An ATR system using an integer based correlation algorithm in a time varying environment

Abstract

In an electro-optical or infrared tracking system, a correlation algorithm offers a robust tracking technique in a time varying scenario. Described in this paper is an Automatic Target Recognition algorithm that employs an operator in the loop as an embedded tracking system. The system described will combine morphological algorithms with an efficient integer based correlation track algorithm. This algorithm will be explored in a time varying search and track scenario using morphological matched filtering to automatically detect and select objects with a handover to a correlation tracker. Development of this algorithm using real and synthetic imagery will be reviewed as well as some preliminary results.

Detection and tracking of targets in infrared images using Bayesian techniques

“…In ATR and tracking systems, correlation algorithm plays a significant role in time varying scenarios. Correlation-based methods include morphological operators along with integer-based correlation [18], shift invariant correlation filters [20] and a fringe-adjusted joint transform correlator [19]. Further, a large number of ATR techniques exploit intensity [21][22][23].…”

“…Other techniques on ATR include, but not limited to, Eigen space-based methods [14], VLSI-based methods [15] and template-based methods [12,[16][17][18][19][20]. In ATR and tracking systems, correlation algorithm plays a significant role in time varying scenarios.…”