Granted Patent B2

US 9,237,332 · App. 13/728,065 · Granted Jan 12, 2016

Abstract

Claims (62)

1. A method of reconstructing a high-density three-dimensional (3D) image, the method comprising:

generating an initial 3D image by matching a first image captured using a first camera and a second image captured using a second camera;

searching for a first area and a second area from the initial 3D image by using a number of characteristic points included in the initial 3D image;

detecting a plane from the first area;

filtering the second area by performing non-linear filtering in which a depth is smoothed while preserving luminance steps;

wherein the filtering comprises smoothing depth, corresponding to Z coordinates of the characteristics points, while preserving contrast equal to or greater than a set value with respect to the second area;

and synthesizing the detected plane and the filtered second area.

2. The method of claim 1 , further comprising, after generating the initial 3D image,

dividing one of the first image and the second image into segments comprising the first and second areas; and

applying information about the dividing to the initial 3D image.

3. The method of claim 2 , wherein the searching for the first and second areas comprises searching for at least one first area including characteristic points a number of which is equal to or greater than a set number.

4. The method of claim 3 , wherein the searching for the first and second areas further comprises searching for at least one second area including the characteristic points a number of which is less than the set number.

5. The method of claim 4 , wherein the filtering comprises smoothing depth, corresponding to Z coordinates of the characteristics points, while preserving contrast equal to or greater than a set value with respect to the second area.

6. The method of claim 4 , wherein the generating the initial 3D image comprises:

extracting the characteristic points from the first image and the second image;

matching the characteristic points of the first image and the characteristic points of the second image; and

calculating 3D coordinates of the matched characteristic points.

7. The method of claim 6 , wherein the matching the characteristic points comprises:

setting a first window including a characteristic point in the first image;

searching for a second window of which luminance differs by the least degree from luminance of the first window; and

selecting a matching characteristic point which is the most similar to the characteristic point in the first window among at least one characteristic point in the second window.

8. The method of claim 7 , wherein the calculating the 3D coordinates comprises:

calculating two-dimensional (2D) coordinates of the matched characteristic points with respect to the first and second images; and

calculating 3D coordinates of the matched characteristic points by using a distance between the first and second cameras and angles of the first and second cameras with respect to the characteristic points.

9. The method of claim 6 , wherein the calculating the 3D coordinates comprises:

calculating two-dimensional (2D) coordinates of the matched characteristic points with respect to the first and second images; and

calculating 3D coordinates of the matched characteristic points by using a distance between the first and second cameras and angles of the first and second cameras with respect to the characteristic points.

10. The method of claim 1 , wherein the generating the initial 3D image comprises:

extracting characteristic points from the first image and the second image;

matching the characteristic points of the first image and the characteristic points of the second image;

calculating 3D coordinates of the matched characteristic points; and

generating the initial 3D image using the 3D coordinates of the matched characteristic points.

11. The method of claim 1 , wherein the searching for the first and second areas comprises searching for at least one first area including characteristic points equal to or greater than a set number and at least one second area including the characteristic points less than the set number.

12. An apparatus for reconstructing a high-density three-dimensional (3D) image, comprising at least one processor which implements:

an image matching unit which matches a first image captured using a first camera and a second image captured using a second camera to generate an initial 3D image; and

a 3D image reconstruction unit which divides the initial 3D image into a first area and a second area to detect a plane from the first area, filters the second area by performing non-linear filtering in which a depth is smoothed while preserving luminance steps;

wherein the filtering comprises smoothing depth, corresponding to Z coordinates of the characteristics points, while preserving contrast equal to or greater than a set value with respect to the second area;

and synthesizes the plane detected from the first area and the filtered second area to generate the high-density 3D image.

13. The apparatus of claim 12 , wherein the image matching unit comprises:

a characteristic point extracting unit which extracts characteristic points from the first image and the second image;

a characteristic point matching unit matching which matches the characteristic points of the first image and the characteristic points of the second image;

a 3D coordinate calculation unit which calculates 3D coordinates of the matched characteristic points; and

an initial 3D image generating unit which generates an initial 3D image by using the 3D coordinates of the matched characteristic points.

14. The apparatus of claim 13 , wherein the 3D image reconstruction unit comprises:

a dividing unit which divides one of the first image and the second image, and applying information about the division to the initial 3D image;

a searching unit which searches for the first area and the second area from the divided image by using a number of characteristic points included in the divided image;

a plane detection unit which detects a plane from the first area; and

a filtering unit which filters the second area.

15. The apparatus of claim 14 , wherein the searching unit searches for at least one first area including a set number of the characteristic points, and at least one second area including less than the set number of the characteristic points.

16. The apparatus of claim 14 , wherein the filtering unit filters the second area to smooth depth, corresponding to Z coordinates of the characteristic points, while the depth discontinuities with respect to the division boundaries.

17. The apparatus of claim 12 , wherein the 3D image reconstruction unit comprises:

a dividing unit which divides one of the first image and the second image, and applies information about the division to the initial 3D image;

a searching unit which searches for the first area and the second area from the divided image by using a number of 3D characteristic points included in the divided image;

a plane detection unit which detects a plane from the first area; and

a filtering unit which filters the second area.

18. A non-transitory computer readable recording medium having embodied thereon a program for executing a method of reconstructing a high-density three-dimensional (3D) image, the method comprising:

generating an initial 3D image by matching a first image captured using a first camera and a second image captured using a second camera;

searching for a first area and a second area from the initial 3D image by using a number of characteristic points included in the initial 3D image;

detecting a plane from the first area;

filtering the second area by performing non-linear filtering in which a depth is smoothed while preserving luminance steps;

wherein the filtering comprises smoothing depth, corresponding to Z coordinates of the characteristics points, while preserving contrast equal to or greater than a set value with respect to the second area;

and synthesizing the detected plane and the filtered second area.

Assignments (9)