Imaging systems have numerous applications in industrial, military, consumer, and medical settings. Assembling a complete imaging system requires the integration of optics, sensing, image processing, and display rendering. This feature issue is aimed at scientists, engineers, and practitioners interested in understanding how different materials, components, and image processing combine to determine and influence image system performance. The design of optical systems must factor the system as an integrated unit and optimize the performance for a given application. There are numerous disciplines that are needed for the design and advancement of an optical system. These disciplines include imaging optics, optical detection, computational, adaptive, and compressive imaging, displays, and usability of information; they all contribute to defining the system.

Scientists and engineers from commercial, academic, and military disciplines came together to share advances in imaging systems at the OSA Imaging Systems and Applications (IS) Topical Meeting in 2012 (Monterey, California). This issue contains a subset of the high-quality work presented at this meeting, as well as some contributions from the wider imaging systems community. These papers cover a broad spectrum of theoretical and experimental investigations in imaging systems and applications. In this collection of papers you will find research on coherent diffractive imaging beyond the Fresnel approximation, passive time-multiplexing superresolved imaging, sensors for lightning detection and imaging, open-access algorithms for commercial videokeratographers, analysis of unitary discrete linear canonical transforms, high precision calibration for fish-eye cameras, performance analysis of millimeter-wave imaging systems, extending depth of focus in tomography systems, information capacity for spectral imagers, in vivo brain imaging systems, pixel scaling for infrared focal plane arrays, a simple digital in-line holographic microscope, and an optical see-through head-mounted display.

In applications to microscopy, the measurement of the diffraction intensity with high numerical aperture beyond the Fresnel approximation is required to obtain the object information at high spatial resolution. Nakajima shows the effectiveness of lensless coherent imaging using a nonholographic and noniterative phase-retrieval method that allows the reconstruction of a complex-valued object from a single diffraction intensity measured with an aperture-array filter [1]. In this experimental paper, he demonstrates reconstruction with about 10 times the resolution of previously reported work and shows that object information in depth direction can be retrieved.

Zalevsky and co-workers present a superresolving approach for detecting an axially moving target that is based upon a time-multiplexing concept [2]. Applying a priori knowledge of the high-resolution background in front of which the target is moving, they overcome the diffraction limit set by the optics of an imaging camera. By recording a set of low-resolution images at different target axial positions, the superresolving algorithm weights each image by demultiplexing it using the high-resolution background image and provides a superresolved image of the target.

Rolando and co-workers present a CMOS image sensor dedicated to lightning detection and imaging [3]. The detector has been designed to evaluate the potentiality of an on-chip lightning detection solution based on a smart sensor. This evaluation is performed in the frame of the predevelopment phase of the lightning detector that will be implemented in the Meteosat Third Generation Imager satellite for the European Space Agency. The lightning detection process is performed by a smart detector combining an in-pixel frame-to-frame difference comparison with an adjustable threshold and on-chip digital processing allowing an efficient localization of a faint lightning pulse on the entire large format array at a frequency of 1 kHz.

Commercial videokeratometers provide useful information about corneal topography and tear film quality. None of the available devices, however, allows full control of the camera and processing algorithms. Espinosa and co-workers present an algorithm to process images of reflected Placido rings captured by a commercial videokeratoscope [4].They obtain raw data without Cartesian-to-polar-coordinate conversion, thus avoiding interpolation and associated numerical artifacts. Their approach processes 6 times more corneal data than commercial software and allows complete control from the capture of corneal images until the computation of curvature radii.

The numerical approximation of the linear canonical transforms (LCTs) is important in modeling coherent wave field propagation through first-order optical systems and in many digital signal processing applications. The continuous LCTs are unitary, but discretization can destroy this property. Liang et al. present a sufficient condition on the sampling rates chosen in the discretization to ensure unitarity [5]. They discuss the various subsets of the unitary matrices examined that have been proposed elsewhere, offer a proof of existence, and examine consequences, particularly in relation to the use of discrete transforms in iterative phase retrieval applications.

Tu et al. present a two-step calibration method to correct distortions of fish-eye cameras by using a global quadratic polynomial projection model fitting followed by local line-fitting calibration optimization for each section of the image to reduce residual error of the first calibration step due to small tangential distortion and asymmetry [6].

Inexpensive millimeter-wavelength (MMW) optical digital imaging raises a challenge of evaluating the imaging performance and image quality because of the large electromagnetic wavelengths and pixel sensor sizes and because of the noisiness of the inexpensive glow discharge detectors that compose the focal-plane array. Shilemay et al. quantify the performance of such MMW imaging systems, including a point-spread function and modulation transfer function [7]. Their experimental results and analysis indicate that the image quality is limited mostly by noise and that the blur is dominated by the pixel sensor size.

Paz et al. describe the development of a computed optical tomography (COT) system with extended depth of focus optics for locating impurities in glass lattice inspection [8]. The authors showed by numerical simulations and preliminary experiments that the proposed method extended the depth of focus for high frequencies compared to conventional optics.

Skauli proposes a way to use the information capacity as a figure of merit for spectral imagers [9]. As an example, he shows how the information capacity can be used to optimize the pixel size in a simple case that can be treated analytically. Generally, he argues that the information capacity is attractive as a fundamental, application-independent figure of merit for the optimization and benchmarking of spectral imagers.

Atchia et al. demonstrate an imaging technique based on vertical cavity surface emitting lasers (VCSELs) with extremely low transient times to enable a multiexposure laser speckle contrast imaging system [10]. Using a time of flight technique, data from multiexposure laser speckle imaging was observed to more closely agree with absolute velocity measurements, when compared to long-exposure laser speckle imaging. In addition, they infer additional depth information of the vasculature morphology by accounting for the change in the scattering from tissue above vessels with respect to the total scattering from blood flow and tissue.

The current trend in focal plane arrays (FPAs) for infrared (IR) imaging systems is toward smaller pixels. Catrysse and Skauli discuss effects that arise in pixels of IR FPAs when pixel size scales down to the wavelength of the incident radiation [11]. Using first-principles electromagnetic simulations of pixel structures based on a mercury cadmium telluride (MCT) photoconductor, they calculate the pixel quantum efficiency and crosstalk as pixel size scales from 16 μm down to 0.75 μm. Their numerical results indicate the possibility of wavelength-size and even subwavelength-size pixels. In addition, they find that the low-pass filtering effect of a metal film aperture can exemplify the impact and possible role that subwavelength structures play in controlling light inside very small pixels.

Ryle et al. describe a simple digital in-line holographic microscope (DIHM) capable of imaging weakly scattering 10 μm diameter microspheres, as well as Hs578T cells, over a depth of 1 mm [12]. This increase by a factor of 100 over the depth of field of a conventional microscope is quantified using the entropy of the resulting images as a figure of merit and the rates of false positive and negatives are examined.

Wang et al. present a design for an optical see-through head-mounted display using free-form design techniques with injection molding fabrication allowing for a fewer number of elements with reduced overall size and weight compared to traditional designs [13]. With this approach, optical performance loss is shown to be minimal with high yield rates.

These papers are a testament to how active and innovative research in this field is today. With the ever-growing integration of optics, image-capture devices, and image processing, we expect that imaging systems science and technology will continue to evolve rapidly in all application areas.