The design of modern imaging systems is intricately concerned with the control of optical aberrations in systems that can
be manufactured at acceptable cost and with acceptable manufacturing tolerances. Traditionally this involves a multi-parameter
optimisation of the lens optics to achieve acceptable image quality at the detector. There is increasing interest
in a more generalised approach whereby digital image processing is incorporated into the design process and the
performance metric to be optimised is quality of the image at the output of the image processor. This introduces the
possibility of manipulating the optical transfer function of the optics such that the overall sensitivity of the imaging
system to optical aberrations is reduced. Although these hybrid optical/digital techniques, sometimes referred as
wavefront coding, have on occasion been presented as a panacea, it is more realistic to consider them as an additional
parameter in the optimisation process. We will discuss the trade-offs involved in the application of wavefront coding to
low-cost imaging systems for use in the thermal infrared and visible imaging systems, showing how very useful
performance enhancements can be achieved in practical systems.
Pupil plane encoding has shown to be a useful technique to extend the depth of field of optical systems. Recently, further studies have demonstrated its potential in reducing the impact of other common focus-related aberrations (such as thermally induced defocus, field curvature, etc) which enables to employ simple and low-cost optical systems while maintaining good optical performance. In this paper, we present for the first time an experimental application where pupil plane encoding alleviates aberrations across the field of view of an uncooled LWIR optical system formed by F/1, 75mm focal length germanium singlet and a 320x240 detector array with 38-micron pixel. The singlet was corrected from coma and spherical aberration but exhibited large amounts of astigmatism and field curvature even for small fields of view. A manufactured asymmetrical germanium phase mask was placed at the front of the singlet, which in combination with digital image processing enabled to increase significantly the performance across the entire field of view. This improvement is subject to the exceptionally challenging manufacturing of the asymmetrical phase mask and noise amplification in the digitally restored image. Future research will consider manufacturing of the phase mask in the front surface of the singlet and a real-time implementation of the image processing algorithms.
Proc. SPIE. 5987, Electro-Optical and Infrared Systems: Technology and Applications II
KEYWORDS: Signal to noise ratio, Infrared imaging, Monochromatic aberrations, Point spread functions, Imaging systems, Image restoration, Wavefronts, Computer programming, Modulation transfer functions, Personal protective equipment
Pupil plane encoding enables extended depth of field and greatly reduced sensitivity to aberrations in an imaging system (field curvature, thermally induced defocus, astigmatism, etc.). The application of pupil plane encoding has potential in thermal imaging where it can enable the use of simple, low-cost, light-weight lens systems. We present numerical and modelling studies of the application of this technique to an uncooled LWIR imaging system, F/1, 75mm focal length, germanium singlet with a detector array size of 240x320 with 50 micron pixel. The initial singlet is corrected from coma and spherical aberration, but its performance across the field of view is greatly limited by astigmatism. The introduction
of an encoding asymmetrical germanium phase mask at the aperture stop of the system, combined with digital image processing, allows the removal of astigmatism and improved imaging performance across the field of view. This improvement is subject to a noise amplification in the digitally restore image. There is as a tradeoff between the maximum correction to astigmatism and reduced signal-to-noise ratio in the recovered image.
We present a two-stage process for target identification and pose estimation. A database of possible target states, i.e. identity and pose, is precomputed by a two-step clustering procedure, reflecting the two stages of the identification process. The current database is based on images generated from 3D CAD models of military ground vehicles on which realistic infrared textures have been applied. At the coarse level, the database is divided into a set of clusters, each represented by a small set of eigenimages, obtained through principal component analysis (PCA). The classification at this level is achieved by measuring the orthogonal distance between the region of interest (ROI) and the eigenspace of each cluster. Each cluster itself contains a few subclusters. A support vector machine is employed for a pairwise discrimination of subclusters. The likelihood that the target belongs to a particular cluster/subcluster is based on histograms, obtained at the time of training of the system. In addition to the classification of individual images it is also possible to handle image sequences where the pose of the target might vary in subsequent image frames. In this situation, the pose is assumed to change according to a first-order Markov process. The overall probability for each target state is accumulated through recursive Bayesian estimation. The performance of the above procedure has been evaluated through the identification of targets in synthetic image sequences, where the targets are placed in realistic backgrounds. Currently , we are able to correctly identify the targets in more than 80 percent of the image sequences. In about 60 (80) percent of the cases the pose can be estimated within an accuracy of 10 (20) degrees. The accuracy of the pose estimation is limited by the size of the sub-clusters.