PySptools

PySptools (Python Spectral Tools) is a Python module that can be used by software developers to integrate spectral and hyperspectral algorithms into their applications, in order to visualize chemistry as well as the spatial relation between chemicals and their proportions.

Some of its key features are represented by convex hull removal, scikit-learn bridge, unmixing process, noise reduction, spectrum-level features extraction, supervised classification, target detection, and endmembers extraction.

To be able to run this library, you must have Python 2.7 or 3.x installed. Numpy, Scipy, scikit-learn, SPy and Matplotlib are also required, while CVXOPT (for running FCLS) and IPython (for using the display feature) are optional.

In addition to the source code, the downloaded package includes examples and files for abundance maps, classification, detection, distance, material count, and noise, among others. Tests are available too.

For installation, it's necessary to take the path of the pysptools-0.xx.x folder and insert it into the PYTHONPATH system variable. Distutils can be used while the library can be called from a Command Prompt window with "python setup.py install" after going to the pysptools-0.xx.x folder.

As far as the algorithm sources are concerned, you can separately download and install Matlab Hyperspectral Toolbox, PWCTools (piecewise constant toolbox), EIA (Endmember Induction Algorithms toolbox), HySime, and scikit-learn. Complete user documentation is available at the developer's website.