Nonstationary image shift discriminators (NCD) can be used as measuring units of optoelectronic tracking systems and provide orthocorrelational estimation of additive shift on the continuous scale of values without using time-consuming extremum search operations.
NCD uses a decomposition of a television signal in terms of orthogonal basis wavelet-functions (BF). Statistical modeling confirms that asymptotic discriminatory characteristic (DC) over a set of image samples is determined by the convolution of the basis functions. It is shown that a proper choice of BFs provides invariance of DC to etalon changes. Standard deviation of DC’s zero may be a measure of noise variations of discriminator’s output characteristics in the neighborhood of zero-shift. The proposed procedure of calculation of this value reveals its linear dependance on DC’s monotonous zone aperture, low output noise level of discriminator and allows the optimal choice of NCD parameters for automatic locking and automatic tracking stages. The considered estimates are convenient when constructing NCD performance assessment procedures which take into account parameters, important for tracking systems, such as the aperture of lock-in zone and standard deviation of DC’s zero as a measure of output noise intensity.