Image analysis – Image transformation or preprocessing
Reexamination Certificate
1998-04-06
2002-04-30
Mehta, Bhavesh (Department: 2621)
Image analysis
Image transformation or preprocessing
C345S648000
Reexamination Certificate
active
06381375
ABSTRACT:
RESERVATION OF COPYRIGHT
The disclosure of this patent document contains material that is subject to copyright protection. The owner thereof has no objection to facsimile reproduction by anyone of the patent document or the patent disclosure, as it appears in the U.S. Patent and Trademark Office patent file or records, but otherwise reserves all copyright rights whatsoever.
BACKGROUND OF THE INVENTION
The invention pertains to image processing and, more particularly, to methods and apparatus for generating a projection of an image.
Image processing refers to the automated analysis of images to determine the characteristics of features shown in them. It is used, for example, in automated manufacturing lines, where images of parts are analyzed to determine placement and alignment prior to assembly. It is also used in quality assurance where images of packages are analyzed to insure that product labels, lot numbers, “freshness” dates, and the like, are properly positioned and legible.
Image processing has non-industrial applications, as well. In biotechnology research, it can be used to identify constituents of microscopically imaged samples, or growth patterns in culture media. On the macroscopic scale, it can be used in astronomical research to find objects in time-lapse images. Meteorologic, agricultural and defense applications of image processing include the detection and analysis of objects in satellite images.
In applications where speed counts, image processing systems typically do not analyze every feature in an image but, rather, infer most from analysis of a significant few.
Thus, for example, the position and orientation of semiconductor chip can usually be inferred from the location and angle of one of its edges. Likewise, the center of an object can often be estimated from the positions of its edges.
One image processing technique for discerning salient features in an image is projection. Though a similarly named technique is commonly used in the visual arts to add depth to paintings and drawings, the image processing counterpart is used to reduce complexity and, thereby, to reduce the computational resources necessary an image. More particularly, image processing projections reduce the dimensions of an image while maintaining relative spacing among features along axes of interest.
In machine vision, the most common form of projection involves reducing an image from two dimensions to one. A projection taken along the x-axis, for example, compresses all features in the y-dimension and, therefore, facilitates finding their positions along the x-axis. Thus, it can be used to find rapidly the width of an imaged object. A projection along the y-axis, on the other hand, speeds finding the height or vertical extent of an object.
These types of projections are typically made by summing the intensities of pixels at each point along the respective axis. For example, the projection of an image along the x-axis is made by summing the intensities of all pixels whose x-axis coordinate is zero; then summing those whose x-axis coordinate is one; those whose x-axis coordinate is two; and so forth. From those sums can be inferred the x-axis locations of significant features, such as edges.
In many instances, it is desirable to project images along arbitrary axes, not merely the pixel grid defined by the x- and y-axes. For example, if a semiconductor chip is not aligned with the x- and y-axes of the video camera which images of it, the locations of the chip leads can best be determined by generating a projection along an axis of the chip itself—not the axis of the camera.
The prior art offers two principal solutions: one that operates quickly, but with low accuracy; the other, that operates more slowly, but with greater accuracy. The former involves summing the intensities of pixels falling between parallel lines normal to the angle of the projection and spaced in accord with the projection bin width. The sum of intensities formed between each pair of neighboring lines is stored in a corresponding projection bin. Though such techniques can be employed in tools that operate sufficiently quickly to permit their use in real time, their accuracy is typically too low for many applications.
The other prior art solution is to transform the image prior to taking its projection. One common transformation tool used for this purpose is referred to as affine transformation, which resizes, translates, rotates, skews and otherwise transforms an image. Conventional affine transformation techniques are typically slow and too computationally intensive for use in real-time applications. Specifically, the prior art suggests that affine transforms can be accomplished by mapping a source image into a destination image in a single pass. For every pixel location in the destination image, a corresponding location in the source image is identified. In a simplistic example, every pixel coordinate position in the destination image maps directly to an existing pixel in the source. Thus, for example, the pixel at coordinate (4,10) in the source maps to coordinate (2,5) in the destination; the pixel at (6,10) in the source, to (3,5) in the destination; and so on.
However, rarely do pixels in the source image map directly to pixel positions in the destination image. Thus, for example, a pixel at coordinate (4,10) in the source may map to a location (
2
.
5
,
5
.
33
) in the destination. This can be problematic insofar as it requires interpolation to determine appropriate pixel intensities for the mapped coordinates. In the example, an appropriate intensity might be determined as a weighted average of the intensities for the source pixel locations (
2
,
5
), (
3
,
5
), (
2
,
6
), and (
3
,
6
).
The interpolation of thousands of such points consumes both time and processor resources. Conventional affine transform tools must typically examine at least four points in the source image to generate each point in the destination image. This is compounded for higher-order transformations, which can require examination of many more points for each interpolation.
Although prior art has suggested the use of multiple passes (i.e., so-called separable techniques) in performing specific transformations, such as rotation, no suggestion is made as to how this might be applied to general affine transforms, e.g., involving simultaneous rotation, scaling, and skew.
An object of this invention is to provide improved systems for image processing and, more particularly, improved methods and apparatus for image projection.
A more particular object is to provide such methods and apparatus as facilitate generating projections of images (or objects) that have been rotated, skewed, scaled, sheared, transposed or otherwise transformed.
Another object of the invention is to provide such methods and apparatus as permit rapid analysis of images, without undue consumption of resources.
Still another object of the invention is to provide such methods and apparatus as are readily adapted to implementation conventional digital data processing apparatus, e.g., such as those equipped with commercially available superscalar processors—such as the Intel Pentium MMX or Texas Instruments C80 microprocessors.
SUMMARY OF THE INVENTION
The foregoing objects are among those attained by the invention, which provides methods and apparatus for generating projections while concurrently rotating, scaling, translating, skewing, shearing, or subjecting the image to other affine transforms. In an exemplary aspect, the invention provides methods for generating a projection of an image by generating an “intermediate” image via a one-dimensional affine transformation of the source along a first axis, e.g., the y-axis. The intermediate image is subjected to a second one-dimensional affine transformation along a second axis, e.g., the x-axis. The resultant image is then projected along a selected one of these axes.
According to related aspects of the invention, there are provided methods as described above in which the first one-dimensional transformation determines a ma
Cognex Corporation
Mehta Bhavesh
Patel Kanji
Powsner David J.
LandOfFree
Methods and apparatus for generating a projection of an image does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Methods and apparatus for generating a projection of an image, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Methods and apparatus for generating a projection of an image will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2918863