1.7 Real Hyperspectral Images to be Used in this Book

Three real hyperspectral image data sets are frequently used in this book for experiments. Two are AVIRIS real image data sets, Cuprite in Nevada and Purdue's Indian Pine test site in Indiana. A third image data set is HYperspectral Digital Imagery Collection Experiment (HYDICE) image scene. Each of these three data sets is briefly described as follows.

1.7.1 AVIRIS Data

Two AVIRIS data sets presented in this section are Cuprite data and Purdue's data, which can be used for different purposes in applications. The Cuprite data set is generally used for endmember extraction and target detection, while the Purdue's data set is mainly used for land cover/land use classification.

1.7.1.1 Cuprite Data

One of the most widely used hyperspectral image scenes available in the public domain is Cuprite mining site, Nevada, as shown in Figure 1.11(a). It is an image scene of 20 m spatial resolution collected by 224 bands using 10 nm spectral resolution in the range of 0.4–2.5 μm. The center region shown in Figure 1.11(b), cropped from the image scene in Figure 1.10(a), has size of img pixel vectors.

Figure 1.11 Cuprite image scene, (a) original Cuprite image scene; (b) the image cropped from the center region of the original scene in (a) (img). (See ...

Get Hyperspectral Data Processing: Algorithm Design and Analysis now with the O’Reilly learning platform.

O’Reilly members experience books, live events, courses curated by job role, and more from O’Reilly and nearly 200 top publishers.