Skip to main content
ARS Home » Research » Publications at this Location » Publication #197203
Title: USING AIRBORNE HYPERSPECTRAL IMAGERY TO IDENTIFY LOCAL INFESTATION OF MUSK THISTLE, CARDUUS NUTANS L., FROM CO-OCCURING VEGETATION

Author
item MIRIK, MUSTAFA - TEXAS AGRIC EXP STATION
item MICHELS, JR., GERALD - TEXAS AGRIC EXP STATION
item KASSYMZHANOVA-MIRIK, SABINA - TEXAS AGRIC EXP STATION
item JONES, DAVID - TEXAS AGRIC EXP STATION
item Elliott, Norman - Norm

Submitted to: Biennial Workshop on Aerial Photography, Videography, and High Resolution Digital Imagery for Resource Assessment Proceedings
Publication Type: Proceedings
Publication Acceptance Date: 9/1/2005
Publication Date: 11/1/2005
Citation: Mirik, M., Michels, Jr., G.J., Kassymzhanova-Mirik, S., Jones, D., Elliott, N.C. 2005. Using airborne hyperspectral imagery to identify local infestation of musk thistle, Carduus nutans L., from co-occuring vegetation. In: Proceedings of the 20th Biennial Workshop on Aerial Photography, Videography, and High Resolution Digital Imagery for Resource Assessment, October 4-6, 2005, Weslaco, Texas. 2005 CDROM.

Interpretive Summary: Invasion by noxious weed species presents a serious threat to natural habitat. Identifying the noxious weeds in natural lands can improve monitoring, planning, and management practices. Methods for reliable, repeatable, quick, and cost effective mapping of invasive weeds are needed to facilitate these practices. Remote sensing has been used to map various plant species including invasive and noxious weeds. Musk thistle, a noxious weed, which is a good candidate for detection by remote sensing because it may reflect incoming solar radiation (light) in a distinctive pattern due its large, purple-reddish flower head. We used imagery acquired by an airborne imaging system that segments the spectrum into narrow wavelength bands (called hyperspectral remote sensing) to map musk thistle infestation in a pasture in Texas. We found that the imagery could differentiate musk thistle from other rangeland plant species with a high degree of accuracy (80%) as verified with ground survey. These results demonstrate the value of hyperspectral remote sensing data for mapping certain noxious weed species within the natural habitats they threaten.

Technical Abstract: Invasion by noxious weed species presents a serious threat to the remaining fragments of the natural habitat. Identifying the population dynamics and extent of spread of noxious weeds in a temporal and spatial perspective improves monitoring, planning, and management practices. Methods for reliable, repeatable, quick, and cost effective mapping of invasion patterns are needed to facilitate these practices. Remote sensing has been used to map various plant species including invasive and noxious weeds. Musk thistle, Carduus nutans L., a noxious weed, is a good candidate for detection by remote sensing platforms because it may produce a unique spectral signature due to a large, purple-reddish flower head. Therefore, airborne hyperspectral imageries acquired at two dates were used to map musk thistle infestation in a pasture, TX, in the second week of April when musk thistle was at the rosette form and mid June when musk thistle was at the flowering stage in 2003. Imageries were classified using the supervised maximum likelihood classifier technique. Overall accuracy was greater than 80% that was verified with ground survey for both imageries. These results demonstrate the value of hyperspectral data for mapping noxious weed species and the habitats they threaten.