Any aspect passive volumetric image processing method

• US 7,233,691 B2
• Filed: 12/29/2000
• Issued: 06/19/2007
• Est. Priority Date: 12/29/1999
• Status: Expired due to Term

• Current Assignee: Paul A. Toretta
• Original Assignee: Geospan Corporation
• Inventors: Setterholm, Jeffrey M.
• Primary Examiner(s): Bella; Matthew C.
• Assistant Examiner(s): TUCKER, WESLEY J
• Application Number:
  US10/168,811
• Publication Number:
  US 20020191838A1
• Time in Patent Office: 2,363 Days
• Field of Search: 382/154, 382/293
• US Class Current: 382/154
• CPC Class Codes: G06T 7/593 from stereo images

Abstract

Recorded images having an overlapping scene transformed into a Quantitative Cylispheric Stereo Pair (QCSP) of images. Pixels corresponding to a common scene object point are located in the same or nearly the same horizontal lines in the QCSP of images. The QCSP'"'"'s produced can take a number of different formats, including traditional stereo pairs, radial stereo pairs, and general case cylispheric stereo pairs. The QCSP'"'"'s can be used for various three-dimensional imaging processes, including passive volumetric surveying and the production of three-dimensional visualization models.

Claims

1. A method of processing recorded images, the method comprising:

selecting a pair of recorded images having pixels representing an overlapping scene, each recorded image having associated with it a camera optical center and a camera orientation;

transforming the pair of recorded images into a pair of transformed images in which pixels corresponding to a common scene object point are located in corresponding horizontal lines in the pair of transformed images; and

performing an image correlation using the pair of transformed images;

wherein transforming the pair of recorded images comprises;

establishing a shared spherical coordinate system having an orientation with a polar axis parallel to a line passing through the camera optical centers of the pair of recorded images; and

transforming the pixels of each of the recorded images via the stored spherical coordinate system to produce the transformed images, such that for each transformed image, each row of pixels represents a line of constant longitude angle beta and each column of pixels represents a line of constant latitude angle alpha.

2. The method of claim 1 wherein each recorded image has associated data identifying a spatial position and orientation of a camera which produced the image, and wherein transforming the pair of recorded images is based upon the associated data.

3. The method of claim 2 wherein the transforming uses a geometric transformation based upon the associated data and camera calibration parameters.

4. The method of claim 1 and further comprising:

determining a range to corresponding pixels in the synthetic pair of images identified by the image correlation.

5. The method of claim 4 and further comprising:

determining a global location for each pixel based upon the range.

6. The method of claim 1 and further comprising:

producing a three-dimensional visualization model based upon the image correlation.

7. The method of claim 1 wherein transforming the pixels comprises:

defining a transformation between a global coordinate system and the shared spherical coordinate system;

generating first and second sets of rays in the shared spherical coordinate system which define pixels of the transformed images;

converting the first and second sets of rays to individual camera coordinate systems; and

populating the transformed images with pixels of the recorded images which are intersected by the first and second sets of rays in the individual camera coordinate systems.

8. The method of claim 7 wherein each ray is defined by shared spherical coordinates and by angles alpha and beta, where alpha is an angle of the ray with respect to the polar axis and beta is an angle of a plane defined by the ray and the polar axis with respect to a reference plane through the polar axis.

9. A method of processing image data, the method comprising:

storing first image data and first camera location data;

storing second image data and second camera location data;

geometrically transforming the first image data and the second image data based upon the first camera location data and the second camera location data to produce first transformed image data and second transformed image data; and

performing an image correlation using the first and second transformed image data;

wherein geometrically transforming comprises;

defining a polar axis of a shared spherical coordinate system based upon the first camera location and the second camera location;

defining a rotation transformation matrix based upon the polar axis and the shared spherical coordinate system;

generating first and second sets of rays which define pixels of the first and second transformed image data, respectively;

transforming the first and second sets of rays into a coordinate system related to the first and second image data; and

forming the first and second transformed image data based upon pixels in the first and second image data intersected by the first and second sets of rays.

10. The method of claim 9 wherein the first sphere is centered on the first camera location.

11. The method of claim 10 wherein the second sphere is centered on the second camera location.

12. The method of claim 9 wherein each ray is defined by angles alpha and beta, wherein alpha is an angle of the ray with respect to the polar axis and beta is an angle of a plane defined by the ray and the polar axis with respect to a reference plane through the polar axis.

13. The method of claim 9 and further comprising:

determining a range to corresponding pixels in the first and second transformed image data identified by the image correlation.

14. The method of claim 13 and further comprising:

determining a global location for each pixel based upon the range.

15. The method of claim 9 and further comprising:

producing a three-dimensional visualization model based upon the image correlation.

16. The method of claim 9 wherein geometrically transforming comprises:

defining a cylindrical axis between a first camera location associated with the first image and a second camera location associated with the second image;

locating selected pixels in each recorded image using a coordinate system based upon the cylindrical axis; and

creating the first and second transformed image data geometrically using the selected pixels.

17. A method of processing recorded images, the method comprising:

selecting first and second recorded images having an overlapping scene;

defining a shared spherical coordinate system having a polar axis parallel to a line between first and second camera locations associated with the first and second recorded images;

generating first and second sets of rays which originate from the first and second camera locations, respectively, each ray being defined by angles ∝ and

β

, where ∝

is a latitude angle of the ray with respect to the polar axis and where β

is a longitude angle of a plane defined by the ray and the polar axis with respect to a reference plane through the polar axis;

defining a transformation between a global coordinate system related to the first and second recorded images and the shared spherical coordinate system;

converting the first and second set of rays from the shared spherical coordinate system to individual camera coordinate systems;

populating a first transformed image with pixels of the first recorded image intersected by the first set of rays;

populating a second transformed image with pixels of the second recorded image intersected by the second set of rays; and

wherein for each of the first and second transformed images, each row of pixels represents a line of constant longitude angle β and

each column of pixels represents a line of constant latitude angle ∝

.

18. The method of claim 17, and further comprising:

performing an image correlation using the first and second transformed images.

19. The method of claim 18, and further comprising:

using the image correlation to determine a location of corresponding pixels in the first and second transformed images.

20. The method of claim 17, and further comprising:

producing a three-dimensional model using the first and second transformed images.

21. The method of claim 17, and further comprising;

determining ray direction increments for rays of the first and second sets of rays.

22. The method of claim 21, wherein the ray direction increments are based upon an arctangent of constant distance increments.

23. The method of claim 21, wherein the ray direction increments are based upon a logarithm of the tangent of ∝

.

24. The method of claim 21, wherein the ray direction increments a constant angular increments.