Reduction of false alarms in pcb inspection

Abstract

A method for automatically optically inspecting an electrical circuit, comprising: acquiring at least one optical image of an electrical circuit; generating at least one first inspection image from the at least one image and determining regions of candidate defects therefrom; generating at least one additional inspection image for regions surrounding candidate defects, said at least one additional inspection image at least partially including optical information not included in the at least one first inspection image; and determining whether the candidate defect is a specious defect by inspecting the at least one additional inspection image.

1 . A method of determining defects in electrical circuits, comprising: 
 acquiring an optical image of an electrical circuit;    generating a representation of the electrical circuit from the image;    analyzing at least part of the representation at a region near a candidate to detect an attribute of the electrical circuit;    revising the representation in response to a detected attribute of the electrical circuit;    processing the revised image to determine the presence of a real defect the electrical circuit.    
   
   
       2 . A method according to  claim 1  wherein the representation is a representation of contours in the electrical circuit.  
   
   
       3 . A method according to  claim 2  wherein the contours are representative of putative edges between metalized portions and substrate portions in the electrical circuit.  
   
   
       4 . A method according to  claim 1  wherein revising includes: 
 evaluating a region in the representation to determine whether it is a metal; and    revising the contours if metal is detected.    
   
   
       5 . A method according to  claim 4  wherein the metal is oxidized or non-oxidized metal.  
   
   
       6 . A method according to  claim 5  wherein the metal is copper.  
   
   
       7 . A method according to  claim 4  wherein the evaluating includes analyzing a monochromatic image of the region.  
   
   
       8 . A method according to  claim 4  wherein the evaluating includes analyzing a polychromatic image of the region.  
   
   
       9 . A method according to  claim 1  wherein revising includes calculating a directed convex hull for a section of the contours, and replacing the section with the directed convex hull.  
   
   
       10 . A method according to  claim 5  wherein revising includes locating points representing corners in a contour located near the candidate defect.  
   
   
       11 . A method according to  claim 10  wherein the directed convex hull extends between corner contour points.  
   
   
       12 . A method according to any of  claim 1  wherein revising includes detecting contour points that are situated on oxidized conductors.  
   
   
       13 . A method according to  claim 12  wherein revising includes identifying at least three contiguous oxidized contour points.  
   
   
       14 . A method according to  claim 13  wherein revising includes identifying from among the at least three contiguous oxidized contour points, at least two oxidized contour points each of which is bounded by a contour point that is not oxidized.  
   
   
       15 . A method according to  claim 13  wherein revising includes identifying from among the at least three contiguous oxidized contour points, at least one oxidized contour point that is also a corner.  
   
   
       16 . A method according to  claim 15  wherein revising includes identifying from among the at least three contiguous oxidized contour points an additional contour point that is bounded by a contour point that is not oxidized.  
   
   
       17 . A method according to  claim 14  wherein revising includes calculating a directed convex hull between any pair of: oxidized contour points which are bounded by a combination of contour points that are not oxidized and oxidized contour points that are a corner  
   
   
       18 . A method according to  claim 1  wherein the attribute comprises an optical attribute characteristic of an oxide.  
   
   
       19 . A method according to  claim 1  and wherein the attribute is an optical attribute comprising a level of reflective intensity in at least two spectral ranges.  
   
   
       20 . A method according to  claim 19  and wherein the attribute is an optical attribute comprising a level of reflective intensity in at least three spectral ranges.  
   
   
       21 . A method according to  claim 19  wherein the spectral ranges comprise at least two of red, green and blue.  
   
   
       22 . A method according to  claim 20  wherein the spectral ranges comprise at least two of red, green and blue.  
   
   
       23 . A method according to  claim 1  wherein processing is performed in response to the detection of metal in the vicinity of the candidate defect.  
   
   
       24 . A method according to  claim 1  wherein processing is performed in response to the detection of an oxide in the vicinity of the candidate defect.  
   
   
       25 . A method according to  claim 24  wherein the detection of an oxide is performed on contour points located along a putative border between metal conductor and substrate.  
   
   
       26 . A method according to  claim 24  wherein the detection of an oxide is performed on the values obtained by pixels for regions in the electrical circuit that are located near the candidate defect.  
   
   
       27 . A method according  claim 1  wherein revising includes revising a putative contour to include a directed convex hull.  
   
   
       28 . A method according to  claim 27  wherein revising includes measuring the distance between putative contours, wherein at least part of the putative contours includes the directed convex hull.  
   
   
       29 . A method for inspecting printed circuit boards for defects, comprising: 
 acquiring an image of a printed circuit board;    performing automated optical inspection on the image to identify candidate defects on the printed circuit board;    providing an image of a region surrounding a candidate defect to a processor, said image including a map of putative contours corresponding to the candidate defect;    analyzing the image of the region surrounding the candidate defect;    revising the putative contour map to provide a revised contour map in response to a predetermined condition occurring in the image, the revised contour map including revised contours corresponding to the candidate defect; and    inspecting the revised contour map to determine the presence of an actual defect in the electrical circuit.    
   
   
       30 . A method according to  claim 29  wherein the image comprises contours that are determined using a first method for determining contours.  
   
   
       31 . A method according to  claim 30  wherein the first method employs a difference of Gaussians calculation.  
   
   
       32 . A method according to  claim 29  wherein the contours are representative of putative edges between metalized portions and substrate portions in the printed circuit board.  
   
   
       33 . A method according to  claim 29  wherein revising includes: 
 evaluating the image to detect whether the region includes a metal; and    revising the contours if metal is detected.    
   
   
       34 . A method according to  claim 33  wherein the metal is oxidized or non-oxidized metal.  
   
   
       35 . A method according to  claim 34  wherein the metal is copper.  
   
   
       36 . A method according to  claim 33  wherein the evaluating includes analyzing a monochromatic image of the region.  
   
   
       37 . A method according to  claim 33  wherein the evaluating includes analyzing a polychromatic image of the region.  
   
   
       38 . A method according to  claim 29  wherein revising contours includes calculating a directed convex hull for a section of the putative contours, and replacing the section with the directed convex hull.  
   
   
       39 . A method according to  claim 29  wherein the predetermined condition comprises an optical attribute characteristic of a metal in the image.  
   
   
       40 . A method according  claim 29  wherein revising includes revising a putative contour to include a directed convex hull.  
   
   
       41 . A method according to  claim 40  wherein revising includes measuring the distance between putative contours, wherein at least part of the putative contours includes the directed convex hull.