Context-Based Target Recognition

• US 20130004017A1
• Filed: 06/30/2011
• Published: 01/03/2013
• Est. Priority Date: 06/30/2011
• Status: Active Grant

• Current Assignee: The Boeing Co.
• Original Assignee: The Boeing Co.
• Inventors: Medasani, Swarup S., Korchev, Dmitriy, Molineros, Jose M., Owechko, Yuri
• Granted Patent: US 9,086,484 B2
• Time in Patent Office: Days
• US Class Current: 382/103
• CPC Class Codes: G01S 13/9027 Pattern recognition for feature extraction
• CPC Class Codes: G01S 7/412 based on a comparison between measured values and known or stored values
• CPC Class Codes: G06V 10/44 Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
• CPC Class Codes: G06V 10/46 Descriptors for shape, contour or point-related descriptors, e.g. scale invariant feature transform [SIFT] or bags of words [BoW]; Salient regional features colour feature extraction G06V10/56
• CPC Class Codes: G06V 10/768 using context analysis, e.g. recognition aided by known co-occurring patterns

Abstract

A method and apparatus for identifying a target object. A group of objects is identified in an image. The group of objects provides a context for identifying the target object in the image. The target object is searched for in the image using the context provided by the group of objects.

Claims

1. A method for identifying a target object, the method comprising:

identifying a group of objects in an image, wherein the group of objects provides a context for identifying the target object in the image; and

searching for the target object in the image using the context provided by the group of objects.

2. The method of claim 1, wherein searching for the target object in the image using the context provided by the group of objects comprises:

identifying a number of regions in the image to search for the target object based on the context provided by the group of objects.

3. The method of claim 2, wherein a region in the number of regions is selected from one of within the group of objects and within a selected distance from the group of objects.

4. The method of claim 1, wherein searching for the target object in the image using the context provided by the group of objects comprises:

responsive to identifying a particular object in the image as the target object, determining whether the particular object is the target object based on a location of the particular object to the group of objects.

5. The method of claim 1, wherein searching for the target object in the image using the context provided by the group of objects comprises:

searching for the target object in the image using the context provided by the group of objects with a number of agents, wherein an agent in the number of agents is configured to search for the target object within a window for the agent.

6. The method of claim 1, wherein searching for the target object in the image using the context provided by the group of objects comprises:

searching for the target object in the image in a number of regions in the image using a particle-swarm optimization process, wherein the number of regions is identified using the context provided by the group of objects.

7. The method of claim 1, wherein identifying the group of objects in the image, wherein the group of objects provides the context for identifying the target object in the image comprises:

identifying the group of objects in the image;

forming a number of masks for the group of objects for use in searching for the target object in the image; and

using the number of masks to identify a number of regions in the image to search for the target object.

8. The method of claim 1, wherein an object in the group of objects is selected from at least one of a building, a road, and vegetation.

9. The method of claim 1, wherein identifying the group of objects in the image, wherein the group of objects provides the context for identifying the target object in the image comprises:

identifying a set of lines in the image;

identifying a set of bright areas in the image;

identifying a set of shadows in the image; and

identifying a set of buildings in the image using at least one of the set of lines, the set of bright areas, and the set of shadows.

10. The method of claim 1, wherein identifying the group of objects in the image, wherein the group of objects provides the context for identifying the target object in the image comprises:

identifying texture features for the image;

identifying a texton map using the texture features and a texton codebook;

applying a number of texton layout filters to the image using a number of texton channels formed using the texton map to identify a number of feature vectors; and

identifying a texture mask using the number of feature vectors and a number of boosted classifiers.

11. The method of claim 1, wherein identifying the group of objects in the image, wherein the group of objects provides the context for identifying the target object in the image comprises:

reducing noise in the image to form a processed image;

registering a number of previous images with the processed image to form registered images;

performing a number of merging operations to form a merged image using the registered images;

identifying a set of lines for roads using the merged image, a line detector, a number of morphological operations, and a Hough transform; and

generating a road mask identifying a set of roads using the set of lines identified.

12. The method of claim 1 further comprising:

receiving the image from a synthetic aperture radar imaging system; and

processing the image to reduce noise present in the image prior to identifying the group of objects in the image.

13. An apparatus comprising:

a context recognition module configured to identify a group of objects in an image, wherein the group of objects provides a context for identifying a target object in the image; and

a target recognition module configured to search for the target object in the image using the context provided by the group of objects.

14. The apparatus of claim 13, wherein the context recognition module comprises:

a building detector configured to identify a set of buildings in the image and generate a building mask for use in searching for the target object in the image.

15. The apparatus of claim 13, wherein the context recognition module comprises:

a vegetation detector configured to identify vegetation in the image and generate a vegetation mask for use in searching for the target object in the image.

16. The apparatus of claim 15, wherein the context recognition module comprises:

a road detector configured to identify a set of roads in the image and generate a road mask for use in searching for the target object in the image.

17. The apparatus of claim 13, wherein the context recognition module is configured to identify a number of regions in the image to search for the target object based on the context provided by the group of objects.

18. The apparatus of claim 17, wherein the target recognition module is configured to search for the target object in the number of regions in the image using a particle-swarm optimization process.

19. The apparatus of claim 13, wherein the image is generated using a synthetic aperture radar imaging system.

20. The apparatus of claim 13, wherein the context recognition module comprises:

a noise filter configured to reduce noise present in the image prior to the context recognition module identifying the group of objects in the image.