DEPTH IMAGE PROCESSING APPARATUS AND METHOD BASED ON CAMERA POSE CONVERSION

• US 20150178900A1
• Filed: 03/09/2015
• Published: 06/25/2015
• Est. Priority Date: 11/29/2012
• Status: Active Grant

• Current Assignee: Korea Advanced Institute of Science and Technology
• Original Assignee: Korea Advanced Institute of Science and Technology
• Inventors: KIM, Jinwook, HONG, Seungpyo
• Granted Patent: US 9,589,326 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06T 2207/10028 Range image; Depth image; 3D point clouds
• CPC Class Codes: G06T 5/50 using two or more images, e.g. averaging or subtraction
• CPC Class Codes: G06T 5/70 Denoising; Smoothing

Abstract

Disclosed is a depth image processing apparatus based on a camera pose conversion, which includes: a depth image obtaining unit for obtaining depth images; a camera pose converting unit for converting camera poses of the depth images into a camera pose of a reference depth image; and a depth image filtering unit for filtering the reference depth image based on the depth images with converted camera poses.

Claims

1. A depth image processing apparatus, comprising:

a depth image obtaining unit configured to obtain depth images;

a camera pose converting unit configured to convert camera poses of the depth images into a camera pose of a reference depth image; and

a depth image filtering unit configured to filter the reference depth image based on the depth images with converted camera poses.

2. The depth image processing apparatus according to claim 1,

wherein the camera pose converting unit is configured to convert the camera poses of the depth images by obtaining relative camera pose values of the depth images based on the reference depth image.

3. The depth image processing apparatus according to claim 2,

wherein the camera pose converting unit is configured to obtain the relative camera pose values using at least one of a sensor device and an Iterative Closest Point (ICP) algorithm.

4. The depth image processing apparatus according to claim 1,

wherein the depth image filtering unit is configured to filter the reference depth image by using an average value for each pixel of the reference depth image.

5. The depth image processing apparatus according to claim 4,

wherein the depth image filtering unit is configured to calculate the average value by generating at least one filtering window for at least one pixel of the reference depth image and calculating the average value within each filtering window.

6. The depth image processing apparatus according to claim 4,

wherein the average is any one of a weighted average, an arithmetic average, and a geometric average.

7. The depth image processing apparatus according to claim 6,

wherein the average is the weighted average, and

wherein the depth image filtering unit is configured to calculate the weighted average value based on at least one of:

a difference between depth values of each pixel of the reference depth image and each pixel of one of the depth images;

a distance value between each pixel of the reference depth image and each pixel of said one of the depth images; and

a difference between camera pose values of the reference depth image and said one of the depth images.

8. The depth image processing apparatus according to claim 7,

wherein the depth image filtering unit is configured to calculate the difference between depth values by comparing a depth value of a pixel of the reference depth image and depth values of pixels within a filtering window of said one of the depth images.

9. The depth image processing apparatus according to claim 7,

wherein the depth image filtering unit is configured to calculate the distance value based on a distance value between a pixel of the reference depth image and a pixel of said one of the depth images corresponding to the pixel of the reference depth image.

10. The depth image processing apparatus according to claim 7,

wherein the camera pose converting unit is configured to convert the camera poses of the depth images by obtaining relative camera pose values of the depth images based on the reference depth image, and

wherein the depth image filtering unit is configured to calculate the difference between camera pose values by log-mapping the relative camera pose value.

11. A depth image processing method comprising:

obtaining depth images;

converting camera poses of the depth images into a camera pose of a reference depth image; and

filtering the reference depth image based on the depth images with converted camera poses.

12. The depth image processing method according to claim 11,

wherein said converting the camera poses of the depth images comprises obtaining relative camera pose values of the depth images based on the reference depth image.

13. The depth image processing method according to claim 12,

wherein the relative camera pose values are obtained using at least one of a sensor device and an Iterative Closest Point (ICP) algorithm.

14. The depth image processing method according to claim 11,

wherein the reference depth image is filtered using an average value for each pixel of the reference depth image in a step of said filtering the reference depth image.

15. The depth image processing method according to claim 14,

wherein said filtering the depth image comprises:

calculating the average value by generating at least one filtering window for at least one pixel of the reference depth image; and

calculating the average value within each filtering window.

16. The depth image processing method according to claim 14,

wherein the average is any one of a weighted average, an arithmetic average, and a geometric average.

17. The depth image processing method according to claim 16,

wherein the average is the weighted average, and

wherein said calculating the weighted average value is based on at least one of:

a difference between depth values of each pixel of the reference depth image and each pixel of one of the depth images;

a distance value between each pixel of the reference depth image and each pixel of said one of the depth images; and

a difference between camera pose values of the reference depth image and said one of the depth images.

18. The depth image processing method according to claim 17,

wherein said calculating the weighted average value comprises comparing a depth value of a pixel of the reference depth image and depth values of pixels within a filtering window of said one of the depth images to calculate the difference between depth values.

19. The depth image processing method according to claim 17,

wherein said calculating the weighted average value comprises calculating the distance value based on a distance value between a pixel of the reference depth image and a pixel of said one of the depth images corresponding to the pixel of the reference depth image.

20. The depth image processing method according to claim 17,

wherein said converting the camera poses of the depth images comprises obtaining relative camera pose values of the depth images based on the reference depth image, and

wherein said calculating the weighted average value comprises log-mapping the relative camera pose value to calculate the difference between camera pose values.