Distinguishing between moving targets and clutter in a video

• US 9,111,361 B1
• Filed: 02/22/2013
• Issued: 08/18/2015
• Est. Priority Date: 02/22/2013
• Status: Active Grant

• Current Assignee: The Boeing Co.
• Original Assignee: The Boeing Co.
• Inventors: Higgins, Robert P.
• Primary Examiner(s): FUJITA, KATRINA R
• Application Number:
  US13/774,445
• Time in Patent Office: 907 Days
• US Class Current: 1/1
• CPC Class Codes: G06T 2207/10016 Video; Image sequence
• CPC Class Codes: G06T 2207/10032 Satellite or aerial image; Remote sensing
• CPC Class Codes: G06T 2207/30212 Military
• CPC Class Codes: G06T 2207/30232 Surveillance
• CPC Class Codes: G06T 7/254 involving subtraction of images
• CPC Class Codes: G06V 10/143 Sensing or illuminating at different wavelengths
• CPC Class Codes: G06V 10/25 Determination of region of interest [ROI] or a volume of interest [VOI]
• CPC Class Codes: G06V 20/00 Scenes; Scene-specific elements control of digital cameras H04N23/60
• CPC Class Codes: G06V 2201/07 Target detection

Abstract

A method and apparatus for an image processing system is provided. A first image and a second image in a sequence of images are registered to a reference image in the sequence of images to form a first registered image and a second registered image, respectively. A double differenced image is formed using the reference image, the first registered image, and the second registered image. A number of moving target signatures in the double differenced image is identified. A moving target signature in the number of moving target signatures comprises a spot of a first type and at least one spot of a second type adjacent to the spot of the first type.

Claims

1. A method comprising:

registering a first image and a second image in a sequence of images to a reference image in the sequence of images to form a first registered image and a second registered image, respectively;

forming a double differenced image using the reference image, the first registered image, and the second registered image; and

identifying a number of moving target signatures in the double differenced image, wherein a moving target signature in the number of moving target signatures comprises a spot of a first type and at least one spot of a second type adjacent to the spot of the first type;

wherein the spot of the first type is a bright spot;

wherein the at least one spot of the second type is a group of dark spots;

wherein distance between the bright spot and the dark spots indicates a velocity of a moving target associated with the moving target signature; and

wherein the group of dark spots comprises two dark spots located symmetrically around the bright spot.

2. The method of claim 1, wherein registering the first image and the second image in the sequence of images to the reference image in the sequence of images comprises:

aligning a background in the first image with the background in the reference image to form the first registered image; and

aligning the background in the second image with the background in the reference image to form the second registered image.

3. The method of claim 2, wherein forming the double differenced image using the reference image, the first registered image, and the second registered image comprises:

differencing the first registered image and the reference image to form a first differenced image in which the background is substantially removed in the first differenced image; and

differencing the second registered image and the reference image to form a second differenced image in which the background is substantially removed in the second differenced image.

4. The method of claim 3, wherein forming the double differenced image using the reference image, the first registered image, and the second registered image further comprises:

summing the first differenced image and the second differenced image to form the double differenced image.

5. The method of claim 1, wherein identifying the number of moving target signatures in the double differenced image comprises:

detecting a set of candidate signatures in the double differenced image.

6. The method of claim 5, wherein identifying the number of moving target signatures in the double differenced image further comprises:

identifying a set of candidate regions in the double differenced image in which each of the set of candidate regions is a region of the double differenced image containing a corresponding candidate signature in the set of candidate signatures.

7. The method of claim 6, wherein identifying the number of moving target signatures in the double differenced image further comprises:

computing a complex moment vector for a candidate region in the set of candidate regions; and

determining whether a candidate signature corresponding to the candidate region is a moving target signature by comparing the complex moment vector to a set of reference vectors.

8. The method of claim 7 further comprising:

forming the set of reference vectors using a set of reference signatures that includes a reference clutter signature and a plurality of reference target signatures.

9. The method of claim 8, wherein forming the set of reference vectors comprises:

computing a mean and a variance of moment vectors for each of the set of reference signatures to form the set of reference vectors.

10. The method of claim 7,

wherein the complex moment vector is computed based on the following:

11. The method of claim 10,

wherein the complex moment vector is computed using a set of intermediate matrices that are pre-computed:

Mat_x,y^pq=(x+iy)^p(x−

iy)^q

where Mat_x,y^pq is one of the set of intermediate matrices, where x and y are row index and column index offsets from a central pixel location, respectively, where i is the imaginary number, and each of the set of intermediate matrices corresponds to a value in the complex moment vector.

