Spatial composition of spectral classes. A structural approach for image analysis of heterogeneous land-use and land-cover types

Journal Article · Tue Feb 01 00:00:00 EST 1994 · Photogrammetric Engineering and Remote Sensing; (United States)

OSTI ID:6530591

Fung, T; Chan, K ^[1]

---

1. Chinese Univ. of Hong Kong, Shatin (China)

Spectral classes generated from image classification are conventional output in land-use and land-cover (LU/LC) mapping with digital remotely sensed data. The spatial composition of these spectral classes within a certain spatial range, or window, can be useful information for image analysis. In this study, the concept of spatial composition of spectral classes (SCSC) was developed and examined. It was found that LU/LC types did exhibit different characteristics of SCSC. Ranges of SCSC were used for post-classification labeling to identify different LU/LC types. Results showed that a 7 by 7 window size was suitable for Hong Kong at Level-II classification. 18 refs.

• OSTI ID: 6530591
• Journal Information: Photogrammetric Engineering and Remote Sensing; (United States), Vol. 60:2; ISSN 0099-1112
• Country of Publication: United States
• Language: English

Related Subjects

54 ENVIRONMENTAL SCIENCES
GROUND COVER
MULTISPECTRAL PHOTOGRAPHY
REMOTE SENSING
PLANTS
SPATIAL DISTRIBUTION
LAND USE
DISTRIBUTION
PHOTOGRAPHY
540210* - Environment
Terrestrial- Basic Studies- (1990-)