Target Handing over System using Pyramid Processing .

Abstract

The image seen by the low resolution (LR) sensor and the image (of the same scene) seen by the operator through the high resolution (HR) sensor are different in spatial resolution. To establish the correlation between these two images, the HR image needs to be resampled and made similar to the LR sensor image by applying a pre-processing technique. The pre-processing will be carried out by handing over system that resamples the HR sensor image making it compatible with LR sensor image and hands over the resampled image to the LR sensor through a serial link. The pre-processing technique reported by Boland has been studied, analysed and implemented using 86 family of processors. Automatic target handing over system discusses the implementation of pre-processing technique, computational complexity and criticality of the execution time. The execution time reported was 1.5 s, whereas the requirement is of the order of few milliseconds for the typical set of conditions. This paper discusses the implementation of a suitable handing over algorithm. Emphasis has been to develop hardware and software to reduce the execution time, which has been brought down from 1.5 s to 40 ms for a typical set of conditions. Emphasis has also been given to reduce the transmission time by applying suitable pyramid processing techniques. The hardware is designed around i486 processor, and the software is developed in PL/M86.