Systems and methods for determining boundaries of features in digital images

• US 9,501,838 B1
• Filed: 08/24/2015
• Issued: 11/22/2016
• Est. Priority Date: 08/24/2015
• Status: Active Grant

• Current Assignee: The Boeing Co.
• Original Assignee: The Boeing Co.
• Inventors: Rivera, Mark Anthony, Baker, Anthony W.
• Primary Examiner(s): Patel, Kanjibhai
• Application Number:
  US14/834,095
• Time in Patent Office: 456 Days
• Field of Search: 382/164, 382/190
• US Class Current: 1/1
• CPC Class Codes: G06F 18/2135 based on approximation criteria, e.g. principal component analysis
• CPC Class Codes: G06T 2207/10024 Color image
• CPC Class Codes: G06T 2207/10036 Multispectral image; Hyperspectral image
• CPC Class Codes: G06T 2207/20048 Transform domain processing
• CPC Class Codes: G06T 2207/20092 Interactive image processing based on input by user
• CPC Class Codes: G06T 2207/30184 Infrastructure
• CPC Class Codes: G06T 7/12 Edge-based segmentation
• CPC Class Codes: G06V 10/26 Segmentation of patterns in the image field; Cutting or merging of image elements to establish the pattern region, e.g. clustering-based techniques; Detection of occlusion
• CPC Class Codes: G06V 10/56 relating to colour
• CPC Class Codes: G06V 10/7715 Feature extraction, e.g. by transforming the feature space, e.g. multi-dimensional scaling [MDS]; Mappings, e.g. subspace methods

Abstract

A boundary determination system for determining boundaries of features in a digital image is provided. The boundary determination system includes a processor coupled to a memory. The processor is configured to analyze a spectral content and a spatial structure of a feature in a digital image, characterize the feature in the digital image, and distinguish the feature from immediate surroundings of the feature in the digital image.

Claims

1. A boundary determination system for determining boundaries of features in a digital image, said boundary determination system comprising a processor coupled to a memory, said processor is configured to:

receive a selection of an area in a digital image;

perform a principal component analysis on spectral signatures of pixels in the selected area in the digital image, including determining a centroid representing an average spectral signature and at least one ellipsoid representing a variance in the spectral signatures;

generate a boundary around the selected area based on the principal component analysis;

identify additional pixels outside the boundary to include within the boundary based on determining that additional spectral signatures of the additional pixels fall within the at least one ellipsoid;

identify an inflection point in an amount of pixels covered by dilation of the boundary; and

dilate the boundary to include at least a subset of the additional pixels without exceeding the inflection point.

2. The boundary determination system of claim 1, wherein said processor is further configured to generate multiple ellipsoids.

3. The boundary determination system of claim 1, wherein said processor is further configured to generate at least one spectral distance image.

4. The boundary determination system of claim 3, wherein said processor is further configured to generate at least one spectral gradient magnitude image.

5. The boundary determination system of claim 4, wherein said processor is further configured to:

normalize the at least one spectral distance image and the at least one spectral gradient magnitude image; and

add the normalized spectral distance image and the normalized spectral gradient magnitude image together.

6. The boundary determination system of claim 1, wherein said processor is further configured to:

determine an asymmetry in spectral signatures of pixels adjacent to an edge of the boundary; and

shift the edge in a direction perpendicular to the edge based upon the determined asymmetry.

7. The boundary determination system of claim 6, wherein said processor is further configured to:

determine an asymmetry inflection point in the determined asymmetry; and

shift the edge to not exceed the asymmetry inflection point.

8. The boundary determination system of claim 1, wherein said processor is further configured to:

determine an asymmetry in spectral signatures of pixels adjacent to an edge of the boundary, the edge including at least two vertices; and

adjust a position of at least one of the vertices based upon the determined asymmetry.

9. The boundary determination system of claim 8, wherein said processor is further configured to:

determine an asymmetry inflection point in the determined asymmetry; and

move the at least one of the vertices to not exceed the asymmetry inflection point.

10. A method for determining boundaries of features in a digital image, said method is performed by a boundary determination computing device, said method comprising:

receiving, by the boundary determination computing device, a selection of an area in a digital image;

performing, by the boundary determination computing device, a principal component analysis on spectral signatures of pixels in the selected area in the digital image, including determining a centroid representing an average spectral signature and at least one ellipsoid representing a variance in the spectral signatures;

generating, by the boundary determination computing device, a boundary around the selected area based on the principal component analysis;

identifying, by the boundary determination computing device, additional pixels outside the boundary to include within the boundary based on determining that additional spectral signatures of the additional pixels fall within the at least one ellipsoid;

identifying, by the boundary determination computing device, an inflection point in an amount of pixels covered by dilation of the boundary; and

dilating, by the boundary determination computing device, the boundary to include at least a subset of the additional pixels without exceeding the inflection point.

11. The method of claim 10, further comprising generating multiple ellipsoids.

12. The method of claim 10, further comprising generating at least one spectral distance image.

13. The method of claim 12, further comprising generating at least one spectral gradient magnitude image.

14. The method of claim 13, further comprising:

normalizing the at least one spectral distance image and the at least one spectral gradient magnitude image; and

adding the normalized spectral distance image and the normalized spectral gradient magnitude image together.

15. The method of claim 10, further comprising:

determining an asymmetry in spectral signatures of pixels adjacent to an edge of the boundary; and

shifting the edge in a direction perpendicular to the edge based upon the determined asymmetry.

16. The method of claim 15, further comprising:

determining an asymmetry inflection point in the determined asymmetry; and

shifting the edge to not exceed the asymmetry inflection point.

17. The method of claim 10, further comprising:

determining an asymmetry in spectral signatures of pixels adjacent to an edge of the boundary, the edge including at least two vertices; and

adjusting a position of at least one of the vertices based upon the determined asymmetry.

18. The method of claim 17, further comprising:

determining an asymmetry inflection point in the determined asymmetry; and

moving the at least one of the vertices to not exceed the asymmetry inflection point.

19. A computer-readable medium having computer-executable instructions stored thereon, wherein when executed by a boundary determination computing device including a processor, said computer-executable instructions cause the boundary determination computing device to:

receive a selection of an area in a digital image;

perform a principal component analysis on spectral signatures of pixels in the selected area in the digital image, including determining a centroid representing an average spectral signature and at least one ellipsoid representing a variance in the spectral signatures;

generate a boundary around the selected area based on the principal component analysis;

identify additional pixels outside the boundary to include within the boundary based on determining that additional spectral signatures of the additional pixels fall within the at least one ellipsoid;

identify an inflection point in an amount of pixels covered by dilation of the boundary; and

dilate the boundary to include at least a subset of the additional pixels without exceeding the inflection point.