Method and apparatus for object positioning by using depth images

• US 9,117,138 B2
• Filed: 12/24/2012
• Issued: 08/25/2015
• Est. Priority Date: 09/05/2012
• Status: Active Grant

• Current Assignee: Industrial Technology Research Institute
• Original Assignee: Industrial Technology Research Institute
• Inventors: Li, Chia-Chang, Liao, Duan-Li
• Primary Examiner(s): AHMED, SAMIR ANWAR
• Application Number:
  US13/726,454
• Publication Number:
  US 20140064602A1
• Time in Patent Office: 974 Days
• Field of Search: 382/154
• US Class Current: 1/1
• CPC Class Codes: G06T 2207/10028 Range image; Depth image; 3D point clouds
• CPC Class Codes: G06T 2207/30196 Human being; Person
• CPC Class Codes: G06T 7/73 using feature-based methods
• CPC Class Codes: G06V 40/107 Static hand or arm

Abstract

According to an exemplary embodiment, a method for object positioning by using depth images is executed by a hardware processor as following: converting depth information of each of a plurality of pixels in each of one or more depth images into a real world coordinate; based on the real world coordinate, computing a distance of each pixel to an edge in each of a plurality of directions; assigning a weight to the distance of each pixel to each edge; and based on the weight of the distance of each pixel to each edge and a weight limit, selecting one or more extremity positions of an object.

Claims

1. A method for object positioning by using depth images, adapted to an apparatus for positioning an object, and the method being executed by a hardware processor to perform:

converting a depth information of each of a plurality of pixels in each of one or more depth images into a real world coordinate;

according to the real world coordinate, computing a distance of each of the plurality of pixels to a plurality of edges of the object in each of a plurality of directions around each of the plurality of pixels;

assigning a weight to the distance of the each pixel to each edge of a plurality of edges; and

according to a weight limit and the weight of the distance of each of the plurality of pixels to each of the plurality of edges, selecting within a specified area range at least one pixel having a largest weight to determine one or more extremity positions of the object.

2. The method as claimed in claim 1, wherein the depth information of the pixel represents a parallax between a left image and a right image, and said parallax between the right image and the left image is a pixel distance between a first image and a second image captured by dual video cameras.

3. The method as claimed in claim 1, wherein computing the distance of the pixel to the edge in each of the plurality of directions further comprises:

computing by searching along a straight direction starting from the pixel in the depth image, until there is a depth information difference between each of at least one pixel and the pixel exceeds a predetermined threshold; and

based on the real-world coordinate, computing at least one distance between the pixel and the at least one pixel.

4. The method as claimed in claim 1, wherein the method assigns different weights to different distances of the pixel to different edges in different directions.

5. The method as claimed in claim 1, wherein the method assigns a largest weight to a distance of the pixel to the edge when the distance is at a specified real distance, and assigns a smaller weight to the distance when a difference between the distance and the specified real distance is larger.

6. The method as claimed in claim 1, wherein said selecting the one or more extremity positions of said object further includes:

checking each of the plurality of pixels in a weighted image after smoothing, and determining pixels with a weight exceeding said weight limit as a candidate extremity position of said object.

7. A method for object positioning by using depth images, adapted to an apparatus for positioning an object, and the method being executed by a hardware processor to perform:

for each of a plurality of pixels in each of one or more depth images, computing a distance of the pixel to a plurality of edges of the object in each of a plurality of directions around each of the plurality of pixels;

converting a depth information of the distance of the each pixel to each edge of a plurality of edges into a real world coordinate;

assigning a weight to the distance of the each pixel to the each edge; and

according to a weight limit and the weight of the distance of each of the plurality of pixels to each of the plurality of edges, selecting within a specified area range at least one pixel having a largest weight to determine one or more extremity positions of the object.

8. The method as claimed in claim 7, wherein the method assigns different weights to different distances of the pixel to different edges in different directions.

9. The method as claimed in claim 7, wherein the method assigns a largest weight to a distance of the pixel to the edge when the distance is at a specified real distance, and assigns a smaller weight to the distance when a difference between the distance and the specified real distance is larger.

10. The method as claimed in claim 7, wherein the method locates at least one pixel having a depth difference greater than a predetermined threshold in each of the plurality of directions around each of the plurality of pixels to determine said edge.

11. An apparatus for positioning an object by using depth images, comprising:

a hardware processor connected to a depth image capture device, wherein said hardware processor further includes;

a real-world coordinate computing module that converts each of a plurality of received depth information into a real world coordinate, and

a computing device configured to compute a distance of each of a plurality of pixels in each of one or more depth images captured by the depth image capture device to a plurality of edges of the object in each of a plurality of directions around each of the plurality of pixels, assign a weight to the distance of each pixel to each edge, and select within a specified area range at least one pixel having a largest weight to determine one or more extremity positions of the object according to a weight limit and the weight of the distance of each of the plurality of pixels to each edge in each of the plurality of directions.

12. The apparatus as claimed in claim 11, wherein each of the plurality of received depth information comes from the one or more depth images captured by said depth image capture device, or is the distance of said each pixel to each edge calculated by said computing device.

13. The apparatus as claimed in claim 11, wherein said depth image capture device is a depth sensor, or an image capture device having dual video camera architecture.

14. The apparatus as claimed in claim 13, wherein the pixel information of each of the plurality of depth images is a real distance with respect to said depth sensor.

15. The apparatus as claimed in claim 13, wherein the pixel information of each of the plurality of depth images is a pixel distance of a first image and a second image captured by dual video cameras.

16. The apparatus as claimed in claim 11, wherein said hardware processor is one of a hardware circuit having a computing ability and a computer system quipped with at least one central processing unit and at least one memory.

17. The apparatus as claimed in claim 11, wherein said apparatus further includes said depth image capture device to capture the one or more depth images.

18. The apparatus as claimed in claim 11, wherein said apparatus locates at least one pixel having a depth difference greater than a predetermined threshold in each of the plurality of directions around each pixel to determine said edge.