Title :
Interactive 4 channels histogram analysis technique for NOAA AVHRR
Author :
Ozawa, Noriaki ; Kudoh, Jun-ichi ; Nemoto, Yoshiaki
Author_Institution :
G.S.I.S., Tohoku Univ., Sendai, Japan
Abstract :
There are several methods for multidimensional satellite image analysis which classify image pixels into some categories or extract areas which have the same properties. Often the automatically generated results from unsupervised methods are unsatisfactory particularly in cloud removal. These methods generally adopt the conventional pattern recognition approach, wherein the distance of sample vectors in the multidimensional spatial distribution is calculated. However, in cases of remote sensing, the class boundaries are not clear. To obtain satisfactory results, manual intervention in the form of defining the boundaries is necessary
Keywords :
edge detection; geophysical signal processing; image classification; remote sensing; NOAA AVHRR; class boundaries; classification; cloud removal; image pixels; interactive 4 channels histogram analysis technique; multidimensional satellite image analysis; multidimensional spatial distribution; pattern recognition; Clouds; Data analysis; Electronic mail; Histograms; Image analysis; Multidimensional systems; Pixel; Remote sensing; Satellites; Telecommunication computing;
Conference_Titel :
Geoscience and Remote Sensing Symposium, 1999. IGARSS '99 Proceedings. IEEE 1999 International
Conference_Location :
Hamburg
Print_ISBN :
0-7803-5207-6
DOI :
10.1109/IGARSS.1999.771629