In the clutter-dominated disturbance modeled as spherically invariant random vectors with the same covariance matrix and possibly correlated texture components, we propose an estimator of covariance matrix, which exploits all secondary data fully and