Coherent correlation addition for increasing match information in scene matching navigation systems

• US 5,146,228 A
• Filed: 01/24/1990
• Issued: 09/08/1992
• Est. Priority Date: 01/24/1990
• Status: Expired due to Fees

• Current Assignee: Johns Hopkins University
• Original Assignee: Johns Hopkins University
• Inventors: Constantikes, Kim T., Shiflett, Gary D., Irani, Geoffrey B.
• Primary Examiner(s): Tarcza, Thomas H.
• Assistant Examiner(s): Cain, David
• Application Number:
  US07/469,875
• Time in Patent Office: 958 Days
• Field of Search: 382/1, 382/30, 364/456, 358/109, 342/62, 342/63, 342/64, 350/320, 356/2
• US Class Current: 342/64
• CPC Class Codes: G01C 21/005 with correlation of navigation data from several sources, e.g. map or contour matching G01C21/30 takes precedence
• CPC Class Codes: G06T 7/74 involving reference images or patches

Abstract

A method and apparatus are described for image matching, particularly for increasing match information in scene matching navigation systems. The invention correlates a sequence of sensed images with a reference map and stores the resultant correlation surfaces. The correlation surfaces are then added so that match point correlations combine coherently by superposing the individual correlation surfaces using short term accurate navigation data. As a result, a composite correlation surface is formed in which the values at the peak (or antipeak) position add up and the effective match area is greatly increased. Both peaks and antipeaks are then detected to determine a match point and provide a navigation position update.

Claims

1. A method for increasing match area in an image matching system comprising the step of combining at least two correlation surfaces to form a composite correlation surface, wherein each correlation surface is formed by;

comparing a sensed image to a reference image at every location in said reference image at which said sensed image is entirely within said reference image; and

computing a correlation amplitude for each location of said sensed image in said reference image, said correlation amplitudes forming said correlation surface.

2. The method of claim 1, wherein said combining step comprises the step of adding coherently said correlation surfaces.

3. The method of claim 2, wherein said adding step comprises the step of superposing said correlation surfaces using data which relate said correlation surfaces to each other.

4. The method of claim 2, wherein the match point correlation amplitude for each of said correlation surfaces are coherently added.

5. The method of claim 2, wherein the anti-match point correlation amplitude for each of said correlation surfaces are coherently added.

6. A method for matching a sequence of sensed images with a reference map comprising the steps of:

correlating at least two of said sensed images with said reference map to generate a correlation surface for each of said sensed images comprising the steps of;

comparing each of said sensed images to said reference map at every location in said reference map at which said sensed image is entirely within said reference map, and

computing a correlation amplitude for each location of said sensed image in said reference map, said correlation amplitudes forming said correlation surface;

combining said correlation surfaces to form a composite correlation surface; and

determining a match point for said sequence of sensed images in said reference map using said composite correlation surface.

8. The method of claim 6 or 7, wherein said combining step comprises the step of coherently adding said correlation surfaces.

9. The method of claim 8, wherein said adding step comprises the step of superposing said correlation surfaces using data which relate said correlation surfaces to each other.

10. The method of claim 8, wherein the match point correlation amplitude for each of said correlation surfaces are coherently added.

11. The method of claim 8, wherein the anti-match point correlation amplitude for each of said correlation surfaces are coherently added.

12. The method of claims 6 or 7, wherein said determining step comprises the steps of:

searching said composite correlation surface for a maximum correlation amplitude that exceeds an upper threshold value and for a minimum correlation amplitude which is below a lower threshold value; and

using either said maximum or minimum correlation amplitude to determine said match point.

13. The method of claim 12, wherein said searching step further comprises the step of adjusting said upper and lower threshold values based on the standard deviation of actual correlation amplitudes excluding the extremes.

14. The method of claims 6 or 7, further comprising the steps of:

forming a new composite correlation surface when said match point cannot be determined using said composite correlation surface; and

redetermining a match point for said sequence of sensed images in said reference map using said new composite correlation surface.

15. The method of claim 14, further comprising the step of repeating said forming and redetermining steps of claim 14 until said match point is determined or until the last acquired of said sensed images located in said reference map is correlated, whichever occurs first.

16. The method of claim 14, wherein said forming step comprises the steps of:

subtracting from said composite correlation surface said correlation surface generated from the oldest acquired of said sensed images; and

adding to said composite correlation surface said correlation surface generated from the latest acquired of said sensed images to form a new composite correlation surface.

