Submitted to: Subtropical Plant Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/27/2009
Publication Date: 12/20/2009
Citation: Everitt, J.H., Summy, K.R., Glomski, L.M., Owens, C.S., Yang, C. 2009. Spectral reflectance and digital image relations among five aquatic weeds. Geocarto International. 61:15-23. Interpretive Summary: Improved techniques are needed to obtain spectral light reflectance measurements and digital imaging under all weather conditions. An artificial quartz lighting source was evaluated under laboratory conditions to acquire reflectance measurements and close range color-infrared and conventional color digital imagery for distinguishing among leaves and leaves/stems of five aquatic weed species: Eurasian watermilfoil, hydrilla, parrotfeather, waterhyacinth, and waterlettuce. Reflectance measurements were studied at five wavelengths: 450 nm, 550 nm, 650 nm, 680 nm, and 850 nm. Reflectance values differed significantly among the species at all five wavelengths, but the best separations occurred at the 550 nm, 650 nm, 680 nm, and 850 nm. Color-infrared and conventional color images of foliage of the five species showed that they differed in image tonal response. Supervised image analysis techniques performed on both types of imagery demonstrated that the computer did an adequate job in identifying the image tonal responses of the weed species. These findings showed that an artificial lighting source can be used successfully to facilitate acquisition of remote sensing data under all weather conditions and should provide insight into using this technology for distinguishing among noxious weeds.
Technical Abstract: This study reports on the use of an artificial quartz halogen lighting source to facilitate the acquisition of spectral light reflectance measurements and digital imaging of invasive aquatic weeds. Spectral leaf or leaf/stem reflectance measurements were made on five aquatic weeds: Eurasian watermilfoil (Myriophyllum spicatum L.), hydrilla [Hydrilla verticillata (L. F.) Royle], parrotfeather [Myriophyllum aquaticum (Vall.), waterhyacinth [Eichhornia crassipes (Mart.) Solms], and waterlettuce (Pistia stratiotes L.). Reflectance measurements were studied at five wavelengths of the electromagnetic spectrum: 450 nm (visible blue), 550 nm (visible green), 650 nm (visible red), 680 nm (visible red edge), and 850 nm (near-infrared). Reflectance values differed significantly (P= 0.05) among the species at all five wavelengths. However, more distinct separations among species occurred at the 550 nm, 650 nm, 680 nm, and 850 nm wavelengths. Reflectance differences among species were attributed to variable foliage coloration and vegetative density. Close range conventional color and color-infrared digital images of leaves or leaves/stems of the five species showed they differed in image tonal response. Reflectance measurements were related to the image tonal response of the plant species on both types of imagery. Supervised image classifications performed on both conventional color and color-infrared images showed the computer generally did an adequate job in identifying the image tonal responses of the weed species.