Title :
A partially supervised change-detection technique
Author :
Prieto, Diego Fernàndez ; Arino, Olivier
Author_Institution :
Earth Obs. Applications Dept., ESA-ESRIN, Frascati, Italy
Abstract :
In this paper, a partially supervised approach to change detection is proposed. It allows the identification of one or few land-cover transitions of interest by using exclusively training samples belonging to the land covers involved. This represents a significant advantage, as avoids the need to label samples for all the land covers present in the area of interest at the two considered dates
Keywords :
geography; image recognition; remote sensing; co-registered remote-sensing images; identification; land use; land-cover transitions; partially supervised change-detection technique; training samples; Classification algorithms; Earth; Forestry; Image analysis; Labeling; Markov random fields; Pixel; Probability density function; Remote sensing; Urban areas;
Conference_Titel :
Geoscience and Remote Sensing Symposium, 2001. IGARSS '01. IEEE 2001 International
Conference_Location :
Sydney, NSW
Print_ISBN :
0-7803-7031-7
DOI :
10.1109/IGARSS.2001.976100
