DocumentCode :
2881977
Title :
Texture Statistics Features of SAR Oil Spills Imagery
Author :
Yongsheng Yang ; Fuyuan Hu ; Juan Xia
Author_Institution :
Coll. of Electr. & Inf. Eng., Suzhou Univ. of Sci. & Technol., Suzhou, China
fYear :
2012
fDate :
1-3 June 2012
Firstpage :
1
Lastpage :
4
Abstract :
In the process of SAR oil spills imagery segment and detection, texture statistics features play a key role, which are based on the gray level co-occurrence matrix. Some of texture statistics features are illuminated, such as contrast, correlation, energy, homogeneity, entropy, max probability, cluster prominence, dissimilarity and difference entropy. The texture statistics features which are varied with the offset and the direction are analyzed by ENVISAT ASAR oil spills imagery. Energy, cluster prominence and contrast features of sea water are varied quickly with the offset and direction. However, cluster prominence of oil slicks has no sensitivity with the direction except 90 degree. These results can be gain helpful for SAR oil spills imagery detection and classification.
Keywords :
geophysical image processing; image segmentation; radar imaging; seawater; synthetic aperture radar; water pollution measurement; ENVISAT ASAR oil spills imagery; SAR oil spills imagery detection; SAR oil spills imagery segment; cluster prominence; difference entropy; gray level co-occurrence matrix; max probability; oil slicks; sea water; texture statistics features; Entropy; Feature extraction; Image segmentation; Probability; Remote sensing; Scattering; Synthetic aperture radar;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Remote Sensing, Environment and Transportation Engineering (RSETE), 2012 2nd International Conference on
Conference_Location :
Nanjing
Print_ISBN :
978-1-4673-0872-4
Type :
conf
DOI :
10.1109/RSETE.2012.6260762
Filename :
6260762
Link To Document :
بازگشت