Method and apparatus for registering color separation film

• US 4,849,914 A
• Filed: 09/22/1987
• Issued: 07/18/1989
• Est. Priority Date: 09/22/1987
• Status: Expired due to Fees

• Current Assignee: Polychrome Systems Incorporated
• Original Assignee: Opti-Copy Incorporated
• Inventors: Medioni, Gerard G. R., Prohaska, Timothy F., Poretta, Lynn R., Wilson, Monti R.
• Primary Examiner(s): Sikes, William L.
• Assistant Examiner(s): Gonzalez, Frank
• Application Number:
  US07/099,800
• Time in Patent Office: 665 Days
• Field of Search: 356/399, 356/400, 356/401, 364/526, 364/474.02, 364/463, 364/524, 364/525, 364/559, 358/75, 358/76, 358/77, 358/78, 358/79, 358/80, 358/101, 358/107, 358/283, 358/93, 377/17, 377/18, 226/3, 226/15, 226/16, 226/20, 226/21, 382/1, 382/8, 382/42, 382/44-46, 382/50, 382/52, 382/54, 83/30, 83/33, 83/13, 83/701
• US Class Current: 382/112
• CPC Class Codes: G03F 9/00 Registration or positioning of originals, masks, frames, photographic sheets or textured or patterned surfaces, e.g. automatically G03F7/22 takes precedence; preparation of photographic masks G03F1/00; within photographic printing apparatus for making copies G03B27/00
• CPC Class Codes: G06T 3/14 Transformations for image registration, e.g. adjusting or mapping for alignment of images
• CPC Class Codes: G06T 7/33 using feature-based methods
• CPC Class Codes: G06V 10/245 by locating a pattern; Special marks for positioning

Abstract

An algorithmic based method and apparatus for registering halftone color separations using edge based registration techniques. Digital pictures are taken at two locations on each film, and, following optimal thresholding of the data, the edges are located by application of a Laplacian of a Gaussian edge detection operator. Sampling is used to reduce the computational requirements, and full resolution is restored by interpolation. The edge locations are corrected for operator edge bias, and the corrected edge points are linked into segments which are registered by the parameter space method. The initial offset is corrected for size differences by application of least squares correction techniques.

Claims

1. In a method of registering a pair of color separation films containing halftone dot detail and macroscopic edges which align when the films are registered, the steps of:

selecting one of the films as a reference film;

recording first and second pictures of the reference film centered at spaced apart first and second locations, respectively;

recording third and fourth pictures of the other film centered at locations thereon which approximately correspond to but may be offset from the respective first and second locations;

extracting from said first and third pictures the locations of corresponding macroscopic edges on the reference film and on the other film;

extracting from said second and fourth pictures the locations of corresponding macroscopic edges on the reference film and on the other film; and

determining the offset between the edge locations on the reference film and the corresponding edge locations on the other film to thereby determine the positional adjustment of the other film necessary for registration with the reference film.

2. The method of claim 1, wherein each of said extracting steps comprises:

optimally thresholding the data of each picture; and

applying to the optimally thresholded data an edge detection operator adapted to detect the edge locations.

3. The method of claim 2, wherein the edge detection operator has a variable length scale determinative of the sensitivity of the operator to detect edges, and including the step of selecting a scale for the operator which is too large to allow the operator to detect edges around individual halftone dots but not too large to detect macroscopic edges in the halftone detail.

4. The method of claim 3, wherein said operator is the Laplacian of a Gaussian.

5. The method of claim 1, wherein each of said extracting steps comprises:

optimally thresholding the data of each picture;

determining the location of a plurality of edge points on each edge from the optimally thresholded data;

linking adjacent pairs of edge points on each edge to create line segments each having a length extending between two adjacent edge points;

establishing a fitting tolerance;

combining adjacent line segments with one another one at a time to create longer segments;

rejecting any line segment which exceeds the fitting tolerance; and

using each rejected line segment as the initial segment of another longer segment.

6. The method of claim 5, wherein said determining step comprises using the parameter space method to match the longer segments of the reference picture and of the other picture.

