Vehicle track detection in synthetic aperture radar imagery

• US 10,467,474 B1
• Filed: 07/11/2016
• Issued: 11/05/2019
• Est. Priority Date: 07/11/2016
• Status: Active Grant

• Current Assignee: National Technology & Engineering Solutions Of Sandia LLC (Honeywell International Inc.)
• Original Assignee: National Technology & Engineering Solutions Of Sandia LLC (Honeywell International Inc.)
• Inventors: Malinas, Rebecca, Quach, Tu-Thach, Koch, Mark W.
• Primary Examiner(s): Windrich, Marcus E
• Application Number:
  US15/207,290
• Time in Patent Office: 1,212 Days
• Field of Search: 342 25 F
• CPC Class Codes: G01S 13/9023 combined with interferometric techniques
• CPC Class Codes: G01S 13/904 SAR modes
• CPC Class Codes: G06F 18/2113 by ranking or filtering the set of features, e.g. using a measure of variance or of feature cross-correlation
• CPC Class Codes: G06V 10/457 by analysing connectivity, e.g. edge linking, connected component analysis or slices
• CPC Class Codes: G06V 10/48 by mapping characteristic values of the pattern into a parameter space, e.g. Hough transformation
• CPC Class Codes: G06V 20/182 Network patterns, e.g. roads or rivers

Abstract

The various technologies presented herein relate to detecting one or more vehicle tracks in radar imagery. A CCD image can be generated from a first SAR image and a second SAR image captured for a common scene, wherein the second SAR image may include a vehicle track that is not present in the first SAR image. A Radon transform (RT) process can be applied to each pixel in the CCD image, and further, a radial derivative (RDRT) can be determined for each pixel from RT values derived for each pixel. Each pixel can be labelled as being related to a track, or not, based upon a unary cost obtained from the RDRT value of that pixel, combined with a probability of the pixel label based upon labels applied to neighboring pixels. A labelled representation of the CCD image can be generated based upon the determination of “track” or “not track”.

Claims

1. A computing system comprising:

at least one processor; and

memory that comprises instructions that, when executed by the at least one processor, cause the at least one processor to perform acts comprising;

determining a radial derivative (RDRT) value for each pixel in a plurality of pixels included in a Radon transform (RT) representation, wherein the RT representation is formed from a coherent change detection (CCD) image, and wherein respective pixel values included in the RT representation are based upon magnitude values for corresponding pixels in the CCD image;

evaluating a first unary cost of labeling a first pixel in the plurality of pixels as being representative of a vehicle track, wherein the first unary cost is based upon a first RDRT value associated with the first pixel;

evaluating a second unary cost of labeling the first pixel as being not representative of a vehicle track, wherein the second unary cost of labeling the first pixel is based upon the first RDRT value associated with the first pixel;

labeling the first pixel in the plurality of pixels as being representative of the vehicle track based upon the first unary cost and the second unary cost; and

outputting an indication that a vehicle track exists in the CCD image, based upon the first pixel in the RT representation being labeled as being representative of a vehicle track.

2. The computing system of claim 1, wherein the labeling of the first pixel in the plurality of pixels is based further upon a pairwise cost, wherein the pairwise cost of labeling the first pixel is based upon a label of an adjacent pixel, wherein the adjacent pixel is labelled as one of being representative of a vehicle track or not being representative of a vehicle track.

3. The computing system of claim 2, the acts further comprising determining, for each pixel in the plurality of pixels, the unary cost of labeling the pixel as being representative of a vehicle track based upon:

applying a window around each respective pixel, wherein the window has a size of I-pixels by h-pixels, wherein l and h are integers;

determining a RDRT value across the window; and

applying the determined RDRT value to the respective pixel around which the window is centered.

4. The computing system of claim 3, wherein the determined RDRT value is the maximum RDRT value determined for all the pixels included in the window.

