Spectrum Processing Technology In Joint Transform Correlator

In the field of optical correlation detection, matched filter correlator (VLC) and joint transform correlator(JTC) are two kind of important correlators.Toward the direction of real-time and programmable,joint transform correlator develops quickly over recent years.But recognition capability of joint transform correlator will be affected by background noise,rotation and size variation of target images.Spectrum processing algorithms of joint transform correlator are studied here.After analysing and comparing several algorithms,minimum average correlation energy(MACE) method is choosed to synthsize training images in order to get synthesized reference image. Laplace sharpening to the joint power spectrum of object image and synthesized reference image are carried out. Improved MACE method has solved the problem which object image is sensitive to scale and rotation variation in traditional joint transform correlator, recognition scope is expanded.Computer simulation experiment and optical experiment of rotation variant object image proves that improved MACE method can be used in joint transform correlator, and target recognition rate is increased.