non zero eigen values of covariance matrix

I refered a research paper and in that i need to find non zer eigen value of covariance matrix denoted by E[hh transpose]. What exactly i need to do here. Also how the values of h (where h is channel coefficients) should be taken.