Home // International Journal On Advances in Systems and Measurements, volume 7, numbers 3 and 4, 2014 // View article
Authors:
David Stork
Patrick Gill
Keywords: Computational sensing, phase grating, diffractive imager, application-specific sensing, face detection, QR code reading
Abstract:
We describe the optical, mathematical and compu- tational foundations for a new class of lensless, ultra-miniature computational imagers and image sensors. Such sensors employ phase gratings that have provably optimal optical properties and are integrated with CMOS photodetector matrices. These imagers have no lens and can thus be made extremely small (∼100 μm) and very inexpensive (a few Euro cents). Because the apertures are small, they have an effective depth of field ranging from roughly 1 mm to infinity. The grating acts as a two-dimensional visual “chirp” and preserves image power throughout the Fourier plane; thus the captured signals preserve image information. The final digital image is not captured as in a traditional camera but is instead computed from raw photodetector signals. The novel representation at the photodetectors demands powerful algorithms such as deconvolution, Bayesian estimation, or matrix inversion with Tikhonov regularization be used to compute the image, each having different bandwidth, space and computational complexities for a given image fidelity. Such imaging architectures can also be tailored to extract application-specific information or compute decisions (rather than compute an image) based on the optical signal. In most cases, both the phase grating and the signal processing can incorporate prior information about the visual field and the imaging or estimation task at hand. Our sensor design methodology relies on modular parallel and computationally efficient software tools for simulating optical diffraction, for CAD design and layout of gratings themselves, and for sensor signal processing. These sensors are so small they should find use in endoscopy, medical sensing, machine inspection, surveillance and the Internet of Things, and are so inexpensive that they should find use in distributed network applications and in a number of single-use or disposable applications, for instance in military, hazardous natural and industrial conditions.
Pages: 201 to 208
Copyright: Copyright (c) to authors, 2014. Used with permission.
Publication date: December 30, 2014
Published in: journal
ISSN: 1942-261x