Advances in Earth Science ›› 2003, Vol. 18 ›› Issue (2): 185-191. doi: 10.11867/j.issn.1001-8166.2003.02.0185

• Articles • Previous Articles     Next Articles

SOME PROBLEMS IN CONSTRUCTING THE GROUND OBJECT SPECTRAL KNOWLEDGE BASE AND ITS SERVICES

Su Lihong 1,Li Xiaowen 1,2,Wang Jindi 1,Tang Shihao 1   

  1. 1. Research Center of Remote Sensing and GIS, Beijing Normal University,Beijing 100875,China;2. Deptartment of Geography and Center for Remote Sensing, Boston University, Boston, MA 02215, USA
  • Received:2002-03-25 Revised:2002-09-24 Online:2003-04-10 Published:2003-04-01

Su Lihong,Li Xiaowen,Wang Jindi,Tang Shihao. SOME PROBLEMS IN CONSTRUCTING THE GROUND OBJECT SPECTRAL KNOWLEDGE BASE AND ITS SERVICES[J]. Advances in Earth Science, 2003, 18(2): 185-191.

The objective of the spectral knowledge base (spectrum library) is to facilitate remote sensing applications. The paper discusses the six problems that are required to solve in constructing the spectral knowledge base and its services. Firstly, in order to share data, the spectral knowledge base should show the measured spectral data and relative information such as the observation criterion and field campaign condition etc. In other words, it is necessary to have the integrated system of spectral and environmental data and the self-contained metadata. Secondly, for the sake of solving the discrepancy between the temporal and spatial scales of the measured spectral data and of remote sensing applications, it is essential to study quantificational descriptions of land surface parameters and approaches to convert the parameters between temporal and spatial scales. In the next place, because it is impractical to measure vegetation spectrum at all times during vegetation growth cycle, remote sensing physical models, which are used to interpolate and extrapolate the measured data, should be collected and the applicable conditions of these models will be evaluated. The remote sensing physical model, which maybe are analytic equations or computer simulation models, are used to extend vegetation parameters along temporal change, and compute the spectrum on three spatial scales such as leaf, canopy and remote sensing pixel from visible light to thermal infrared based on the measured spectral data, the extended vegetation parameters, and pixel components. Fourthly, the model metadata are defined and collected so that the users can know which models in the spectral knowledge base is suitable to their tasks and why they are. Moreover, building the model run-time support software based on metadata of data and models is an effectual approach to extract the parameters for the models and run the models automatically. Finally, in order to sharing data and model on Internet, it is need to research how manage the data and models so that the users can obtain the data and models on Internet easily and the interpolated and extrapolated the spectral data real-timely.

No related articles found!
Viewed
Full text


Abstract