5. The computing system of claim 3, wherein the window size is 32 pixels by 32 pixels.

6. The computing system of claim 2, the acts further comprising determining a pairwise cost for each pixel in the plurality of pixels by:

determining one or more pairwise values for a pixel based upon a label applied to each pixel located adjacent to the respective pixel; and

determining the pairwise cost for the respective pixel based upon the identified one or more pairwise values.

7. The computing system of claim 6, wherein the adjacent pixels comprise of at least one of a second pixel located adjacent and to the right of the respective pixel, a third pixel located adjacent and to the left of the respective pixel, a fourth pixel located adjacent and below the respective pixel, or a fifth pixel located adjacent and above the respective pixel.

8. The computing system of claim 1, further comprising a display device, wherein outputting the indication that the vehicle track exists in the CCD image comprises displaying the CCD image with the vehicle track indicated thereon on the display device.

18. The computing system of claim 1, wherein evaluating the first unary cost is based upon a probability of the first pixel being associated with the first RDRT value given the first pixel being labeled as representative of the vehicle track.

9. A computer-executable method for identifying a vehicle track in a radar image, comprising:

determining a radial derivative (RDRT) value for each pixel in a plurality of pixels included in a Radon transform (RT) representation, wherein the RT representation is formed from a coherent change detection (CCD) image, and wherein respective pixel values included in the RT representation are based upon magnitude values for corresponding pixels in the CCD image;

labeling a first pixel in the plurality of pixels as being representative of a vehicle track based upon a first unary cost of labeling the first pixel as being representative of the vehicle track and a second unary cost of labeling the first pixel as not being representative of the vehicle track, wherein the first unary cost and the second unary cost are based upon a first RDRT value associated with the first pixel; and

outputting an indication that a vehicle track exists in the CCD image, based upon the first pixel in the RT representation being labeled as being representative of a vehicle track.

10. The method of claim 9, wherein the labeling of the first pixel in the plurality of pixels is based further upon a pairwise cost of labeling the first pixel as being representative of the vehicle track, the pairwise cost based upon a label assigned to a second pixel that is adjacent to the first pixel, wherein the label assigned to the second pixel is indicative of whether the second pixel is representative of a vehicle track.

11. The method of claim 9, wherein the first RDRT value is the maximum RDRT value determined for all the pixels included in a window of pixels centered around the first pixel.

12. The method of claim 11, wherein a size of the window of pixels is 32 pixels by 32 pixels.

13. The method of claim 10, wherein the pairwise cost of labeling the first pixel as being representative of the vehicle track is based upon labels assigned to a plurality of pixels that are adjacent to the first pixel, the second pixel included in the plurality of pixels that are adjacent to the first pixel.

14. The method of claim 13, wherein the plurality of adjacent pixels further comprise at least one of a third pixel located adjacent and to the right of the first pixel, a third pixel located adjacent and to the left of the first pixel, a fourth pixel located adjacent and below the first pixel, or a fifth pixel located adjacent and above the first pixel.

15. The method of claim 9, wherein outputting the indication that the vehicle track exists in the CCD image comprises displaying the CCD image with the track indicated thereon.

16. A computer-readable storage medium comprising instructions that, when executed by a processor, cause the processor to perform acts comprising:

determining a radial derivative (RDRT) value for each pixel in a plurality of pixels included in a Radon transform (RT) image, wherein the RT image is formed from a coherent change detection (CCD) image depicting a scene;

labeling a first pixel in the plurality of pixels as being representative of a vehicle track, the labelling is based upon a combination of a unary cost of labeling the first pixel based upon the RDRT for the first pixel and a pairwise cost of labeling the first pixel based upon a label of a pixel that is adjacent to the first pixel, wherein the pixel that is adjacent to the first pixel is labelled as one of being representative of a vehicle track or not being representative of a vehicle track; and

outputting an indication that a vehicle track exists in the CCD image based upon the first pixel being labeled as being representative of a vehicle track.

17. The computer-readable storage medium of claim 16, wherein outputting the indication comprises displaying the CCD image with the first pixel identified thereon.