17. The method of claim 12, wherein said using step comprises the steps of:

computing peak-to-sidelobe ratio (PSR) and antipeak-to-sidelobe (APSR) ratio for said composite correlation surface using said maximum and minimum correlation amplitudes; and

using either said PSR or APSR to determine said match point.

18. The method of claims 6 or 7, wherein said correlating step further comprises the step of generating only that portion of said correlation surface for each of said sensed images needed to form said composite correlation surface.

7. An image matching method comprising the steps of:

acquiring a sequence of sensed images with at least two of said sensed images being located within a reference map;

correlating each of said sensed images with said reference map to generate a correlation surface for each of said sensed images comprising the steps of;

comparing each of said sensed images to said reference map at every location in said reference map at which said sensed image is entirely within said reference map, and

computing a correlation amplitude for each location of said sensed image in said reference map, said correlation amplitudes forming said correlation surface;

combining said correlation surfaces to form a composite correlation surface; and

determining a match point for said sequence of sensed images in said reference map using said composite correlation surface.

19. An image matching method comprising the steps of:

acquiring a sequence of sensed images with at least three of said sensed images being located within a larger reference map;

correlating each of said sensed images with said reference map to generate a correlation surface for each of said sensed images comprising the steps of;

comparing each of said sensed images to said reference map at every location in said reference map at which said sensed image is entirely within said reference map, and

computing a correlation amplitude for each location of said sensed image in said reference map, said correlation amplitudes forming said correlation surface;

adding coherently said correlation surfaces to form a composite correlation surface;

searching said composite correlation surface for a maximum correlation amplitude that exceeds an upper threshold value and for a minimum correlation amplitude which is below a lower threshold value wherein said upper and lower threshold values are adjusted based on the standard deviation of actual correlation amplitudes excluding the extremes;

computing peak-to-sidelobe ratio (PSR) and antipeak-to-sidelobe ratio (ASPR) for said composite correlation surface using said maximum and minimum correlation amplitudes;

using said PSR or APSR to determine a match point of said sequence of sensed images in said reference map;

forming a new composite correlation surface when said match point cannot be determined using said composite correlation surface comprising the steps of;

subtracting from said composite correlation surface said correlation surface generated form the oldest acquired of said sensed images; and

adding to said composite correlation surface said correlation surface generated from the latest acquired

of said sensed images to form said new composite correlation surface;

redetermining said match point of said sequence of sensed images in said reference map using said new composite correlation surface; and

repeating said above forming and redetermining steps until said match point with greatest PSR or APSR is determined or until the last acquired of said sensed images located in said reference map is correlated, whichever occurs first.

20. A correlation processor comprising:

means for correlating a sequence of sensed images with a reference map to generate a correlation surface for each of said sense images, said correlating means comprising;

means for comparing each of said sensed images to said reference map at every location in said reference map said correlation amplitudes forming said correlation surface;

means for combining said correlation surfaces to form a composite correlation surface; and

means for determining a match point for said sequence of sensed images in said reference map using said composite correlation surface.

21. The correlation processor as recited in claim 20, wherein said combining means comprises means for adding coherently said correlation surfaces.

22. The correlation processor as recited in claim 21, wherein said adding means comprises means for superposing said correlation surfaces using data which relate said correlation surfaces to each other.

23. The correlation processor as recited in claim 22, wherein said determining means comprises:

means for searching said composite correlation surface for a maximum correlation amplitude which exceeds an upper threshold value and for a minimum correlation amplitude which is below a lower threshold value; and

means for using either said maximum or minimum correlation amplitudes to determine said match point.

24. The correlation processor as recited in claim 23, wherein said determining means further comprises means for adjusting said upper and lower threshold values based on the standard deviation of actual correlation amplitudes excluding the extremes.

25. The correlation processor as recited in claim 24, further comprising:

means for forming a new composite correlation surface when said match point cannot be determined using said composite correlation surface, said forming means comprising;

means for subtracting from said composite correlation surface said correlation surface generated from the oldest acquired of said sensed images; and

means for adding to said composite correlation surface said correlation surface generated from the latest acquired of said sensed images to form said new composite correlation surface;

means for redetermining said match point of said sequence of sensed images in said reference map using said new composite correlation surface; and

means for repeating said above forming and redetermining steps until said match point is determined or until the last acquired of said sensed images located in said reference map is correlated, whichever occurs first.

26. The correlation processor as recited in claim 25, wherein said correlating means further comprises means for generating only that portion of said correlation surface for each of said sensed images needed to form said composite correlation surface.