Image change detection systems, methods, and articles of manufacture

• US 20050047678A1
• Filed: 09/03/2003
• Published: 03/03/2005
• Est. Priority Date: 09/03/2003
• Status: Active Grant

• Current Assignee: Battelle Energy Alliance LLC (Battelle Memorial Institute)
• Original Assignee: Battelle Energy Alliance LLC (Battelle Memorial Institute)
• Inventors: Jones, James L., Lancaster, Gregory D., Lassahn, Gordon D.
• Granted Patent: US 7,643,703 B2
• Time in Patent Office: Days
• US Class Current: 382/294
• CPC Class Codes: G06T 7/246 using feature-based methods, e.g. the tracking of corners or segments
• CPC Class Codes: G06T 7/33 using feature-based methods

Abstract

Aspects of the invention relate to image change detection systems, methods, and articles of manufacture. According to one aspect, a method of identifying differences between a plurality of images is described. The method includes loading a source image and a target image into memory of a computer, constructing source and target edge images from the source and target images to enable processing of multiband images, displaying the source and target images on a display device of the computer, aligning the source and target edge images, switching displaying of the source image and the target image on the display device, to enable identification of differences between the source image and the target image.

Claims

1. A method of identifying differences between a pair of images, comprising:

loading a source image and a target image into memory of a computer;

constructing source and target edge images from the source and target images to enable processing of multiband images;

displaying the source and target images on a display device of the computer;

aligning the source and target edge images; and

switching displaying of the source image and the target image on the display device to enable identification of differences between the source image and the target image.

2. The method of claim 1, wherein the aligning comprises automatically aligning the source and target edge images.

3. The method of claim 2, wherein the aligning is automatically performed by the computer.

4. The method of claim 2, wherein the automatically aligning the source and target edge images comprises:

a) computing figure of merit values for the source and target edge images;

b) computing separation and ratio values for the source and target edge images;

c) constructing small image versions of the source and target edge images;

d) computing remap coefficients for the small image versions of source and target edge images;

e) remapping the small image versions using the remap coefficients;

f) determining intensity peaks in the small image versions of the source and target edge images; and

g) matching intensity peaks in the source and target edge images.

5. The method of claim 4, further comprising:

h) performing non-linear squares fitting;

i) computing new remap coefficients of the source and target edge images; and

j) computing new figure of merit values using the new remap coefficients;

k) comparing the new figure of merit values with the figure of merit values computed in step a); and

l) iterating steps b)-k) until either the separation or figure of merit values are within predetermined limits.

6. The method of claim 4 wherein the separation comprises smoothing width for the source and target edge images, and further wherein the smoothing width is inversely proportional to a degree of smoothing and local peaks present in the source and target edge images.

7. The method of claim 1, wherein the aligning comprises manually aligning the source and target images.

8. The method of claim 7, wherein the manually aligning the source and target edge images comprises:

enabling a user selecting a tie-point in the target image and a tie-point in the source image to perform alignment of the source and target images; and

computing remap coefficients using the selected tie-points in the source and target images.

9. The method of claim 1, wherein the constructing the source and target edge images comprises:

constructing an edge image for individual color components of the source image;

constructing an edge image for individual color components of the target image;

averaging the edge images of individual color components of the source image to create a single source edge image corresponding to the source image; and

averaging the edge images of individual color components of the target image to create a single target edge image corresponding to the target image.

10. The method of claim 9, wherein the constructing the source edge image comprises:

for each interior pixel of the source edge image, examining a corresponding pixel of the source image and four neighboring pixels of the corresponding pixel of the source image;

determining largest and smallest intensity values of the corresponding pixel and the four neighboring pixels;

establishing a difference between the largest and the smallest intensity values of the corresponding pixel and the four neighboring pixels as intensity value of the interior pixel in the source edge image; and

assigning a value of 0 to pixels at edges of the source edge image.

11. The method of claim 9, wherein the constructing the target edge image comprises:

for each interior pixel of the target edge image, examining a corresponding pixel of the target image and four neighboring pixels of the corresponding pixel of the target image;

determining largest and smallest intensity values of the corresponding pixel and the four neighboring pixels;

establishing a difference between the largest and the smallest intensity values of the corresponding pixel and the four neighboring pixels as intensity value of the interior pixel in the target edge image; and

assigning a value of 0 to pixels at edges of the target edge image.

12. The method of claim 1, wherein the source and target images each comprise color images having a plurality of color bands.

13. The method of claim 1, wherein performing alignment using the target and source edge images enables alignment of source and target images which are not of similar types.

14. The method of claim 1, wherein performing alignment using the target and source edge images enables merging of information from several bands of a multi-band image into a single-band image to obtain efficient processing of images.

15. A method of identifying differences between a pair of images, comprising:

receiving source and target images in a computer;

constructing source and target edge images from the respective source and target images to enable processing of multiband images;

performing alignment of the source and target edge images displayed on a display device of the computer; and

periodically switching displaying of the source and target images on the display device.

