P. Gong, INTEGRATED ANALYSIS OF SPATIAL DATA FROM MULTIPLE SOURCES - USING EVIDENTIAL REASONING AND ARTIFICIAL NEURAL-NETWORK TECHNIQUES FOR GEOLOGICAL MAPPING, Photogrammetric engineering and remote sensing, 62(5), 1996, pp. 513-523
As the availability of digital spatial data, other than from remote se
nsing, increases, it becomes increasingly important to develop algorit
hms to handle both remote sensing and other spatial data. For classifi
cation purposes, commonly used remote sensing algorithms such as the m
aximum-likelihood classifier and the minimum-distance classifier can o
nly be used to deal with spatial data of interval and ratio scales. Th
ey are not applicable to spatial data of nominal or ordinal scale as e
xemplified by data digitized from a categorical map. Bayesian theory,
mathematical theory of evidence, and artificial neural networks, on th
e other hand, are capable of handling data with any measurement scale.
In this paper, we introduce an evidential reasoning and a back-propag
ation feed-forward neural network algorithm and evaluate their applica
tions to classification problems. A multisource data set including Lan
dsat Thematic Mapper, aeromagnetic, radiometric, and gravity data has
been used in the classification of four rock types in Melville Peninsu
la, Northwest Territories, Canada. The highest overall accuracy of 96.
0 percent and average accuracy of 92.1 percent were achieved with the
neural network algorithm while the evidential reasoning method produce
d an overall accuracy of 94.7 percent and average accuracy of 89.3 per
cent. The evidential reasoning method resulted in three highest indivi
dual class accuracies out of the four classes.