Abstract
We propose an algorithm for automated type recognition of dynamic objects using observations performed by an optoelectronic device against a natural background. This algorithm is invariant with respect to the trajectories of the objects, is based on the ratio of the likelihood functions for simple alternative hypotheses, and implements an unbiased maximum-power criterion for object recognition with indeterminate a priori knowledge of the target environment. The likelihood functions are calculated over certain samples of wavelet-spectrum energies, fractal dimensions, and maximum auto-correlation-matrix eigenvalues for instrumentally measured elevation and azimuth, and the calculated maximum range of the object during a finite time interval. Modeling was used to establish that the algorithm has high computational efficiency for real-time use on modern PCs.
© 2017 Optical Society of America
PDF Article
More Like This
Cited By
You do not have subscription access to this journal. Cited by links are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.
Contact your librarian or system administrator
or
Login to access Optica Member Subscription