Image Analysis, Classification and Change Detection in Remote Sensing, 4th Edition

The output is the same as in Section 1.3.4 apart from small rounding errors:

[[   407.13229638         442.18038333         -78.32374081]

   [   442.18038333         493.57036037      -120.6419761   ]

   [   -78.32374081      -120.6419761            438.95704379]]

2.3.3    Real and complex multivariate sample distributions

For the purposes of interval estimation and hypothesis testing, and also for image classification and change detection, we require not only sample vector estimators, but also their distributions. We concluded, from Theorem 2.3 in Section 2.1.4, that the mean of m normally distributed samples $Z_i\sim N\left(\mu ,{\sigma }^2\right)$,

$\overline{Z}=\frac{1}{m}{\displaystyle \sum _{i=1}^m{Z}_i,}$

is normally distributed with mean μ and variance σ²/m. Similarly, according to Theorem 2.8, (m − 1) S/σ²