Doamusic is a python library that implements the MUltiple SIgnal Classification (or MUSIC) algorithm for direction of arrival estimation as described by R.O. Schmidt..
MUSIC, as applied to an N-element array, uses the NxN joint expectation matrix (whose ij element is the expectation of the product of the signals at the ith and jth elements) to discriminiate between up to N-1 uncorrelated incident signals. Doamusic was developed for microwave direction finding, but it should be usable for any array of coherent sensors.
Critical computations are multithreaded and using the BLAS in compiled native code, and non-crtical computations use Numpy. On 2006 hardware, Doamusic can render a 512x512 pseudospectrum image for a 64 element array in around 2.3 seconds.