A lightweight hyperspectral imaging library that provides developers with spectral algorithms for the Python programming language. New for v0.14.x: a scikit-learn bridge (alpha and partial).
The functions and classes are organized by topics:
* abundance maps: FCLS, NNLS, UCLS
* classification: AbundanceClassification, NormXCorr, KMeans SAM, SID, SVC
* detection: ACE, CEM, GLRT, MatchedFilter, OSP
* distance: chebychev, NormXCorr, SAM, SID
* endmembers extraction: ATGP, FIPPI, NFINDR, PPI
* material count: HfcVd, HySime
* noise: Savitzky Golay, MNF, whiten
* sigproc: bilateral
* sklearn: HyperEstimatorCrossVal, HyperSVC and others
* spectro: convex hull quotient, features extraction (tetracorder style), USGS06 lib interface
* util: load_ENVI_file, load_ENVI_spec_lib, corr, cov and others
The library do an extensive use of the numpy numeric library and can achieve good speed. The library is mature enough and is very usable even if the development is at a beta step.
Great job! Always good to see more hyperspectral tools. I think you should seriously consider "going into business" with the developer of SPy. You both have great packages and I believe that together you will get more recognition in the community as a "go to" toolbox. Keep up the good work.