16. The method of claim 15, wherein the periodically switching further comprises:

obtaining remap coefficients using the source and target edge images; and

computing a remapped source edge image using the remap coefficients to correct alignment errors between the source and target edge images.

17. The method of claim 15, wherein the performing the alignment is automatically performed by the computer.

18. The method of claim 17, wherein performing the alignment by the computer comprises:

selecting control points in the source and target edge images by a user; and

providing the selected control points to the computer for alignment of the source and target edge images by the computer.

19. The method of claim 17, wherein performing the alignment by the computer comprises:

selecting a region of interest in the source and target images by a user; and

providing the selected region of interest to the computer for alignment of the source and target edge images by the computer.

20. The method of claim 15, wherein the alignment is manually performed by a user.

21. The method of claim 20, wherein performing the alignment manually comprises:

selecting control points in the source and target images; and

aligning the source and target images using the selected control points.

22. The method of claim 20, wherein performing the alignment manually comprises:

selecting a region of interest in the source and target images; and

aligning the source and target images using the selected region of interest.

23. The method of claim 15, wherein the periodically switching the displaying of the source and target images is performed automatically performed by the computer.

24. The method of claim 15, wherein the periodically switching the displaying of the source and target images is performed by a user.

25. The method of claim 15, wherein the performing the alignment of the source and target edge images comprises:

identifying intensity peaks in the source and target edge images; and

using a least squares computation to match corresponding intensity peaks in the source and target edge images.

26. The method of claim 15, wherein the constructing the source edge image comprises:

for each interior pixel of the source edge image, examining a corresponding pixel of the source image and four neighboring pixels of the corresponding pixel of the source image;

determining largest and smallest intensity values of the corresponding pixel and the four neighboring pixels;

establishing a difference between the largest and the smallest intensity values of the corresponding pixel and the four neighboring pixels as intensity value of the interior pixel in the source edge image; and

assigning a value of 0 to pixels at edges of the source edge image.

27. The method of claim 15, wherein the constructing the target edge image comprises:

for each interior pixel of the target edge image, examining a corresponding pixel of the target image and four neighboring pixels of the corresponding pixel of the target image;

determining largest and smallest intensity values of the corresponding pixel and the four neighboring pixels;

establishing a difference between the largest and the smallest intensity values of the corresponding pixel and the four neighboring pixels as intensity value of the interior pixel in the target edge image; and

assigning a value of 0 to pixels at edges of the target edge image.

28. An image change detection system, comprising:

a storage device configured to receive and store a pair of images, the pair of images files including a source image and a target image; and

processing circuitry configured to receive the source and target images, compute source and target edge images from the respective source and target images to enable processing of multiband images, execute alignment of the source and target edge images, and execute periodic switching of the source and target images to enable a user to determine changes between the source and target images.

29. The image change detection system of claim 28, wherein the source and target images are automatically aligned by the processing circuitry.

30. The image change detection system of claim 28, wherein the source and target images are manually aligned by a user of the image change detection system.

31. An article of manufacture comprising:

processor-usable media embodying programming configured to cause a processing circuitry of an image change detection system to;

load a source image and a target image into memory of a computer;

construct source and target edge images from the source and target images to enable processing of multiband images;

display the source and target images on a display device of the computer;

align the source and target edge images; and

switch displaying of the source image and the target image on the display device, to enable identification of differences between the source image and the target image.

32. The article of manufacture of claim 31, wherein the processing circuitry is configured to:

a) first compute figure of merit values for the source and target edge images;

b) compute separation and ratio values for the source and target edge images;

c) construct small image versions of the source and target edge images;

d) compute remap coefficients for the small image versions of the source and target edge images;

e) remap the small image versions using the remap coefficients;

f) determine intensity peaks in the small image versions of the source and target edge images; and

g) match intensity peaks in the source and target edge images.

33. The article of manufacture of claim 32, further comprising:

h) perform non-linear squares fitting;

i) compute new remap coefficients of the source and target edge images; and

j) compute new figure of merit values using the new remap coefficients;

k) compare the new figure of merit values with the first computed figure of merit values; and

l) iterate steps b)-k) until either the separation or figure of merit values are within predetermined limits.

34. An image change detection system, comprising:

means for loading a source image and a target image into memory of a computer means;

means for constructing source and target edge images from the source and target images to enable processing of multiband images;

means for displaying the source and target images on a display means of the computer means;

means for aligning the source and target edge images; and

means for switching displaying of the source image and the target image on the display means, to enable identification of differences between the source image and the target image.

35. The image change detection system of claim 34, wherein the means for aligning the source and target edge images comprises:

means for computing figure of merit values for the source and target edge images;

means for computing separation and ratio values for the source and target edge images;

means for constructing small image versions of the source and target edge images;

means for computing remap coefficients for the small image versions of the source and target edge images;

means for remapping the small image versions using the remap coefficients;

means for determining intensity peaks in the small image versions of the source and target edge images; and

means for matching intensity peaks in the source and target edge images;

36. The image change detection system of claim 35, further comprises:

means for performing non-linear squares fitting; and

means for computing new remap coefficients of the source and target edge images.