7. The method of claim 6, wherein said determining step further comprises applying a least squares correction offset to the segment matches.

8. The method of claim 1, wherein each picture recorded is a digital picture and including the step of optimally thresholding the data of each picture prior to extracting the edge locations therefrom.

9. The method of claim 8, wherein each of said extracting steps comprises applying an edge detector operator to the optimally thresholded data in a manner to obtain a convolution output from which edge points on the macroscopic edges can be derived.

10. The method of claim 9, wherein each of said extracting steps further comprises sampling the convolution output systematically to obtain a reduced resolution determination of the locations of edge points on the edges of the pictures and the edge direction at each edge point.

11. The method of claim 10, including the step of interpolating to obtain an interpolated full resolution determination of the locations of edge points on the edges of the pictures.

12. The method of claim 11, wherein each of said extracting steps further comprises the step of correcting the location of each edge point for the edge bias of the operator.

13. The method of claim 12, wherein each of said extracting steps comprises producing linear segments from the edge points of each picture.

14. The method of claim 13, wherein the step of producing linear segments comprises the steps of:

linking the edge points from the reduced resolution determinations of the edge point locations in a manner to obtain line segments;

tracing the line segments to obtain distinct chains of linked edge points; and

fitting said chains to segments defined by the corrected edge points of the interpolated full resolution determination.

15. The method of claim 14, wherein said fitting step comprises:

establishing a fitting tolerance;

combining neighboring line segments to create longer segments so long as each segment which is combined with others is within said tolerance;

rejecting line segments which exceed said tolerance; and

using each rejected line segment as the initial segment of another longer segment obtained by combining said initial segment with additional segments within said tolerance.

16. The method of claim 15, wherein said determining step comprises using the parameter space method to match the longer segments of the reference picture and of the other picture.

17. The method of claim 16, wherein said determining step further comprises applying a least squares correction offset to the segment matches.

18. The method of claim 14, wherein said determining step comprises using the parameter space method to match the longer segments of the reference picture and of the other picture.

19. The method of claim 14, wherein said determining step comprises:

determining the length, center coordinate and angular orientation of each fitted chain of linked edge points;

comparing the angular orientations of each fitted chain of the reference picture with each possibly matching fitted chain of the other picture;

accepting pairs of fitted chains that do not differ in angular orientation by more than a prescribed threshold value;

constructing a rectangular painting window centered at a location determined by the offset between the center coordinates of each accepted pair of fitted chains, each window having a preselected width and a length and angular orientation determined by the lengths and angular orientations of the fitted chains in each accepted pair;

assigning to each coordinate point in a preselected coordinate frame a score based on the number of painting windows the coordinate falls within; and

determining the offset by ascertaining the coordinate point having the highest score.

20. The method of claim 19, including the step of correcting said offset for size differences between the reference picture and the other picture.

21. The method of claim 20, wherein said correcting step comprises applying to said offset a least squares correction offset.

22. Apparatus for effecting registration between a pair of color separation films containing halftone dot detail and macroscope edges which align when the films are registered, said apparatus comprising:

a digital camera operable to record digital pictures at locations on each film with the locations on the respective films corresponding but possibly being offset;

means for extracting from each digital picture the locations of macroscopic edges thereon; and

means for calculating the offset between the edges on corresponding pictures of the films thereby determining the relative movement necessary to bring the films into registration.

23. Apparatus as set forth in claim 22, wherein said extracting means comprises:

means for optimally thresholding the data of each picture; and

means for applying to the optimally thresholded data an edge detection operator characterized by the ability to detect the locations of edges in the pictures.

24. Apparatus as set forth in claim 23, wherein the edge detection operator is characterized by a variable length scale determinative of the sensitivity of the operator to detect edges, said scale being too large to detect edges around individual halftone dots but not too large to detect macroscopic edges in the halftone detail.

25. Apparatus as set forth in claim 24, wherein the edge detection operator is the Laplacian of a Gaussian.

26. Apparatus as set forth in claim 22, wherein said extracting means comprises:

means for optimally thresholding the data of each picture;

means for locating a plurality of edge points on each edge of the optimally thresholded data;

means for using the edge points to create line segments extending between the edge points; and

means for combining the line segments with one another one at a time to create longer segments so long as each added segment is within a predetermined fitting tolerance.

