Skip to main content
ARS Home » Research » Publications at this Location » Publication #208685
Title: Mapping Wild Taro with Color-infrared Aerial Photography and Image Processing

Author
item Everitt, James
item Yang, Chenghai
item Davis, Michael

Submitted to: Journal of Aquatic Plant Management
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 3/5/2007
Publication Date: 7/1/2007
Citation: Everitt, J.H., Yang, C., Davis, M.R. 2007. Mapping wild taro with color-infrared aerial photography and image processing. Journal of Aquatic Plant Management. 45:106-110.

Interpretive Summary: The invasion and spread of weeds in wetlands is a major deterrent to management of these areas. Wild taro is an exotic ornamental plant that has escaped cultivation and invaded many freshwater wetlands in the southeastern United States where it has reduced the diversity of native vegetation. A study was conducted on the Rio Grande in southwest Texas to determine the potential of using remote sensing technology to distinguish wild taro. Field reflectance measurements showed that wild taro had distinct visible and near-infrared reflectance. Color-infrared aerial photography and supervised image analysis were used to distinguish and map wild taro infestations. Accuracy assessments performed on classification maps of photographs from three sites had producer’s and user’s accuracies ranging from 83.3% to 100%. These results should be of interest to weed specialists and wetland resource managers.

Technical Abstract: Wild taro [Colocasia esculenta (L.) Schott.] is an exotic ornamental plant that has escaped cultivation and invaded many freshwater wetlands in the southeastern United States. Remote sensing techniques were evaluated for distinguishing wild taro along the Rio Grande in southwest Texas. Field reflectance measurements showed that wild taro had significantly different (p = 0.05) visible and near-infrared reflectance from associated plant species. Wild taro could be distinguished on color-infrared photographs where it had a bright red image response. Supervised image analysis techniques were used to classify the imagery. Accuracy assessments performed on classification maps of photographs from three sites had producer’s and user’s accuracies ranging from 83.3% to 100%.