Pattern recognition system

• US 4,823,394 A
• Filed: 04/24/1986
• Issued: 04/18/1989
• Est. Priority Date: 04/24/1986
• Status: Expired due to Fees

• Current Assignee: Kulicke and Soffa Industries Incorporated
• Original Assignee: Kulicke and Soffa Industries Incorporated
• Inventors: Mayer, Ralph, Clarke, Jeffrey, Raghavan, Asuri, Chlebda, Joseph, Berkin, George, Heady, Richard F., Foran, Raymond
• Primary Examiner(s): Boudreau, Leo H.
• Assistant Examiner(s): NOT, DEFINED
• Application Number:
  US06/855,760
• Time in Patent Office: 1,090 Days
• Field of Search: 382/8, 382/22, 382/18, 382/51, 358/107, 356/376, 235/454, 235/483, 235/486
• US Class Current: 382/145
• CPC Class Codes: G06T 2207/30148 Semiconductor; IC; Wafer
• CPC Class Codes: G06T 7/70 Determining position or orientation of objects or cameras camera calibration G06T7/80

Abstract

A system and method for determining the condition and location of an electronic die on a die holding means is disclosed. The system includes means for illuminating the die and means for generating a video image in which the appearance of the die is enhanced relative to the die holding means. This is accomplished by providing a sensing means for sensing primarily only that light reflected from the candidate die. Means for converting the video image into a digital image and processing said digital image to identify the edges of the die is also disclosed. More particularly, means for generating digital data representing the video image, means for defining a corridor in said digital image, said corridor having a width and a length, and means for processing the digital image to obtain a signature element within the digital data are provided. A signature element is the widthwise summation of data within the corridor at a point along the lengthy of the corridor and represents the relative blackness or whiteness in a particular region of the digital image. Also provided are control means for defining a corral within the digitized image of the die, means for obtaining a black count within the corral, and means for analyzing the black count to determine the condition of the die. Means for automatically selecting the light intensity and gray level used in generating the digital image is provided.

Claims

1. A pattern recognition system for determining the position of a candidate die mounted on flexible die holding means comprising:

(a) a light source for illuminating the candidate die;

(b) means for presenting the candidate die such that the flexible die holding means surrounding said die is inclined at an angle greater than about zero degrees with respect to the plane of the candidate die;

(c) means for sensing substantially only that light reflected at an angle of about 90°

from the candidate die to provide a video image, whereby the image of the candidate die is enhanced with respect to the image of the die holding means;

(d) means for generating digital data representing said video image;

(e) memory means for storing said digital data in binary format;

(f) means for defining a corridor across said die in said digital data, said corridor having a width y and a length x;

(g) means for processing said digital data to obtain a signature element, said signature element being the widthwise summation of said y binary data within said corridor at a plurality of x points along the length of said corridor; and

(h) means for analyzing said signature element to distinguish the surface of the candidate die from the surface of the holding means, whereby the position of said candidate die is determined within said corridor.

2. The system as described in claim 1 wherein said means for generating said digital data comprises means for generating a digital data array comprising orthogonal rows and columns of individual pixels, each pixel having a value representative of the intensity of said reflected light.

3. The system as described in claim 2 further comprising means for binarizing said digital data.

4. The system as described in claim 3 wherein said means for defining said corridor comprises means for determining an expected dimension of the candidate die and defining a pair of said columns or rows of said binarized data spaced apart by said expected dimension.

5. The system of claim 1 wherein said means for analyzing comprises means for comparing the value of said signature element to a threshold signature element value.

6. The system of claim 5 wherein said threshold signature element value is approximately 65% of the width of said corridor.

7. The system as described in claim 5 further comprising control means for sequencing the operation of said processing means and said analyzing means so as to obtain a signature element value less than said threshold signature element value and thereby identify an edge of the candidate die.

8. The system as described in claim 7 wherein said control means sequence said operation so as to identify two parallel edges of said die, and further comprising means for analyzing said identified edges to determine a dimension of said die.

