Modeling of the widefield space variant security systems

Abstract

The most important methods for security image data processing contain person recognition, suspicious events detection, objects and area of interest analysis and their classification, MSE, SNR and subjective evaluation. Image data processing of UWFC (Ultra Wide Field Cameras) security systems is very difficult in general. Precision of image data post-processing and analyses, which is very important for these systems, decreases because of a lot of different kinds of optical aberrations and distortions. Influence of optical aberrations increases at most at the frontiers of field of view (high angular distance from optical axis of the system). These aberrations distort the PSF of optical system and rapidly cut the accuracy of measurements. The third order optical aberrations models for LSI and LSV systems have been described. These models are based on Seidel and Zernike approximating polynomials. Optical aberrations effect to the transfer function (PSF) of optical imaging systems has been described. Higher quality and precision of image data can be obtained with deconvolution of the acquired images and system point spread function. The point spread function can be modeled as a space variant function from the estimation of optical system wavefront aberrations.