Infrared Search System Image Registration Algorithm For Hardware Implementation Of A Number Of Studies

In the infrared search system, image registration is the important assurance for complex processing such as reliable detection and infrared target tracking. In order to improve the real-time performance of the system, the processing time needs to be shorten for the image registration algorithm to provide ample time allowance for subsequent processing. For more than the actual demand, this topic does research on image registration algorithm and optimization technique, and implies the algorithms on the high-speed hardware platform for real-time.Based on the analysis of the gray scale and character, this topic is mainly focused on the phase correlation method for image registration, the KLT corner tracking method, and the method of scale invariant features. Phase correlation method is mainly studied in the two-dimensional Fourier transform processing module and the mutual power spectrum calculation module. The parallel optimization design is implied, and algorithm processing time is greatly reduced and the effect of registration is better. KLT corner tracking method which is based on the point of KLT feature detection and tracking method completes infrared image registration. Less hardware resources, higher real-time performance. Scale invariant features method does the main research on the gaussian pyramid and the establishment of gaussian difference scale space and simplify, the hardware design method of characteristic vector generation and description, finally on the premise of guaranteeing the scale-invariant features accuracy the processing time is greatly reduced, and the method provides ample time allowance for subsequent algorithms.This topic completes the three kinds of image registration algorithm design, debug and validation on the field programmable logic array hardware platform. The experimental results show that by the improved image registration algorithm and hardware optimization, the real-time has improved greatly. Under the premise that guarantees the precision of image registration, this topic solves the problem of real-time performance about target detection and tracking module in the infrared search system.