Surface extraction method, surface extraction device, and program

• US 8,447,080 B2
• Filed: 03/27/2008
• Issued: 05/21/2013
• Est. Priority Date: 03/28/2007
• Status: Expired due to Fees

• Current Assignee: Sony Corporation (Sony Group Corp.)
• Original Assignee: Sony Corporation (Sony Group Corp.)
• Inventors: Muquit, Mohammad Abdul
• Primary Examiner(s): Akhavannik, Hadi
• Application Number:
  US12/532,479
• Publication Number:
  US 20100054608A1
• Time in Patent Office: 1,881 Days
• Field of Search: 382/124
• US Class Current: 382/128
• CPC Class Codes: G06T 7/55 from multiple images
• CPC Class Codes: G06T 7/596 from three or more stereo images
• CPC Class Codes: G06V 40/107 Static hand or arm
• CPC Class Codes: G06V 40/14 Vascular patterns

Abstract

A surface extraction method capable of improving accuracy of extraction of surfaces of an object, a surface extraction device, and a program are proposed. A voxel space that is divided into voxels having a lattice shape, and whether or not object constituting voxels that constitute an object which is shown in the voxel space is investigated in units of voxel rows of the voxels. Ends of the continuous object constituting voxels in the voxel rows are extracted as surfaces of the object. Straight-line rows are used as units, and intermediate portions of continuous constituting regions in the straight-line rows are ignored. Thus, even when a region that does not constitute the object exists inside the object, the region that does not constitute the object can be prevented from being falsely extracted as a surface of the object without investigation of the same region two times or more.

Claims

1. A surface extraction method comprising:

a first step of investigating, in a three-dimensional space that is divided into regions having a lattice shape, in units of straight-line rows of the regions, whether or not constituting regions that constitute an object which is shown in the three-dimensional space exist; and

a second step of extracting, as surfaces of the object, the constituting regions that exist at ends of a straight-line row and ignoring one or more regions between the end regions,

wherein in the first step,

when a length, width, and height of the three-dimensional space that is occupied by the object are denoted by “

M_X”

, “

M_Y”

, and “

M_Z”

, straight-line rows in a length direction and a width direction on M_Z X-Y planes, M_X Y-Z planes, and M_Y Z-X planes are investigated.

5. The method according to claim 1, further comprising:

forming the three-dimensional space by projecting a plurality of two-dimensional images.

6. The method according to claim 5, wherein the two-dimensional images are based on a first set of two-dimensional images taken from different perspectives of the object.

7. The method according to claim 6, further comprising:

detecting contours of the object in the first set of two-dimensional images;

rotating the first set of two-dimensional images based on the detected contours of the object; and

cutting the rotated first set of two-dimensional images in a predetermined size.

8. The method according to claim 5, wherein the two-dimensional images are binarized.

2. A surface extraction method comprising:

a first step of investigating, in a three-dimensional space that is divided into regions having a lattice shape, in units of straight-line rows of the regions, whether or not constituting regions that constitute an object which is shown in the three-dimensional space exist; and

a second step of extracting, as surfaces of the object, the constituting regions that exist at ends of a straight-line row and ignoring one or more regions between the end regions,

wherein in the second step,

when either of or both of the constituting regions that exist at ends of a straight-line row are in an isolated state, in a case in which distances between the constituting regions in the isolated state and constituting regions that are closest to the constituting regions in the isolated state in the straight-line row are equal to or longer than a predetermined threshold, the constituting regions that are closest to the constituting regions in the isolated state are extracted as surfaces of the object.

11. The method according to claim 2, further comprising:

forming the three-dimensional space by projecting a plurality of two-dimensional images.

12. The method according to claim 11, wherein the two-dimensional images are based on a first set of two-dimensional images taken from different perspectives of the object.

13. The method according to claim 12, further comprising:

detecting contours of the object in the first set of two-dimensional images;

rotating the first set of two-dimensional images based on the detected contours of the object; and

cutting the rotated first set of two-dimensional images in a predetermined size.

