Bookmark and Share

Computer Vision Metrics: Chapter One (Part B)

Register or sign in to access the Embedded Vision Academy's free technical training content.

The training materials provided by the Embedded Vision Academy are offered free of charge to everyone. All we ask in return is that you register, and tell us a little about yourself so that we can understand a bit about our audience. As detailed in our Privacy Policy, we will not share your registration information, nor contact you, except with your consent.

Registration is free and takes less than one minute. Click here to register, and get full access to the Embedded Vision Academy's unique technical training content.

If you've already registered, click here to sign in.

See a sample of this page's content below:


For Part A of Chapter One, please click here.

Bibliography references are set off with brackets, i.e. "[XXX]". For the corresponding bibliography entries, please click here.


2D Computational Cameras

Novel configurations of programmable 2D sensor arrays, lenses, and illuminators are being developed into camera systems as computational cameras [424,425,426], with applications ranging from digital photography to military and industrial uses, employing computational imaging methods to enhance the images after the fact. Computational cameras borrow many computational imaging methods from confocal imaging [419] and confocal microscopy [421, 420]—for example, using multiple illumination patterns and multiple focal plane images. They also draw on research from synthetic aperture radar systems [422] developed after World War II to create high-resolution images and 3D depth maps using wide baseline data from a single moving-camera platform. Synthetic apertures using multiple image sensors and optics for overlapping fields of view using wafer-scale integration are also topics of research [419]. We survey here a few computational 2D sensor methods, including high resolution (HR), high dynamic range (HDR), and high frame rate (HF) cameras.

The current wave of commercial digital megapixel cameras, ranging from around 10 megapixels on up, provide resolution matching or exceeding high-end film used in a 35mm camera [412], so a pixel from an image sensor is comparable in size to a grain of silver on the best resolution film. On the surface, there appears to be little incentive to go for higher resolution for commercial use, since current digital methods have replaced most film applications and film printers already exceed the resolution of the human eye.

However, very high resolution gigapixel imaging devices are being devised and constructed as an array of image sensors and lenses, providing advantages for computational imaging after the image is taken. One configuration is the 2D array camera, composed of an orthogonal 2D array of image sensors and corresponding optics; another configuration is the spherical camera as shown in Figure 1-8 [411, 415], developed as a DARPA research project at Columbia University CAVE.


Figure 1-8. (...