ARS | Publication request: TEXTURE ANALYSIS FOR CLASSIFICATION AND SEGMENTATION OF AIRBORNE IMAGES OF PLANT COVER

Submitted to: International Society for Photogrammetry and Remote Sensing Proc
Publication Type: Proceedings/Symposium
Publication Acceptance Date: December 14, 2004
Publication Date: March 7, 2005
Citation: Walthall, C.L., Pachepsky, L., Daughtry, C.S. 2005. Texture analysis for classification and segmentation of airborne images of plant cover. In: Proceedings of the 2005 American Society for Photogrammetry and Remote Sensing Annual Conference, March 7-11, 2005, Baltimore, Maryland. I:180-191.

Technical Abstract: Image texture is an element of the spatial information domain of remote sensing information. Image texture analysis is a useful tool that can provide complementary information to spectral data analysis. Crops and natural ecosystem plants have different quantitative texture characteristics that offer potential opportunities for detection, classification, and segmentation analyses. However, image texture analysis is rarely used for agricultural applications of remote sensing despite the availability of basic texture analysis routines via popular GIS and image processing systems. The availability of high spatial resolution aircraft and satellite imagery is increasing, and thus provides suitable data for texture analysis. Texture characteristics of images of various agricultural and natural plant covers imaged with hand-held cameras from various distances were analyzed as a training set for use with algorithms to conduct classification and segmentation of airborne images. Classification was successful for 80% of cases. Merits of the approach with suggestions for future directions are discussed.