Method for inspection of protruding features

• US 5,058,177 A
• Filed: 08/31/1990
• Issued: 10/15/1991
• Est. Priority Date: 01/04/1988
• Status: Expired due to Fees

• Current Assignee: Motorola, Inc. (Motorola Solutions, Inc.)
• Original Assignee: Motorola, Inc. (Motorola Solutions, Inc.)
• Inventors: Chemaly, Ephrem A.
• Primary Examiner(s): Razavi, Michael
• Assistant Examiner(s): Couso, Jose L.
• Application Number:
  US07/577,845
• Time in Patent Office: 410 Days
• Field of Search: 382/1, 382/8, 382/30, 382/34, 382/41, 382/48, 382/49, 358/101, 358/106, 358/107, 356/388, 356/375, 356/392, 356/394, 356/398, 364/488
• US Class Current: 382/149
• CPC Class Codes: G06T 2207/30148 Semiconductor; IC; Wafer
• CPC Class Codes: G06T 7/0006 using a design-rule based approach

Abstract

An automatic system for the inspection of a plurality of protruding features of an object employing computer vision. The system examines the position of a tip of the feature and compares it with an ideal position to determine if the object is defective or not. The position of the tip is calculated from a cluster of perceived pixels above a certain threshold of light intensity. In one embodiment, mathematic morphological manipulations of the gray scale perceived pixels assists in the analysis of the positions of the protruding features. The system provides an objective, fast and economical method of inspecting objects, such as electronic packages having a plurality of leads protruding therefrom.

Claims

1. A method for inspecting a plurality of similar protruding features on objects to determine whether the object is "good", or "bad", comprising:

providing a plurality of objects individually at an inspection station, where each object has a plurality of protruding features having tips as a terminal end thereof;

illuminating only the tips of the features of a single object with a light source to generate a pattern of tip images corresponding to the package;

perceiving the pattern of tip images by means of a computer imaging system that receives a digital array of gray scale pixels and sets a brightness threshold for pixels corresponding to the lead tip images;

determining from the pattern of resolved tip images an array of calculated (x,y) image space coordinates corresponding to the package by grouping the pixels above the threshold together in a cluster for each tip and mathematically calculating the centroid of each cluster of pixels;

translating the array of ideal (x,y) coordinates to minimize the means squared error of the (x) coordinate deviation between the ideal array and the calculated array;

computing the difference in the (x) coordinates between the ideal array and the calculated array and computing the best fitting straight line through the (y) coordinates of the calculated array and determining the difference of each (y) value from the best fitting line;

deciding (1) whether the array of calculated (x,y) image space coordinates is within a predetermined acceptable deviation range from the array of ideal (x,y) image space coordinates and denoting the corresponding objects as "good" or (2) whether the array of calculated (x,y) image space coordinates is outside the predetermined acceptable deviation range and then denoting the objects as "bad"; and

sorting the "good" objects from the "bad" objects.

2. The method of claim 1 wherein the step of fitting the image space coordinates within the predetermined range is accomplished by a clustering technique.

3. A method for inspecting a plurality of similar leads of electronic packages to determine whether the package is "good", or "bad", comprising:

providing a plurality of integrated circuit packages individually at an inspection station;

illuminating only the tips of the features of a single package with a light source to generate a pattern of lead tip images corresponding to the package;

perceiving the pattern of lead tip images by means of a computer imaging system that receives a digital array of gray scale pixels and sets a brightness threshold for pixels corresponding to the lead tip images;

determining from the pattern of lead tip images an array of calculated (x,y) image space coordinates corresponding to the package by grouping the pixels above the threshold together in a cluster for each lead and mathematically calculating the centroid of each cluster of pixels;

translating the array of ideal (x,y) coordinates to minimize the mean squared error of the (x) coordinate deviation between the ideal array and the calculated array;

computing the difference in the (x) coordinates between the ideal array and the calculated array and computing the best fitting straight line through the (y) coordinates of the calculated array and determining the difference of each (y) value from the best fitting line;

deciding (1) whether the array of calculated (x,y) image space coordinates is within a predetermined acceptable deviation range from the array of ideal (x,y) image space coordinates and denoting the corresponding objects as "good" or (2) whether the array of calculated (x,y) image space coordinates is outside the predetermined acceptable deviation range and then denoting the object as "bad"; and

sorting "good" objects from the "bad" objects.

4. The method of claim 3 wherein in the step of illuminating the tips, high intensity light is directed onto the tips through a slit in a light baffle.

5. The method of claim 3 further comprising creating the array of ideal (x,y,), image space coordinates by inspecting a plurality of similar objects known to be at least "good", by the method of this invention and taking an average thereof.

6. The method of claim 3 wherein the step of fitting the image space coordinates within the predetermined range is accomplished by a clustering technique.

7. A method for rejecting an object if a feature extending therefrom deviates from a predetermined position relative thereof by more than a selected amount, comprising the steps of:

positioning the object at a precise location;

illuminating the feature to emphasize a distal end thereof;

developing a gray scale image of the distal end of the feature with respect to the precise location of the object;

analyzing the gray scale image using mathematical morphology to locate the distal end of the feature and resolve the distal end to one pixel to determine an actual position of the feature with respect to the object;

translating the analyzed image to minimize the mean squared error of the spatial deviation between the analyzed image and the predetermined position;

computing the difference in between the predetermined position and the analyzed image; and

rejecting the object if the deviation exceeds the selected amount.

8. The method of claim 7 where in the step of illuminating the features to emphasize a distal end thereof, a high intensity light is directed onto the distal end through a slit in a light baffle.