In this collection of papers, you will find research on computationally efficient video restoration for Nyquist-sampled imaging sensors; spectral imaging by combining transverse-field-detector and color-filter arrays; continuum fusion solutions for replacement target models in electro-optic detection; image enhancement of objects with highly dynamic brightness and large rotational motion; a video-rate nine-band multispectral SWIR sensor; design of pictorial stimuli for perceptual experiments; characterization of a 2-Gpixel-wide field-of-view visible imager; a compact camera for multispectral and conventional imaging based on patterned filters; and analytical direct solutions of the Risley prism systems for tracking and pointing.
Schaum discusses the general problem of detecting target signals that are not additive, but instead replace some part of a background signal. This is a relevant signal model in imaging, for example, in the case of detecting an opaque subpixel target in a hyperspectral image. It is then necessary to test the hypothesis of the target being present, but without knowing the fraction of the target in the pixel. Schaum employs his original “continuum-fusion” concept to tackle this problem and derives families of new detectors that can be expressed in closed form.
Kanaev et al. report on the application of multiframe superresolution to sampling limited imagery that models space objects. They analyze the motion-estimation errors from the standpoint of an optical-flow-interpolation-error metric and show dependence of the object-tracking estimation on brightness changes and on the pixel displacement values between subsequent images. Spatial-acuity enhancement was obtained with a resolution enhancement.
Farnand and Fairchild studied the optimal design of pictorial stimuli for effective and efficient perceptual experiments with the goal of understanding the impact of image content on visual attention and consistency of experimental results. They apply this understanding to develop guidelines for pictorial-target design for perceptual-image-comparison experiments. It is shown that, along with scene complexity, the image modifications and the difficulty of the image-equivalency decisions played a role in the experimental response.
Marks et al. present a 2-Gpixel camera built around a single objective lens of monocentric design, i.e., with only concentric spherical surfaces. An array of up to 382 subcameras is placed around the objective lens, each covering a part of the full field of view. MTF measurements show good performance, and the paper also presents simple observer tests. The image presentation concept is radically different from conventional cameras in that only the part of the image requested by the user is actually rendered. Apart from their primary use in surveillance, these cameras are likely to be used for detailed recordings of large scenes of interest such as sports events and landscapes.
Skauli et al. describe a design for a compact camera that allows for multispectral and conventional imaging based on utilization of striped patterned filters in the focal plane combined with scanning of the field of view. This approach allows for both multispectral and conventional 2D imaging in a single, compact, low-cost architecture. The potential volume-size reduction is of the order of several hundred for comparable traditional imaging systems.
Kutteruf et al. take the use of patterned filters a step further and present a nine-band multispectral camera capable of video-rate-snapshot imaging. The camera employs a filter array with repeating cells of different pixel-sized bandpass filters attached directly to an InGaAs image sensor operating in the 1–1.7 μm spectral range. The paper discusses the spectral response, pixel alignment, and radiometric calibration of the camera. The recorded images are in good agreement with hyperspectral images of the same scene. This technology has a clear potential for use in military systems as well as many emerging applications using the short wave infrared spectral range.