determine the composition of 136 plots sown with six Festuca spp. cultivars and three weed species (Trifolium repens L., Bellis perennis L., Achillea millefolium L.) with object based image analysis (OBIA). We collected over 600 overlapping multispectral images of a field trial with a Parrot Sequoia (Parrot Drones SAS, Paris, France) camera. The images were taken from 3 m above ground at four spectral bands: green (530-570 nm), red (640-680 nm), red edge (730-740 nm) and near infrared (770-810 nm). An orthomosaic of the study area was created in Agisoft Metashape and OBIA and a Random Forest Model (RFM) were used to classify the complete area of the field experiment in three classes (grass, weed, and soil) and five classes (clover, daisy, grass, soil, and yarrow), respectively. Separation into three classes produced an overall accuracy of 92%, whereas separation into five classes resulted in an overall accuracy of 77%. We used the classification results to calculate the cover percentage of each class per plot and compared this to results obtained with the point quadrant method and with vegetation cover results processed in software package 'turf analyzer'. Three classes (r=0.79, p<0.0001) and five classes (r= 0.74, p< 0.0001) correlated well with results of cover. Weed cover correlated well with three (r= 0.63, p<0.0001) and five classes (r= 0.64, p<0.0001) compared to point quadrant estimations."
Note:This item is an abstract only!
"Poster #1637"
ASA/CSSA/SSSA Citation (Crop Science-Like - may be incomplete):
Hahn, D. 2019. Object based image analysis of high resolution multi-spectral imagery for classifying and quantifying weeds in turfgrass areas. Agron. Abr. p. 121248.
Fastlink to access this record outside TGIF: https://tic.msu.edu/tgif/flink?recno=309573
If there are problems with this record, send us feedback about record 309573.
Choices for finding the above item:
Web URL(s):
    Last checked: 12/04/2019
    Requires: JavaScript
Find from within TIC:
   Digitally in TIC by record number.
Request through your local library's inter-library loan service (bring or send a copy of this TGIF record)