Abstract
A method is introduced for combining multispectral and ancillary data in remote sensing and geographic information processing to improve a digital classification and the accuracy of extracting information. The method can represent the uncertainty associated with observed data and provide plausible reasoning in data analysis. This method is viewed from the standpoint of knowledge engineering, and each source is considered as a piece of evidence that provides a certain measure of support to hypotheses. Each body of evidence also has a degree of reliability that represents the relative quality of the corresponding data set in data analysis. The authors focus on formal approaches to the quantitative reprentations of the degree of reliability and measure of support and examine existing combining functions of evidence.
| Original language | English |
|---|---|
| Pages | 1543-1547 |
| Number of pages | 5 |
| Publication status | Published - 1986 |