Change is intrinsic to ecosystems, but is also the essence of instability and the outgrowth of situations that lack sustainability. It must also be recognized that change can be associated with either restoration or degradation. Compressed multiband image data provides increased flexibility and practicality for systematic change detection on a regional basis. Combining such capability with conceptual extensions of spatial pattern analysis represents a methodology for systematically monitoring spatial structure of spectral change across landscapes in order to profile characteristic broad scale regimes of change and to indicate trends in those regimes. Sustainability and ecosystem health are watchwords of contemporary ecosystem management. To solve urgent needs in application of remote sensing data, environmental change must be detected based on monitoring spatial and temporal regimes across landscapes. Environmental landscape level indices are used to examine land cover transitions. Based on classified TM images for North-Western Black Sea, Constantza urban area, Romania, conditional probability matrices of land cover transition are compared to measures of landscape structure. Based on proper algorithms for structural composition modeling were defined landscape elements being estimated the probabilistic behavior of pixels. This approach suggests a means of linking the probabilistic behavior of the fine scale dynamics to the pattern observed at larger spatial scales.

Date Published: 21 October 2004
PDF: 9 pages
Proc. SPIE 5581, ROMOPTO 2003: Seventh Conference on Optics, (21 October 2004); doi: 10.1117/12.583054

Published in SPIE Proceedings Vol. 5581:
ROMOPTO 2003: Seventh Conference on Optics
Valentin I. Vlad, Editor(s)