Title :
Infrared satellite image segmentation
Author :
Brad, Remus ; Popovici, Zeno O.
Author_Institution :
Comput. Sci. & Autom. Dept., Lucian Blaga Univ. of Sibiu, Sibiu, Romania
Abstract :
Satellite images are difficult to segment due to the complexity of phenomena captured within. Based on a rarely employed technique presented by Ahuja, we proposed a method for satellite infrared image segmentation. The previously introduced transform is extracting a pixel cohesion force field based on their similarities. We have found that the forces convergence points forms median lines of uniform regions. Combining the features provided by the transform with an adapted segmentation method, a successful region extraction is performed.
Keywords :
Artificial satellites; Automation; Clouds; Computer science; Image segmentation; Infrared detectors; Infrared imaging; Karhunen-Loeve transforms; Optical fiber communication; Pixel; Ahuja transform; cloud detection; infrared satellite image; segmentation;
Conference_Titel :
Roedunet International Conference (RoEduNet), 2010 9th
Conference_Location :
Sibiu, Romania
Print_ISBN :
978-1-4244-7335-9
Electronic_ISBN :
2068-1038