14. The method according to claim 11, wherein the two-dimensional images are binarized.

3. A surface extraction device comprising:

a work memory; and

an image processing section that performs image processing with the work memory,

wherein the image processing section performs;

investigating, in a three-dimensional space that is divided into regions having a lattice shape, in units of straight-line rows of the regions, whether or not constituting regions that constitute an object which is shown in the three-dimensional space exist, wherein when a length, width, and height of the three-dimensional space that is occupied by the object are denoted by “

M_X”

, “

M_Y”

, and “

M_Z”

, straight-line rows in a length direction and a width direction on M_Z X-Y planes, M_X Y-Z planes, and M_Y Z-X planes are investigated, and

extracting, as surfaces of the object, the constituting regions that exist at ends of a straight-line row and ignoring one or more regions between the end regions.

15. The surface extraction device according to claim 3, wherein the image processing section further performs forming the three-dimensional space by projecting a plurality of two-dimensional images.

16. The surface extraction device according to claim 15, wherein the two-dimensional images are based on a first set of two-dimensional images taken from different perspectives of the object.

17. The surface extraction device according to claim 16, wherein the image processing section further performs:

detecting contours of the object in the first set of two-dimensional images;

rotating the first set of two-dimensional images based on the detected contours of the object; and

cutting the rotated first set of two-dimensional images in a predetermined size.

18. The surface extraction device according to claim 15, wherein the two-dimensional images are binarized.

4. A non-transitory computer-readable memory storing a program which, when executed by a control section, causes the control section to perform:

investigating, in a three-dimensional space that is divided into regions having a lattice shape, in units of straight-line rows of the regions, whether or not constituting regions that constitute an object which is shown in the three-dimensional space exist, wherein when a length, width, and height of the three-dimensional space that is occupied by the object are denoted by “

M_X”

, “

M_Y”

, and “

M_Z”

, straight-line rows in a length direction and a width direction on M_Z X-Y planes, M_X Y-Z planes, and M_Y Z-X planes are investigated; and

extracting, as surfaces of the object, the constituting regions that exist at ends of a straight-line row and ignoring one or more regions between the end regions.

9. A surface extraction device comprising:

a work memory; and

an image processing section that performs image processing with the work memory,

wherein the image processing section performs;

investigating, in a three-dimensional space that is divided into regions having a lattice shape, in units of straight-line rows of the regions, whether or not constituting regions that constitute an object which is shown in the three-dimensional space exist, and

extracting, as surfaces of the object, the constituting regions that exist at ends of a straight-line row and ignoring one or more regions between the end regions, wherein when either of or both of the constituting regions that exist at ends of a straight-line row are in an isolated state, in a case in which distances between the constituting regions in the isolated state and constituting regions that are closest to the constituting regions in the isolated state in the straight-line row are equal to or longer than a predetermined threshold, the constituting regions that are closest to the constituting regions in the isolated state are extracted as surfaces of the object.

19. The surface extraction device according to claim 9, wherein the image processing section further performs forming the three-dimensional space by projecting a plurality of two-dimensional images.

20. The surface extraction device according to claim 19, wherein the two-dimensional images are based on a first set of two-dimensional images taken from different perspectives of the object.

21. The surface extraction device according to claim 20, wherein the image processing section further performs:

detecting contours of the object in the first set of two-dimensional images;

rotating the first set of two-dimensional images based on the detected contours of the object; and

cutting the rotated first set of two-dimensional images in a predetermined size.

22. The surface extraction device according to claim 19, wherein the two-dimensional images are binarized.

10. A non-transitory computer-readable memory storing a program which, when executed by a control section, causes the control section to perform:

investigating, in a three-dimensional space that is divided into regions having a lattice shape, in units of straight-line rows of the regions, whether or not constituting regions that constitute an object which is shown in the three-dimensional space exist; and

extracting, as surfaces of the object, the constituting regions that exist at ends of a straight-line row and ignoring one or more regions between the end regions, wherein when either of or both of the constituting regions that exist at ends of a straight-line row are in an isolated state, in a case in which distances between the constituting regions in the isolated state and constituting regions that are closest to the constituting regions in the isolated state in the straight-line row are equal to or longer than a predetermined threshold, the constituting regions that are closest to the constituting regions in the isolated state are extracted as surfaces of the object.