No abstract

No abstract

Lineaments derived from remote sensing data are analyzed with respect to ground‐water exploration in the Voltaian Basin of central Ghana. The lineament data were collected using both Landsat TM and SPOT imagery, multiple interpreters, and multiple trials. Three types of reproducibility tests are analyzed: (1) azimuth‐length comparisons, disregarding location; (2) lineament coincidence using a raster‐based comparison method; and (3) feature agreement using a rule‐based approach. The reproducibility tests show that there are clear differences in length and location of individual line segments between interpreters, but that a large proportion of the inferred structural features are detected by all interpreters. Fifty percent of the features on both imagery types are detected by any two interpreters and 40 percent by all interpreters. Lineaments classed as more hydrologically significant show a reproducibility of up to 90 percent between interpreters and justify the use of classifiers in lineament mapping. The application of the feature agreement approach, including classification by hydrological significance, shows the greatest promise for targeting successful well sites.

No abstract

Efficient exploitation of hyperspectral imagery is of great importance in remote sensing. Artificial intelligence approaches have been receiving favorable reviews for classification of hyperspectral data because the complexity of such data challenges the limitations of many conventional methods. Artificial neural networks (ANNs) were shown to outperform traditional classifiers in many situations. However, studies that use the full spectral dimensionality of hyperspectral images to classify a large number of surface covers are scarce if non-existent. We advocate the need for methods that can handle the full dimensionality and a large number of classes to retain the discovery potential and the ability to discriminate classes with subtle spectral differences. We demonstrate that such a method exists in the family of ANNs. We compare the maximum likelihood, Mahalonobis distance, minimum distance, spectral angle mapper, and a hybrid ANN classifier for real hyperspectral AVIRIS data, using the full spectral resolution to map 23 cover types and using a small training set. Rigorous evaluation of the classification accuracies shows that the ANN outperforms the other methods and achieves ≈90% accuracy on test data.