27. Apparatus as set forth in claim 26, wherein said means for locating comprises means for applying the parameter space method to the data.

28. Apparatus as set forth in claim 27, including means for applying to the data a least squares correction offset.

29. Apparatus as set forth in claim 22, wherein said extracting means comprises:

means for optimally thresholding the data of each picture;

means for applying an edge detection operator to the optimally thresholded data to locate points on the edge; and

means for systematically sampling the points located by application of said operator to provide a reduced resolution determination of the edge point locations and the edge direction at each edge point location.

30. Apparatus as set forth in claim 29, including means for interpolating said reduced resolution determination to provide an interpolated full resolution determination of the edge point locations.

31. Apparatus as set forth in claim 30, including means for correcting the location of each edge point in said full resolution determination for the edge bias of said operator.

32. Apparatus as set forth in claim 29, including means for producing linear segments from the edge points of each picture.

33. Apparatus as set forth in claim 32, wherein said linear segment producing means comprises:

means for linking the edge points from said reduced resolution determination in a manner to provide line segments from the linked edge points;

means for tracing said line segments to provide distinct chains of linked edge points;

means for interpolating said reduced resolution determination to provide an interpolated full resolution determination of the edge point locations; and

means for fitting said chains to line segments defined by the edge point locations of said full resolution determination.

34. Apparatus as set forth in claim 33, wherein said calculating means comprises:

means for determining the length, center coordinate and angular orientation of each fitted chain;

means for accepting as possibly matching pairs all pairs of chains that are within a prescribed angular orientation of one another;

means for constructing at a location determined by the offset between the center coordinates of each possibly matching pair of chains a rectangular painting window having a preselected length and width and an angular orientation dependent upon the lengths and angular orientations of the fitted chains in each possibly matching pair; and

means for assigning a score to each coordinate point in a preselected coordinate frame based on the number of said painting windows the coordinate point falls within, whereby the coordinate point having the highest score can be considered to be said offset.

35. Apparatus as set forth in claim 34, including means for correcting said offset for size differences between the corresponding pictures of the two films.

36. Apparatus as set forth in claim 35, including least squares correction means for correcting said offset for size differences between the corresponding pictures of the two films.

37. In a method of registering a pair of color separation films containing halftone dot detail and macroscopic edges which align when the films are registered, the step of:

selecting one of the films as a reference film;

recording first and second pictures of the reference film centered at spaced apart first and second locations, respectively;

recording third and fourth pictures of the other film centered at locations thereon which approximately correspond to but may be offset from the respective first and second locations;

extracting from said first and third pictures the locations of corresponding macroscopic edges on the reference film and on the other film;

extracting from said second and fourth pictures the locations of corresponding macroscopic edges on the reference film and on the other film;

determining the offsets between the edge locations on the reference film and the corresponding edge locations on the other film required to separately move the first picture into registration with the third picture and the second picture into registration with the fourth picture;

constructing a imaginary line vector extending from a tail point at one selected location on the reference film and a tip point at another selected location on the reference film;

computing a tail point on the other film by subtracting the corresponding offset from the coordinates of said one selected location;

computing a tip point on the other film by subtracting the corresponding offset from the coordinates of said other selected location; and

constructing an imaginary line vector extending from said tail point on the other film to said tip point on the other film whereby the positional adjustment of the other film necessary to register it with the reference film can be effected by aligning the vector on the other film with the vector on the reference film.

38. The method of claim 37, wherein one of said vectors is shorter than the other, and including the step of effecting registration of said other film with said reference film by positioning the other film such that the shorter vector is aligned with and centered on the other vector, thereby equalizing the error in the vector lengths at both ends of the vectors to split the difference in the error.

39. The method of claim 37, wherein the vectors differ in length and including the steps of:

determining the difference in length of the vectors; and

aborting the registration if the difference in length of the vectors exceeds a preselected tolerance.