12. The method of claim 1,

wherein the spot of the first type

has at least a first level of brightness, and

comprises one or more contiguous pixels having an intensity value greater than a threshold;

wherein the at least one spot of the second type

has less than a second level of brightness, which is a level of darkness;

wherein positions of the dark spots relative to the bright spot is determined by a velocity of a moving target associated with the moving target signature and a path along which the moving target is moving;

wherein as the velocity of the moving target increases the distance between the bright spot and the dark spots increases; and

wherein an orientation of the dark spots relative to the bright spot indicate a direction in which the moving target is moving.

13. An apparatus comprising:

a processor unit of a computer system, the computer system configured to;

register a first image and a second image in a sequence of images to a reference image in the sequence of images to form a first registered image and a second registered image, respectively, and

form a double differenced image using the reference image, the first registered image, and the second registered image; and

identify a number of moving target signatures in the double differenced image, wherein a moving target signature in the number of moving target signatures comprises a spot of a first type and at least one spot of a second type adjacent to the spot of the first type;

wherein the spot of the first type is a bright spot;

wherein the at least one spot of the second type is a group of dark spots;

wherein distance between the bright spot and the dark spots indicates a velocity of a moving target associated with the moving target signature; and

wherein the group of dark spots comprises two dark spots located symmetrically around the bright spot.

14. The apparatus of claim 13, wherein the computer system is further configured to align a background in the first image with the background in the reference image to form the first registered image and align the background in the second image with the background in the reference image to form the second registered image.

15. The apparatus of claim 14, wherein the computer system is further configured to difference the first registered image and the reference image to form a first differenced image in which the background is substantially removed in the first differenced image and difference the second registered image and the reference image to form a second differenced image in which the background is substantially removed in the second differenced image.

16. The apparatus of claim 15, wherein the computer system is further configured to sum the first differenced image and the second differenced image to form the double differenced image.

17. The apparatus of claim 13, wherein the computer system is configured to detect a set of candidate signatures in the double differenced image.

18. The apparatus of claim 17, wherein a candidate signature is one of a moving target signature and a clutter signature in which the clutter signature comprises one spot of the first type and no spots of the second type adjacent to the one spot of the first type.

19. The apparatus of claim 17, wherein the computer system is further configured to identify a set of candidate regions in the double differenced image in which each of the set of candidate regions is a region of the double differenced image containing a corresponding candidate signature in the set of candidate signatures.

20. The apparatus of claim 19, wherein the computer system is further configured to compute a complex moment vector for a candidate region in the set of candidate regions and determine whether a candidate signature corresponding to the candidate region is a moving target signature by comparing the complex moment vector to a set of reference vectors.

21. The apparatus of claim 20, wherein the computer system is further configured to form the set of reference vectors using a set of reference signatures that includes a reference clutter signature and a plurality of reference target signatures.

22. The apparatus of claim 21, wherein the computer system is configured to compute a mean and a variance of moment vectors for each of the set of reference signatures to form the set of reference vectors.

23. An image processing system of a computer system, the computer system comprising:

a processor unit;

wherein the computer system is configured to;

register a first image and a second image in a sequence of images to a reference image in the sequence of images to form a first registered image and a second registered image, respectively;

form a double differenced image using the reference image, the first registered image, and the second registered image;

form a set of reference vectors using a set of reference signatures in which the set of reference vectors includes a reference clutter signature and a plurality of reference target signatures;

identify a set of candidate regions in the double differenced image in which each of the set of candidate regions contains a candidate signature;

compute a complex moment vector for each of the set of candidate regions; and

determine whether the candidate signature in each of the candidate regions is a moving target signature or a clutter signature using the complex moment vector computed for each of the set of candidate regions and the set of reference vectors;

wherein the moving target signature comprises a spot of a first type and at least one spot of a second type adjacent to the spot of the first type;

wherein the spot of the first type is a bright spot;

wherein the at least one spot of the second type is a group of dark spots;

wherein distance between the bright spot and the dark spots indicates a velocity of a moving target associated with the moving target signature; and

wherein the group of dark spots comprises two dark spots located symmetrically around the bright spot.