Title :
Analysis to the relationship of classification accuracy, segmentation scale, image resolution
Author :
Huiping, Huang ; Bingfang, Wu ; Jinlong, Fan
Author_Institution :
Inst. of Remote Sensing Applications, Chinese Acad. of Sci., Beijing, China
Abstract :
Information extraction encountered a new challenge while the spatial resolution is increasing quickly. People suppose that the higher the spatial resolution is, the better is the result of classification. To prove this guess we use two approaches: pixel-based classification and object-oriented classification. The former site test shows one class has different accuracy from various resolution images. The object-oriented approach is an advanced solution for image analysis. The accuracy of object-oriented approach is much higher than those of the pixel-based approach. The site result shows that each class has its optimal image segmentation scale.
Keywords :
feature extraction; geophysical techniques; image classification; image resolution; image segmentation; object-oriented methods; remote sensing; classification accuracy; image analysis; image resolution; information extraction; object-oriented classification; optimal image segmentation scale; pixel-based classification; segmentation scale; spatial resolution; Image analysis; Image color analysis; Image generation; Image resolution; Image segmentation; Maximum likelihood detection; Pixel; Remote sensing; Spatial resolution; Testing;
Conference_Titel :
Geoscience and Remote Sensing Symposium, 2003. IGARSS '03. Proceedings. 2003 IEEE International
Print_ISBN :
0-7803-7929-2
DOI :
10.1109/IGARSS.2003.1295233
