Facsimile and static presentation processing – Facsimile – Picture signal generator
Reexamination Certificate
1999-07-30
2003-04-29
Lee, Cheukfan (Department: 2722)
Facsimile and static presentation processing
Facsimile
Picture signal generator
C358S444000
Reexamination Certificate
active
06556315
ABSTRACT:
FIELD OF INVENTION
This invention relates generally to digital image scanning and more specifically to image scanners having segmented photosensor arrays.
BACKGROUND OF THE INVENTION
Image scanners convert a visible image on a document or photograph, or an image in a transparent medium, into an electronic form suitable for copying, storing or processing by a computer. Reflective image scanners typically have a controlled source of light, and light is reflected off the surface of a document, through an optics system, and onto an array of photosensitive devices. The photosensitive devices convert received light intensity into an electronic signal. Some imaging devices, for example digital cameras, have a two-dimensional photosensor array. Other imaging devices have a one-dimensional photosensor array, and the overall image is formed one raster line at a time by providing relative movement between the object being scanned and the photosensor array. In some scanners, light from an object being scanned is reduced by the optics system onto a photosensor array having a length that is quite small relative to the width of the object being scanned. Small photosensor arrays may be fabricated as a single integrated device. In other scanners, the object is focused onto an array of photosensors using optics having a one-to-one magnification, so that the length of the photosensor array is the same as the width of the object being scanned. These relatively long arrays are typically manufactured by assembling multiple segments. Segments may abut end-to-end, or segments may have gaps between ends. The present patent document is primarily concerned with one-dimensional segmented photosensor arrays, but it is not important whether the ends abut, or whether the ends have gaps.
A picture element (pixel) may be defined as an area on the image being scanned, or as an area on a photosensor array, or as a set of numbers in the data representing an image. For document scanners and transparent film scanners, a pixel is commonly defined as an area on the surface of the document being scanned. For example, for document and transparent film scanners, a common specification is “pixels per inch” (or mm) as measured on the surface of the document being scanned.
For segmented photosensor arrays, alignment of the segments is important. Visually, an error in photosensor position of about one pixel is minor. Therefore, if visual appearance is the only requirement, the segments may be mounted with a spatial accuracy that ensures that actual pixel locations will not deviate from ideal locations by more than about the pitch of the photosensor elements. In some applications, however, precise pixel positioning is needed for reasons other than visual appearance. For example, hand scanners require an operator to manually move a sensor array over an original image, capturing image data in relatively narrow swathes. Image processing software then stitches the various swathes together. Stitching error can result in accumulation of error in the computed position of the scanner on the page. The stitching algorithms may require pixel accuracy to be, for example, less than half the pitch of the photosensor elements.
Placing photosensor array segments over a relatively long length with high precision is technically challenging. High precision adds cost due to the alignment equipment and mounting techniques required, and more importantly, adds cost due to scrapping of photosensor assemblies that do not meet specifications. If the specifications push the state of the art, yield may drop substantially, and relatively small changes in the specification may make significant improvements in yield. There is a need to be able to relax mechanical alignment requirements for photosensor segment alignment, while still providing accurate information needed for image stitching and scanner positioning.
The following discussion provides additional background for a prior art hand held scanner to provide an example of a device where the invention may be used, and to facilitate later discussion of the invention in light of the prior art. A example of a scanning device for which the invention is applicable may be found in U.S. Pat. No. 5,578,813 (Allen '813). Allen '813 describes a hand-held scanner in which a scan-line is defined by the position of a linear array of photosensors. The linear sensor array measures image intensity at uniform positions between the ends of a scan-line. The scan-line may be skewed relative to the sides of the document being scanned and the skew may change as the scanner is moved. In general, the intensity samples measured by the scanner do not fall on a uniform rectangular grid. Separate position sensors monitor the positions of the ends of the scan-line, relative to a starting reference position, as the scanner is moved across the page. Position-tagged intensity data is used to compute image pixels at discrete points on a uniform rectangular pixel grid. The process of transforming a digital image from one set of pixel coordinates to another set of pixel coordinates is called geometric image resampling. Resampling from non-rectangular coordinates to rectangular coordinates is called image rectification.
Ideally, the rectangular pixel grid has grid lines that are parallel to the sides of the document. For example, if a complete image is saved in memory, image analysis may be performed to determine margins, or edges of photographic images, or other information indicating the orientation of the document. Alternatively, the first scan-line may be used to define a reference position and one of two orthogonal directions. All scan-lines after the first may then be resampled to a rectangular grid that is defined relative to the first scan-line. In the following discussion, for convenience of illustration, pixel grids are depicted as square, having the same pitch for each of the two orthogonal directions. Likewise, for convenience of illustration, the optical sampling rate of the scanner along the direction of movement is depicted as the same as the optical sampling rate along the scan-line. In general, however, each of the two orthogonal directions of the pixel grid may have a different pixel pitch. Likewise, the scanner may have one optical sampling rate along the scan-line and a different optical sampling rate orthogonal to the scan line.
FIGS. 1-4
further illustrate prior art hand-held scanning and image resampling. In
FIG. 1
, a hand-held scanner
100
is manually moved across the surface of a document
102
along a serpentine path
102
.
FIGS. 2
,
3
A and
3
B illustrate prior art resampling.
FIG. 2
illustrates a scan-line
200
with image intensities measured at discrete points
202
along the scan-line. Also illustrated in
FIG. 2
is a grid of pixel locations
204
. The lines of the pixel grid are ideally parallel to the sides of the document, but may be determined by the first scan-line as discussed above. Separate position sensors (see
FIG. 4
) are used to measure the relative movement of the ends of the scan-line
200
. Given a starting reference position, and given the relative movements of the ends of the scan-line, and given the positions of the measured intensities
202
relative to the ends of the scan-line, the position of each measured intensity
202
may be determined relative to the reference position. The measured intensities
202
, and sufficient information to determine the position of each measured intensity
202
, may be stored in a memory for later rectification, or rectification may be computed in real time as the intensity data is generated.
FIG. 3A
illustrates a series of scan-lines (
300
,
302
,
304
,
306
) determined by the position of the linear sensor array of the scanner as the scanner is moved across the document being scanned. Each intensity measurement from the scanner sensor array has been assigned a lower-case letter.
FIG. 3A
also illustrates a grid of pixel locations. The pixel locations have been assigned location notations relative to an arbitrary reference locat
Crane Randy T
Kommrusch Steven J
Hewlett--Packard Company
Lee Cheukfan
Winfield Augustus W.
LandOfFree
Digital image scanner with compensation for misalignment of... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Digital image scanner with compensation for misalignment of..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Digital image scanner with compensation for misalignment of... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3110869