William W Silver

US Patent:

6408109, Jun 18, 2002

Filed:

Oct 7, 1996

Appl. No.:

08/727721

Inventors:

William Silver - Weston MA
Arman Garakani - Wellesley MA
Aaron Wallack - Natick MA

Assignee:

Cognex Corporation - Natick MA

International Classification:

G06K 982

US Classification:

382300, 382190, 382199, 382266

Abstract:

The invention provides a fast, computationally inexpensive, and highly accurate method and apparatus for edge detection in a digital image, even for edges that are not substantially parallel to the axes of the pixel grid, by exploiting computationally inexpensive estimates of gradient magnitude and direction. In particular, the method includes the steps of: estimating gradient magnitude and direction at a plurality of regularly-spaced pixel points in the image so as to provide a plurality of estimates of gradient magnitude and direction, each such estimate being associated with a respective gradient point of a regularly-spaced gradient grid; using gradient direction associated with each gradient point to select a respective set of neighboring gradient points; comparing gradient magnitude associated with each gradient point with each gradient magnitude of the respective set of neighboring gradient magnitudes so as to determine which of the gradient magnitudes is a local maximum in approximately the gradient direction; and using the local maximum of gradient magnitude and a set of neighboring gradient magnitudes to determine an interpolated edge position along a one-dimensional gradient magnitude profile. Another aspect of the invention for providing two-dimensional edge position interpolation further includes the step of determining a plane position line normal to the gradient direction of a gradient point associated with the local maximum of gradient magnitude, the plane position line also passing through the interpolated edge position, along which plane position line at least one two-dimensional interpolated position of the edge can be determined.

US Patent:

6457032, Sep 24, 2002

Filed:

Nov 15, 1997

Appl. No.:

08/971241

Inventors:

William Silver - Weston MA

Assignee:

Cognex Corporation - Natick MA

International Classification:

G06F 1710

US Classification:

708315, 708319

Abstract:

A system is provided for digitally processing a one-dimensional digital signal, including convolving the one-dimensional digital signal with a function that is the (n+1) difference of an n order discrete piecewise polynomial kernel so as to provide a second one-dimensional digital signal. Here, ânâ is at least 1, the polynomial kernel has a plurality of non-zero elements, the function has a plurality of non-zero elements and at least one zero element, and the function has fewer non-zero elements than the polynomial kernel has non-zero elements. Then, the second one-dimensional digital signal is discretely integrated n+1 times. Also, multi-dimensional signals are dimensionally separated and processed using a function for each dimension. In addition to providing parabolic kernels and approximate Gaussian kernels, the invention can provide many other filter shapes, thereby allowing unprecedented flexibility in matching filter specifications to particular signal processing applications, while also providing high computational efficiency.

US Patent:

6469734, Oct 22, 2002

Filed:

Apr 29, 2000

Appl. No.:

09/563010

Inventors:

Sanjay Nichani - Natick MA
Robert Wolff - Sherborn MA
William Silver - Weston MA
David A. Schatz - Needham MA

Assignee:

Cognex Corporation - Natick MA

International Classification:

H04N 718

US Classification:

348152, 348164

Abstract:

A two-dimensional (2-D) machine-vision safety-solution involving a method and apparatus for performing high-integrity, high efficiency machine vision. The machine vision safety solution digitally filters a configurable number of time-sequenced images of a viewed area to create a filtered image. A textured background is used as a static target and viewed under ambient light, or alternatively a texture is observed under IR light. An artificial texture can be created with a particularized pattern. Each new image is then compared to the previously created model and a pixel-oriented difference image is calculated. The difference image is then processed for segmentation features attributable to an intruder object. For segmentation results that exceed a predetermined threshold, an alarm condition is set.

US Patent:

6639624, Oct 28, 2003

Filed:

Apr 11, 2000

Appl. No.:

09/547590

Inventors:

Ivan Bachelder - Newton MA
Karen Roberts - Natick MA
Aaron Wallack - Natick MA
William Silver - Weston MA
Edward Collins - Smithfield RI
Karen Sarachik - Newtonville MA

Assignee:

Cognex Corporation - Natick MA

International Classification:

H04N 718

US Classification:

348 87, 348 94, 348 95, 348126, 348129, 348130

Abstract:

A machine vision method for locating a leaded electronic device, such as integrated circuit chips, includes searching the image with a template depicting only a portionâe. g. , one or two âsideâ edgesâof each of one or more of the leads expected on the device. The template can include portions representing expected leads, as well as unexpected leads or unexpected lead positions. Positive and negative weightings are applied to such portions, as appropriate. A location of the device, or of one or more of its leads, is based on the results of the search.

