IMAGE PROCESSING METHOD

• US 20090324062A1
• Filed: 11/06/2008
• Published: 12/31/2009
• Est. Priority Date: 06/25/2008
• Status: Active Grant

• Current Assignee: Samsung Electronics Co. Ltd.
• Original Assignee: Samsung Electronics Co. Ltd.
• Inventors: Kang, Byong Min, LIM, Hwa Sup, Sim, Jae Young
• Granted Patent: US 8,781,256 B2
• Time in Patent Office: Days
• US Class Current: 382/162
• CPC Class Codes: G06T 7/33 using feature-based methods

Abstract

The present invention relates to a method of matching a color image and depth image using feature points corresponding to each other. An image processing method, including: selecting a first pixel group including at least one pixel from an input color image; selecting a second pixel group including at least one pixel corresponding to each of the at least one pixel included in the first pixel group from an input depth image; calculating a transformation function of the first pixel group and second pixel group based on a coordinate value of each of the at least one pixel included in the first pixel group and a coordinate value of each of the at least one pixel included in the second pixel group; and matching each pixel of the input color image with each pixel of the input depth image based on the transformation function. Image processing apparatus, including: a first selection unit selecting a first pixel group including at least one pixel from an input color image; a second selection unit selecting a second pixel group including at least one pixel corresponding to each of the at least one pixel included in the first pixel group from an input depth image; a calculation unit calculating a transformation function of the first pixel group and second pixel group based on a coordinate value of each of the at least one pixel included in the first pixel group and a coordinate value of each of the at least one pixel included in the second pixel group; and an image matching unit matching each pixel of the input color image with each pixel of the input depth image based on the transformation function.

Claims

1. An image processing apparatus, comprising:

a first selection unit selecting a first pixel group including at least one pixel from an input color image;

a second selection unit selecting a second pixel group including at least one pixel corresponding to each of the at least one pixel included in the first pixel group from an input depth image;

a calculation unit calculating a transformation function of the first pixel group and second pixel group based on a coordinate value of each of the at least one pixel included in the first pixel group and a coordinate value of each of the at least one pixel included in the second pixel group; and

an image matching unit matching each pixel of the input color image with each pixel of the input depth image based on the transformation function.

2. The image processing apparatus of claim 1, wherein the first selection unit selects the first pixel group including the at least one pixel by applying a corner detection algorithm to the input color image.

3. The image processing apparatus of claim 2, wherein the corner detection algorithm uses a Harris operator.

4. The image processing apparatus of claim 1, wherein the second selection unit selects at least one pixel of the input depth image by applying a corner detection algorithm to the input depth image, matches each of the at least one pixel of the input depth image with each of the at least one pixel included in the first pixel group, and selects the second pixel group including the at least one pixel of the matched input depth image.

5. The image processing apparatus of claim 1, wherein the transformation function is a transformation matrix obtained by using a Direct Linear Transformation (DLT) algorithm.

6. The image processing apparatus of claim 1, wherein, when a deviation of Z coordinate value of each of the at least one pixel included in the second pixel group is less than a predetermined value, the calculation unit calculates the transformation function excluding information about the Z coordinate value of each of the at least one pixel included in the second pixel group.

7. The image processing apparatus of claim 1, wherein the image matching unit matches each pixel of the input color image with each pixel of the input depth image by applying the transformation function to each of the pixels of the input depth image.

8. The image processing apparatus of claim 1, wherein the image matching unit applies an Interpolation to each pixel of the input depth image, when a resolution of the input color image is not identical to that of the input depth image, each of the pixels of the input depth image being matched with each pixel of the input color image.

9. An image processing apparatus, comprising:

a first selection unit selecting a first pixel group including at least one pixel from an input color image;

a second selection unit selecting a second pixel group including at least one pixel corresponding to each of the at least one pixel included in the first pixel group from an intensity image matched with an input depth image;

a calculation unit calculating a transformation function of the first pixel group and second pixel group based on a coordinate value of each of the at least one pixel included in the first pixel group and a coordinate value of each of the at least one pixel included in the second pixel group; and

an image matching unit matching each pixel of the input color image with each pixel of the input depth image based on the transformation function.

10. The image processing apparatus of claim 9, wherein the second selection unit selects at least one pixel of the intensity image by applying a corner detection algorithm to the intensity image matched with the input depth image, matches each of the at least one pixel of the intensity image with each of the at least one pixel included in the first pixel group, and selects the second pixel group including the at least one pixel of the matched input depth image.

11. The image processing apparatus of claim 9, wherein the transformation function is a transformation matrix obtained by using a DLT algorithm.

12. The image processing apparatus of claim 9, wherein, when a deviation of Z coordinate value of each of the at least one pixel included in the second pixel group is less than a predetermined value, the calculation unit calculates the transformation function excluding information about the Z coordinate value of each of the at least one pixel included in the second pixel group.

13. An image processing apparatus, comprising:

a first selection unit selecting a first pixel group including at least one pixel from an input color image;

a second selection unit receiving information about a second pixel group including at least one pixel corresponding to each of the at least one pixel included in the first pixel group from an intensity image matched with an input depth image;

a calculation unit calculating a transformation function of the first pixel group and second pixel group based on a coordinate value of each of the at least one pixel included in the first pixel group and a coordinate value of each of the at least one pixel included in the second pixel group; and

an image matching unit matching each pixel of the input color image with each pixel of the input depth image based on the transformation function.

14. The image processing apparatus of claim 13, wherein the first selection unit selects the first pixel group including the at least one pixel by applying a corner detection algorithm to the input color image.

15. The image processing apparatus of claim 13, wherein the transformation function is a transformation matrix obtained by using a DLT algorithm.

16. The image processing apparatus of claim 13, wherein the calculation unit calculates the transformation function excluding information about a Z coordinate value of each of the at least one pixel included in the second pixel group.

17. The image processing apparatus of claim 13, wherein the of claim 16, wherein the image matching unit applies an Interpolation to each pixel of the input depth image, when a resolution of the input color image is not identical to that of the input depth image, each of the pixels of the input depth image being matched with each pixel of the input color image.

18. An image processing method, comprising:

selecting a first pixel group including at least one pixel from an input color image;

selecting a second pixel group including at least one pixel corresponding to each of the at least one pixel included in the first pixel group from an input depth image;

calculating a transformation function of the first pixel group and second pixel group based on a coordinate value of each of the at least one pixel included in the first pixel group and a coordinate value of each of the at least one pixel included in the second pixel group; and

matching each pixel of the input color image with each pixel of the input depth image based on the transformation function.

19. The image processing method of claim 18, wherein the selecting of the second pixel group comprises:

selecting at least one pixel of the input depth image by applying a corner detection algorithm to the input depth image;

matching each of the at least one pixel of the input depth image with each of the at least one pixel included in the first pixel group; and

selecting the second pixel group including the at least one pixel of the matched input depth image.

20. A computer-readable recording medium having stored thereon instructions for implementing an image processing method, the instructions comprising:

an instruction set of selecting a first pixel group including at least one pixel from an input color image;

an instruction set of selecting a second pixel group including at least one pixel corresponding to each of the at least one pixel included in the first pixel group from an input depth image;

an instruction set of calculating a transformation function of the first pixel group and second pixel group based on a coordinate value of each of the at least one pixel included in the first pixel group and a coordinate value of each of the at least one pixel included in the second pixel group; and

an instruction set of matching each pixel of the input color image with each pixel of the input depth image based on the transformation function.