Imaging system for correction of perceptual distortion in wide angle images

• US 5,999,660 A
• Filed: 07/26/1995
• Issued: 12/07/1999
• Est. Priority Date: 07/26/1995
• Status: Expired due to Term

• Current Assignee: California Institute of Technology
• Original Assignee: California Institute of Technology
• Inventors: Zorin, Denis, Barr, Alan H.
• Primary Examiner(s): COUSO, YON JUNG
• Application Number:
  US08/506,764
• Time in Patent Office: 1,595 Days
• Field of Search: 382/293, 382/285, 382/275, 382/154, 382/276, 395/118, 395/119, 395/125, 395/127
• US Class Current: 382/276
• CPC Class Codes: G06T 3/12 Panospheric to cylindrical image transformations

Abstract

An imaging system which corrects perceptual distortion in wide angle picture balances between direct view distortion and linear perspective distortion. The two kinds of distortion are balanced according to a transformation ratio. The balancing can be expressed as: and φ=.o slashed.

where ρ and φ are the transformed polar coordinates corresponding to the original polar coordinates r and .o slashed., λ is the transformation ratio defining the amount of each transform to be used, and R is the size of the new image. This can also be effected by a lens having the transfer characteristic whereby B=tan β/K; where K is between 1 and 2.

Claims

1. A method of correcting an image, comprising:

obtaining an image in a form where image components can be moved to different locations in the image;

obtaining first and second transformation parameters for the image, said first transformation parameter being a transformation for preserving straightness of lines, and said second transformation parameter being a transformation minimizing distortion of shapes; and

correcting said image by moving its components to different locations, said correcting of at least one of said components being based on both said first transformation parameter and said second transformation parameter.

2. A system for processing images, comprising:

a first element which receives a signal indicative of a content of an image;

a memory, operating to store the content of the image;

an image processor, investigating the image content in said memory to determine kinds of information in the image, and to obtain first and second transformation parameters for the image, said first transformation parameter correcting for a first straightness of lines distortion in the image, and said second transformation parameter correcting for a second distortion of shape distortion in the image and to correct said image by moving its components to different locations based on both said first transformation parameter and said second transformation parameter at the same time, in addition to a transformation ratio that sets a ratio between said first and second transformation parameters.

3. A method of correcting an image, comprising:

obtaining an image in a form where image components can be moved to different locations in the image;

obtaining first and second transformation parameters for the image, said first transformation parameter being a transformation for preserving straightness of lines, and said second transformation parameter being a transformation minimizing distortion of shapes; and

correcting said image by moving its components to different locations, said correcting of at least one of said components being based on both said first transformation parameter and said second transformation parameter, wherein said first transformation corrects as ρ

=r/R, where r is the polar coordinate of an initial image, and R is a dimension of the photographic plate, and said second transformation corrects as ##EQU10## where ρ and

φ

are transformed polar coordinates corresponding to original polar coordinates r and .o slashed., λ

is a transformation ratio defining the amount of each transform to be used, and R is the size of an image; and

said first and second transformation are combined as ##EQU11## .

5. A method as in claim 3 wherein said transformation ratio λ

is a function of coordinates of the image.

4. A method of correcting an image, comprising:

obtaining an image in a form where image components can be moved to different locations in the image;

obtaining first and second transformation parameters for the image, said first transformation parameter being a transformation for preserving straightness of lines, and said second transformation parameter being a transformation minimizing distortion of shapes; and

correcting said image by moving its components to different locations, said correcting of at least one of said components being based on both said first transformation parameter and said second transformation parameter, further comprising;

determining a first part of the image;

determining kinds of distortion in said first part of said image, and a first transformation ratio between said straightness of lines transformation and said distortions of shape transformation which will appropriately correct for said distortion in said first part of said image;

determining a second part of said image;

determining kinds of distortion in said second part of said image, and a second transformation ratio, different than said first transformation ratio, between said straightness of lines transformation and said distortions of shape transformation which will appropriately correct for said distortion in said second part of said image; and

correcting distortion in said first part of said image using said first transformation ratio and correcting distortion in said second part of said image using said second part of said image using said second transformation ratio.

6. A method as in claim 4 wherein said obtaining an image obtains an image on a photographic plate.

7. A method as in claim 4 wherein said obtaining an image comprises obtaining an electronic view of the image.

8. A method of obtaining a distortion-free image, comprising:

obtaining a two-dimensional image of a three dimensional object as projected onto a flat image plane;

first carrying out an operation which has the effect of converting said image from said first flat image plane into an image projected onto a second, curved image plane; and

second carrying out an operation which has the effect of transforming said image from said second image plane to a third, flat image plane, wherein said projections are carried out such that ##EQU12## and φ

=.o slashed.where ρ and

φ

are transformed polar coordinates corresponding to the original polar coordinates r and .o slashed., λ

is a transformation ratio defining the amount of each transform to be used, and R is the size of the image.

10. A method as in claim 8 where λ

is a function of coordinates of the image on the screen.

11. A method as in claim 8 wherein said first and second transformations are carried out by a computer which computes new pixel locations for pixels of the image.

12. A method as in claim 8 wherein said transformations are carried out by a lens.

9. A method of correcting an image, comprising:

acquiring an image that has distortion therein;

determining kinds of distortions which exist in different parts of the image, and classifying said distortions based on whether they will be more easily corrected by a linear transformation or by a direct view transformation;

setting a first transformation ratio λ

1 for a first area of the image;

setting a second transformation ratio μ

2 for a second area of the image;

correcting each said area of the image using a combination of said linear transformation and direct view transformation as weighted by said first and second transformation ratio respectively and a compliment of said transformation ratio respectively, wherein said correcting is carried out such that ##EQU13## and φ

=.o slashed., where ρ and

φ

are the transformed polar coordinates corresponding to the original polar coordinates r and .o slashed., λ

is the transformation ratio defining the amount of each transform to be used with λ

1 being used for said first area, and λ

2 being used for said second area, and R is the size of the new image.

13. A method of correcting an image, comprising:

acquiring an image that has distortion therein;

determining kinds of distortions which exist in different parts of the image, and classifying said distortions based on whether they will be more easily corrected by a linear transformation or by a direct view transformation;

setting a first transformation ratio λ

1 for a first area of the image;

setting a second transformation ratio λ

2 for a second area of the image;

correcting each said area of the image using a combination of said linear transformation and direct view transformation as weighted by said first and second transformation ratio respectively and a compliment of said transformation ratio respectively.

17. A method as in claim 13 wherein said transformation causes some of the areas of the transformed image to be closer together than the original, and causes some areas of the transformed image to be further apart than the original, and further comprising

interpolating pixels to form a new image with constant pixel density.

14. A method of correcting distortion in images, comprising:

obtaining a two dimensional image of a three dimensional object; and

transforming said two dimensional image in a way to reduce distortion therein, using a transformation ratio which is greater than zero, and less than one, to balance linear transformation and direct view transformations and to use both linear transformation and direct view transformations at the same time, wherein said transformation ratio is different for different parts of the image.

15. A method as in claim 14 wherein said transformation ratio is different for different parts of the image.

16. A method as in claim 14 wherein said transformation ratio is a function of position in the image.