Television – Camera – system and detail – Swing driven
Reexamination Certificate
2000-02-24
2003-05-27
Garber, Wendy R. (Department: 2612)
Television
Camera, system and detail
Swing driven
C348S218100, C348S273000
Reexamination Certificate
active
06570613
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention relates to devices and methods to increase the effective resolution of an array of spaced-apart light-sensitive transducers used to record an image in a digital imaging device.
2. Description of the Prior Art
Devices that record a visual scene by using an array of photoelectric sensors are well known in the art. Such a device configured as a scanner may be used to record digitally an existing transparency or print, or, in the form of a camera, may record a scene. The photoelectric-sensor array is typically a charge-coupled device (CCD) that outputs analog information that, in turn, is commonly digitized through the use of an analog-to-digital converter (ADC) and stored as bits of data in computer memory or other form of electronic storage. For this discussion we will refer to camera-type devices utilizing CCD array sensors, though the principles laid out here apply equally well to cameras constructed with other types of photoelectric sensors.
Often the CCD is the single most expensive part of the digital camera. The quality of the recorded image is related in part to the total number of sensing elements (pixels) that are present in the CCD. The greater this number, the higher the quality of the resulting image. In fact, it is the number of pixels that the camera can capture that will often distinguish one camera from another in the marketplace. Thus any device or method that can increase the effective pixel count of a CCD without causing the actual count to increase is of interest to camera manufacturers since such a device can achieve higher performance with minimal cost increase.
Professional digital cameras used in the studio tend to fall into two categories: (1) those that can record action scenes, but sacrifice color accuracy and suffer from color aliasing; (2) those that can only record still-life scenes, but have higher accuracy and resolution enhancement. Since most studios must photograph both types of scenes, most studios are compelled to own both types of cameras at a considerable financial burden. A single camera that can function in both modes would be very desirable.
Definition of Terms
An “array sensor” is a sensor containing an array of tiled sensor pixels, i.e., the sensor pixels are arranged in contiguous horizontal rows and vertical rows. A “sensor pixel” is defined as the smallest unit of area on the sensor that can be tiled to create an array. By this definition, a sensor pixel may include both light-sensitive and non-light-sensitive regions. Referring to
FIG. 1
, array
10
is comprised of many instances of pixel
11
, pixel
11
being further comprised of light-sensitive region
12
and non-light-sensitive region
14
. “Sensor pixel pitch” in either the X or Y direction is equal to the dimension of sensor pixel
10
in the corresponding direction. (Alternatively, it can be described as the center-to-center distance between pixels in the stated direction.) “Aperture ratio” is defined as the area of light-sensitive region
12
divided by the total pixel area (the sum of the area of non-light-sensitive region
14
and light-sensitive region
12
). Aperture ratio can be decomposed into an X component and a Y component.
FIG. 1
represents an array with an aperture ratio of about 25% (50% in X and 50% in Y).
FIG. 2
represents an array with an aperture ratio of about 50% (50% in X and about 100% in Y).
FIG. 3
represents an array with a nearly 100% aperture ratio. The term “resolution” refers to a total number of pixels being fixed. Resolution is a measure of information. “Native resolution” refers to the actual number of pixels in the imager. “Effective resolution” produced from some resolution-enhancement technique is defined to be numerically equal to the native resolution that would be otherwise needed to create the same amount and quality of pixel information without said technique.
The terms “image pixel” and “sensor pixel” distinguish the image realm from the sensor realm. An image is typically comprised of image pixels that are computed from sensor pixel data using a correspondence that is often, but not necessarily one-to-one.
Many devices and methods have been proposed to make the effective sensor resolution higher than the native sensor resolution. In general, these approaches can be divided into three main categories.
Three Approaches to Increase Sensor Resolution.
The most straightforward approach is the tiling approach, which is to move, between successive exposures, the entire sensing array once or multiple times a distance equal to the width and/or height of the whole array and tiling the resulting image pixels together. The effective sensor resolution is higher than the native resolution by the number of such moves that occur in the course of producing the net image of such tiles.
The tiling approach has several disadvantages. First, it is difficult to implement, since the translation device must rapidly and with sub-pixel pitch accuracy displace the array a distance equal to its full width and/or height with repeatable results. Since the typical imager will have thousands of pixels in each dimension, the positioning accuracy must be, say, one part in ten thousand or better to get good tiling. Also, the imaging array must not be displaced in a direction normal to the focal plane or a deleterious focus shift will occur. Finally, such an approach has the effect of making the angle of view dependent on the size of the tiling for a given focal length lens. This is a very awkward situation since a non-tiled image does not yield a preview of a final higher resolution image but rather only a section of the higher resolution image. Thus the whole tiling procedure must be completed before viewing a high resolution image is possible. If the photographer completes his or her image at a low resolution (say a single tile for example) and is satisfied with the aesthetics, but decides the image lacks sufficient resolution, then the new exposure must be tiled demanding that the photographer change lenses and recompose. This change of lens is slow, and usually requires re-focusing and resetting the aperture. A zoom lens can be used to overcome the lens substitution problem, but not the re-composition problem, and zoom lenses usually possess a lower resolving power than the same fixed focal length lens.
FIGS. 5A and 5B
illustrate the tiling approach.
FIG. 5A
shows a 3×3 array with a native resolution f9 and angle of view theta;
FIG. 5B
shows the 3×3 array tiled 3 times, with a native resolution of 9 and an enhanced resolution of 27. The angle of view is (2)×theta.
It would be much more desirable to maintain the same angle of view independent of effective resolution. This would allow a quick preview for aesthetic composition for example.
The second approach is the interstitial approach. This approach depends on the fact that the aperture ratio of some imagers is or can be made to be equal to or less than 50% in at least one (X or Y) direction, i.e., part of each pixel is non-sensitive. For example, an interline transfer CCD has columns of non-sensitive regions interleaved with the sensitive regions. Such imagers are common in video cameras. A way to increase the effective sensor resolution with such imagers is to shift light-sensitive regions into the non-sensitive regions. By shifting such a “sparse” CCD array by an amount sufficient to re-position the light sensitive areas to a new position previously occupied entirely by non-sensitive areas, a new array of image data is produced that interleaves with the native data, thus increasing the effective sensor pixel count. The success of this approach depends in part on the fact that the native and shifted data sets are not convoluted because there is no spatial overlap between the unshifted and shifted light-sensitive positions.
Many such shifting methods exist and all depend on the existence of insensitive regions of the array into which the sensitive areas may be shifted. Torok et al. (U.S. Pat. No. 5,489,994
Bohan Thomas L.
Garber Wendy R.
Mathers Patricia M.
Nguyen Luong
LandOfFree
Resolution-enhancement method for digital imaging does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Resolution-enhancement method for digital imaging, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Resolution-enhancement method for digital imaging will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3090539