Towards a Novel Approach for Texture Segmentation of SAR Sea Ice Imagery

• Main Authors: David Clausi, M. Ed Jernigan
• Other Authors: The Pennsylvania State University CiteSeerX Archives
• Format: Text
• Language: English
• Subjects: Sea ice
• Online Access: http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.52.3397; http://monet.uwaterloo.ca/~dclausi/crss96_gabor.ps

: Texture is an important aspect of identifying sea ice types in SAR imagery. The traditional grey level cooccurrence matrix has limitations that prevent its use for segmentation purposes. The Gabor filter, based on characteristics of the human visual system, is an alternative approach that can generate an improved texture feature set. Texture segmentation applied to a difficult image demonstrates the versatility and appropriateness of the Gabor approach when compared to the cooccurrence features. I. INTRODUCTION Exceptional volumes of data transmitted from satellite and aerial radar platforms demand the automated interpretation of remotely sensed imagery in a cost and time-effective manner. There is no known algorithm capable of performing consistent identification of the various visually distinct formations found in SAR sea ice imagery. The remote sensing community recognizes that texture is an important aspect of automated segmentation of such imagery. A common texture feature appro.