Labeled projection of digital images

• US 5,901,241 A
• Filed: 12/02/1996
• Issued: 05/04/1999
• Est. Priority Date: 05/02/1994
• Status: Expired due to Term

• Current Assignee: Cognex Corporation
• Original Assignee: Cognex Corporation
• Inventors: Michael, David J., Koljonen, Juha
• Primary Examiner(s): Mancuso, Joseph
• Application Number:
  US08/759,853
• Time in Patent Office: 883 Days
• Field of Search: 382/146, 382/158, 382/168, 382/209, 382/291, 382/286, 382/194, 348/126
• US Class Current: 382/150
• CPC Class Codes: G06T 2207/30152 Solder
• CPC Class Codes: G06T 7/0004 Industrial image inspection
• CPC Class Codes: G06T 7/62 of area, perimeter, diameter or volume
• CPC Class Codes: G06T 7/66 of image moments or centre of gravity
• CPC Class Codes: G06T 7/70 Determining position or orientation of objects or cameras camera calibration G06T7/80
• CPC Class Codes: G06V 10/48 by mapping characteristic values of the pattern into a parameter space, e.g. Hough transformation

Abstract

The invention automatically inspects the bond of a wire to a contact pad on a semiconductor chip. The apparatus includes a movable platform for holding semiconductor chips situated in lead frames; a video camera for sensing images; illumination means for illuminating a chip in a lead frame; an image processor to digitize and analyze the images; a bonding mechanism; and a host controller electronically connected to bonding mechanism, movable platform, video camera, and image processor. Image processor locates a bond on a pad in a digitized image and provides a first nominal center of ball bond image. The invention aligns the center of a polar coordinate transform image having one or more segments with the nominal center of ball bond image and evaluates ball bond image using the polar coordinate transform image to create a polar projection histogram array and store it. An edge filter is applied to histogram array to detect peaks and store their number and values. Polar coordinate transform image is aligned with a next nominal ball center location until a predetermined number of potential ball center locations is exhausted. The maximum peak in the list of stored peaks is selected as the radius of bond from which the size and position of bond is computed and reported to host controller for further action.

Claims

1. In a computer system having a means for sensing objects and digitizing and analyzing images of said objects, a method for automatically measuring designated geometric properties of expected features in said objects, comprising the steps of:

a. constructing a projection formula to define the intensity of the pixels in a labeled transform image having the maximum dimensions of the designated geometric properties of an expected feature to be measured;

b. creating an array in which to store results, said array having a number of elements equal to the maximum dimensions of the designated geometric properties of said expected feature wherein each element has an index, each index corresponding to a particular pixel intensity of the labeled transform image;

c. specifying a number of potential expected features to be evaluated within a digitized image;

d. acquiring a digitized image of said object;

e. storing said digitized image of said object as an intensity image in which pixels are assigned grey level values corresponding to intensities sensed from said object;

f. receiving a potential expected feature location;

g. aligning said labeled transform image with said potential expected feature location such that designated coordinates of said potential expected feature location correspond to an x,y location in said labeled transform image;

h. evaluating said potential expected feature location and said labeled transform image, such that each pixel value in said intensity image having said potential expected feature location is summed into a corresponding element of said array indexed by the pixel value at the corresponding x,y coordinate in said labeled transform image, such that a one-dimensional histogram is created and stored in said array;

I. analyzing peaks in said histogram for correspondence to said designated geometric properties and storing the values thereof;

j. selecting a next potential expected feature location and applying steps g through i to said next potential expected feature location until the specified number of potential expected feature locations is exhausted;

k. locating the best result from said step of analyzing peaks in said histogram such that the x,y coordinates of said best result are identified as said expected feature location and used in computing the designated geometric properties of said expected feature.

2. The method of claim 1, wherein the step of constructing a labeled transform image further comprises the step of dividing said labeled transform image into a plurality of n labeled segments according to a predetermined method until n labeled segments have been created.

3. The method of claim 2, wherein the step of evaluating said potential expected feature location and said labeled transform image further comprises the steps of:

a. generating n labeled segments; and

b. creating n labeled histograms, each representing one of said labeled segments, such that n labeled subsets of said expected feature location are located and their characteristics can be used in computing the designated geometric properties of said expected feature.

4. In a computer system having a means for sensing objects and digitizing and analyzing images of said objects, an apparatus for automatically measuring designated geometric properties of expected features in said objects, comprising:

a. a means for constructing a projection formula to define the intensity of the pixels in a labeled transform image having the maximum dimensions of the designated geometric properties of an expected feature to be measured;

b. a means for creating an array in which to store results, said array having a number of elements equal to the maximum dimensions of the designated geometric properties of said expected feature wherein each element has an index, each index corresponding to a particular pixel intensity of the labeled transform image;

c. a means for specifying a number of potential expected features to be evaluated within a digitized image;

d. a means for acquiring a digitized image of said object;

e. a means for storing said digitized image of said object as an intensity image in which pixels are assigned grey level values corresponding to intensities sensed from said object;

f. a means for receiving a potential expected feature location;

g. a means for aligning said labeled transform image with said potential expected feature location such that designated coordinates of said potential expected feature location correspond to an x,y location in said labeled transform image;

h. evaluating said potential expected feature location and said labeled transform image, such that each pixel value in said intensity image having said potential expected feature location is summed into a corresponding element of said array indexed by the pixel value at the corresponding x,y coordinate in said labeled transform image, such that a one-dimensional histogram is created and stored in said array;

I. a means for analyzing peaks in said histogram for correspondence to said designated geometric properties and storing the values thereof;

j. a means for selecting a next potential expected feature location and applying means g through i to said next potential expected feature location until the specified number of potential expected feature locations is exhausted;

k. a means for locating the best result from said step of analyzing peaks in said histogram such that the x,y coordinates of said best result are identified as said expected feature location and used in computing the designated geometric properties of said expected feature.

5. The apparatus of claim 4, wherein said means for constructing a labeled transform image further comprises a means for dividing said labeled transform image into a plurality of n labeled segments according to a predetermined method until n labeled segments have been created.

6. The apparatus of claim 5, wherein said means for evaluating said potential expected feature location and said labeled transform image further comprises:

a. a means for generating n labeled segments; and

b. a means for creating n labeled histograms, each representing one of said labeled segments, such that n labeled subsets of said expected feature location are located and their characteristics can be used in computing the designated geometric properties of said expected feature.