9. A pattern recognition system for determining the position of a candidate die on an adhesive film comprising:

(a) a light source for illuminating the candidate die;

(b) means for presenting the candidate die such that the adhesive film surrounding said die is inclined at an angle greater than about a zero degrees with respect to the plane of the candidate die;

(c) means for sensing primarily only that light reflected from the candidate die at a first angle with respect to the plane of the candidate die;

(d) means for generating a digital image of said reflected light, said digital image being comprised of a two dimensional array of orthogonal rows and columns, the component element of the rows and columns being a pixel, each pixel having one value in a range of a values representing the digital intensity of the corresponding portion of said digital image, said range of values comprising more than two values;

(e) memory means for storing said digital image;

(f) means for defining an y corridor in said digital image, said y corridor having a width in pixel rows equal to about the expected y dimension of said candidate die and a location coinciding with the approximate expected location of said candidate die;

(g) means for binarizing aid digital image such that each of said pixels has a value of either 1 or 0;

(h) means for outlining an x signature element of said binarized data, said x signature being the summation of a column of pixel values within the y corridor;

(i) means for obtaining an x edge threshold value representative of the minimum expected x signature element value within the digital image of the candidate die; and

(j) means for indicating whether said signature element is within or without the digital image of said candidate die by comparing said x signature element to said x edge threshold value.

10. The pattern recognition system of claim 9 further

(a) means for defining an x corridor in said digital image, said x corridor having a width in pixel columns equal to the expected x dimension of said candidate die and a location coinciding with the approximate expected location of said candidate die;

(b) means for obtaining a y signature element of said binarized data, said y signature element being the summation of a row of the pixel values within the x corridor; and

(c) means for comparing said y signature element to a y edge threshold value, where said y edge threshold value represents the minimum expected y signature element value within the digital image of the candidate die.

11. The pattern recognition system of claim 10 further comprising control means for sequencing the operation of said x signature element obtaining means, said y signature element obtaining means, said x signature element comparing means, and said y signature element comparing means so as to obtain two x signature elements having a value less than said x edge threshold value and two y signature elements having a value less than said y edge threshold value, and thereby identify the dimensions and location of said candidate die.

12. The pattern recognition system of claim 9 wherein means for sensing comprises means for sensing primarily only that light reflected from the candidate die.

13. A method of determining the position and condition of a candidate die held on an adhesive film with respect to a die transfer tool comprising:

(a) illuminating the candidate die with light from a light source;

(b) presenting the canditate die such that the adhesive film surrounding said die is inclined at an angle greater than zero degrees with respect to the plane of the candidate die;

(c) sensing primarily only that light reflected from the candidate die at a first angle with respect to the plane of the candidate die;

(d) generating digital data representing said reflected light;

(e) storing said digital data in a memory;

(f) defining a y corridor in said memory, said y corridor having a width equal to about the y dimension of the candidate die;

(g) defining an x corridor in said memory, said x corridor having a width equal to about the x dimension of the candidate die;

(h) obtaining an x signature element, said x signature element being the widthwise summation of said data at a point within aids y corridor;

(i) obtaining a y signature element, said y signature element being the widthwise summation of said data at a point within said x corridor; and

(j) comparing the value of said y signature element or said x signature element to a threshold signature element value.

14. The method of claim 13, wherein said digital data comprises a data array of orthogonal rows and columns of individual pixels, each pixel having a digital value representative of the intensity of said reflected light.

15. The method as described in claim 14 further comprising the step of binarizing said digital data prior to the first obtaining step.

16. The method of claim 15 wherein said y corridor defining step comprises determining the expected y dimension of the candidate die and defining a pair of said rows of said binarized data spaced apart by said expected dimension and wherein said x corridor defining step comprises determining the expected x dimension of the candidate die and defining a pair of said columns of said binarized data spaced apart by said expected dimensions.

17. The method as defined in claim 13 wherein said threshold signature element value is approximately 65% of the corridor width.

18. The method as defined in claim 17, further comprising repeating said x or y obtaining step and said comparing step until the value of said y signature element or said x signature element is less than said threshold signature element value.

19. The method as defined in claim 18 further comprising controlling said repeating step so as to identify two parallel edges of said die.