A Synthesis of Target-Localizing Algorithms Tolerant to Image Brightness Variations in Two-dimensional Correlative Extremal Navigation Systems Part II. Choosing a Primary Standard of Maximal Correlation and Optimal Algorithm Synthesis
Alexandr Sergeevich Sultanov, Valeriy Ivanovich Antyufeev, Boris Ivanovich Makarenko · Telecommunications and Radio Engineering · 2000
The problem [1] devoted to the selection of the most correlated reference image, represented in absolute, ratio, difference, interval, order, hyperorder, and nominal scales, has been solved. Target-locating algorithms optimal for the maximum likelihood criterion have been synthesized for the maximal correlation systems of navigation. The results of statistical tests of these algorithms have shown that under substantial brightness variations of a current image, the algorithm most resistant to these variations is based on representation of the reference image on the order scale. This algorithm slightly increases calculation costs when compared to the algorithm on the absolute scale.