US Patent:

6658145, Dec 2, 2003

Filed:

Dec 22, 2000

Appl. No.:

09/746147

Inventors:

William Silver - Weston MA
Aaron S. Wallack - Natick MA
Adam Wagman - Framingham MA

Assignee:

Cognex Corporation - Natick MA

International Classification:

G06K 900

US Classification:

382149, 382141, 382181, 382190, 382197, 382209, 382276, 348 92, 348125

Abstract:

A method and apparatus are provided for identifying differences between a stored pattern and a matching image subset, where variations in pattern position, orientation, and size do not give rise to false differences. The invention is also a system for analyzing an object image with respect to a model pattern so as to detect flaws in the object image. The system includes extracting pattern features from the model pattern; generating a vector-valued function using the pattern features to provide a pattern field; extracting image features from the object image; evaluating each image feature, using the pattern field and an n-dimensional transformation that associates image features with pattern features, so as to determine at least one associated feature characteristic; and using at least one feature characteristic to identify at least one flaw in the object image. The invention can find at least two distinct kinds of flaws: missing features, and extra features. The invention provides pattern inspection that is faster and more accurate than any known prior art method by using a stored pattern that represents an ideal example of the object to be found and inspected, and that can be translated, rotated, and scaled to arbitrary precision much faster than digital image re-sampling, and without pixel grid quantization errors.

US Patent:

6671049, Dec 30, 2003

Filed:

Oct 30, 2000

Appl. No.:

09/699540

Inventors:

William Silver - Weston MA

Assignee:

Cognex Corporation - Natick MA

International Classification:

G01B 1100

US Classification:

356401, 356400, 356399

Abstract:

An article of manufacture bearing an alignment target having at least one of the following attributes: it has no significant rotational symmetry; it has no significant orientation self-similarity or size self-similarity; it is not primarily composed of circles, circular arcs, straight lines, or right angles; and it may have other unique attributes. In one embodiment, the specific alignment targets are shapes consisting of at least one âgeneralized polar polygon. â Such a figure will include radial lines, circular arcs, and spiral segments when drawn in the Cartesian plane. In another embodiment, the specific alignment targets are members of a family of âfanâ shapes.

US Patent:

6690842, Feb 10, 2004

Filed:

Apr 2, 2002

Appl. No.:

10/114357

Inventors:

William Silver - Weston MA
Arman Garakani - Wellesley MA
Aaron Wallack - Natick MA

Assignee:

Cognex Corporation - Natick MA

International Classification:

G06L 932

US Classification:

382300, 382199

Abstract:

The invention provides a fast, computationally inexpensive, and highly accurate method and apparatus for edge detection in a digital image, even for edges that are not substantially parallel to the axes of the pixel grid, by exploiting computationally inexpensive estimates of gradient magnitude and direction. In particular, the method includes the steps of: estimating gradient magnitude and direction at a plurality of regularly-spaced pixel points in the image so as to provide a plurality of estimates of gradient magnitude and direction, each such estimate being associated with a respective gradient point of a regularly-spaced gradient grid; using gradient direction associated with each gradient point to select a respective set of neighboring gradient points; comparing gradient magnitude associated with each gradient point with each gradient magnitude of the respective set of neighboring gradient magnitudes so as to determine which of the gradient magnitudes is a local maximum in approximately the gradient direction; and using the local maximum of gradient magnitude and a set of neighboring gradient magnitudes to determine an interpolated edge position along a one-dimensional gradient magnitude profile. Another aspect of the invention for providing two-dimensional edge position interpolation further includes the step of determining a plane position line normal to the gradient direction of a gradient point associated with the local maximum of gradient magnitude, the plane position line also passing through the interpolated edge position, along which plane position line at least one two-dimensional interpolated position of the edge can be determined.

US Patent:

6812933, Nov 2, 2004

Filed:

Oct 30, 2000

Appl. No.:

09/699590

Inventors:

William Silver - Weston MA

Assignee:

Cognex Technology and Investment - Mt. View CA

International Classification:

G09G 500

US Classification:

345619, 345419, 382199, 382203

Abstract:

A method of accurately rendering algebraically defined shapes that can be used to render alignment targets, or for other purposes, is disclosed. The method can render shapes as a bitmap image for machine vision alignment training, for computer display for human observation, and for computer printout for documentation. The method can aid in rendering targets on the objects to be located by providing input to a computer-aided design (CAD) system, or by directly controlling devices such as laser engravers. The rendering method is fast, extremely accurate, and can handle substantially arbitrary shapes. Targets are defined by a real-valued function of position in the real plane that gives the distance from that position to the nearest point on a target boundary. The function gives distance as a positive value for points inside the figure, and as a negative value for points outside the figure. An edge model is used to map distance to real-valued image intensity.

Address:

10 Edgewater Dr APT 5G, Coral Gables, FL 33133

VIN:

1G6DM57T170147671

Make:

CADILLAC

Model:

CTS

Year:

2007