Progress in Camera-Based Document Image Analysis
ICDAR '03 Proceedings of the Seventh International Conference on Document Analysis and Recognition - Volume 1
| Hi-index | 0.00 |
A critical issue in Automatic Target Recognition is the detection and identification of low contrast targets in Forward Looking Infra-Red (FLIR) images at sufficiently long ranges. In this paper, we describe a technique basedon determining the best magnification to zoom in onto a target object while it is still at a long range. We model the zoom process based on the expected signal from the object and the background, and show that the variance in a region of interest can be used in a feedback control loop to determine the optimal zoom. The method is highly suitable for real-time implementation and is demonstrated through a series of experiments.