Method and apparatus for labeling point cloud data

• US 10,460,172 B2
• Filed: 01/19/2017
• Issued: 10/29/2019
• Est. Priority Date: 09/14/2016
• Status: Active Grant

• Current Assignee: Beijing Baidu Netcom Science and Technology Company Limited (Baidu Incorporated)
• Original Assignee: Beijing Baidu Netcom Science and Technology Company Limited (Baidu Incorporated)
• Inventors: Feng, Kaiwen, Chen, Zhuo, Liu, Bocong, Deng, Chengliang, Liang, Yibing, Ma, Yu
• Primary Examiner(s): Misleh, Justin P.
• Application Number:
  US15/409,667
• Publication Number:
  US 20180075666A1
• Time in Patent Office: 1,013 Days
• Field of Search: None
• CPC Class Codes: G06F 18/00 Pattern recognition
• CPC Class Codes: G06T 19/00 Manipulating 3D models or images for computer graphics
• CPC Class Codes: G06T 2207/10028 Range image; Depth image; 3D point clouds
• CPC Class Codes: G06T 2219/004 Annotating, labelling
• CPC Class Codes: G06V 20/20 in augmented reality scenes
• CPC Class Codes: G06V 20/40 in video content extracting overlay text G06V20/62; video retrieval G06F16/70; processing of video elementary streams in video servers H04N21/234; processing of video elementary streams in video clients H04N21/44
• CPC Class Codes: G06V 20/58 Recognition of moving objects or obstacles, e.g. vehicles or pedestrians; Recognition of traffic objects, e.g. traffic signs, traffic lights or roads
• CPC Class Codes: G06V 2201/12 Acquisition of 3D measurements of objects

Abstract

The present application discloses a method and an apparatus for processing point cloud data. The method of an embodiment comprises: recognizing an object in a to-be-labeled point cloud frame by using an obstacle recognition algorithm, to obtain a recognition result; presenting the recognition result as an initial labeling result of the point cloud frame; and updating the labeling result in response to a correction operation by a user on the labeling result. According to the embodiment, the speed and accuracy of point cloud data labeling are improved.

Claims

1. A method for processing a point cloud frame of point cloud data, comprising:

recognizing an object in the point cloud frame by using an obstacle recognition algorithm;

presenting an initial labeling result of the point cloud frame on a user interface of an electronic device, wherein the initial labeling result comprises the recognized object and a three-dimensional border added to the recognized object; and

updating the initial labeling result in response to a correction operation by a user on the initial labeling result, wherein the correction operation comprises;

a border adjusting operation, the border adjusting operation including a translation operation, zooming operation, and rotation operation on only the three-dimensional border to fit the three-dimensional border to the recognized object with the recognized object remaining unadjusted by the border adjusting operation.

2. The method according to claim 1, wherein the correction operation further comprises:

a checking operation by the user on an object type information and updating the initial labeling result according to the checked object type information.

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

training the obstacle recognition algorithm by using the updated initial labeling result as sample data, to update the obstacle recognition algorithm.

4. The method according to claim 3, further comprising:

recognizing the object in the point cloud frame by using the updated obstacle recognition algorithm.

5. The method according to claim 1, wherein presenting the initial labeling result of the point cloud frame further comprises:

synchronously and nonoverlappingly presenting, on the user interface, the point cloud frame, the recognized object labeled with the three-dimensional border, and an recognized object type information.

6. The method according to claim 5, wherein the user interface is a cloud visual tool.

7. The method according to claim 5, wherein label options for labeling object types are presented on the user interface.

8. An apparatus for processing a point cloud frame of point cloud data, comprising:

at least one processor; and

a memory storing instructions, which when executed by the at least one processor, cause the at least one processor to perform operations, the operations comprising;

recognizing an object in the point cloud frame by using an obstacle recognition algorithm;

presenting an initial labeling result of the point cloud frame on a user interface of an electronic device, wherein the initial labeling result comprises the recognized object and a three-dimensional border added to the recognized object; and

updating the initial labeling result in response to a correction operation by a user on the initial labeling result, wherein the correction operation comprises;

a border adjusting operation, the border adjusting operation including a translation operation, zooming operation, and rotation operation on only the three-dimensional border to fit the three-dimensional border to the recognized object with the recognized object remaining unadjusted by the border adjusting operation.

9. The apparatus according to claim 8, wherein the updating the correction operation further comprises:

a checking operation of the user on the an object type information and updating the initial labeling result according to the checked object type information.

10. The apparatus according to claim 8, wherein the operations further comprise:

training the obstacle recognition algorithm by using the updated initial labeling result as sample data, to update the obstacle recognition algorithm.

11. The apparatus according to claim 10, wherein the operations further comprise:

recognizing the object in the point cloud frame by using the updated obstacle recognition algorithm.

12. The apparatus according to claim 8, wherein presenting the initial labeling result of the point cloud frame operation further comprises:

synchronously and nonoverlappingly presenting, on the user interface, the point cloud frame, the recognized object labeled with the three-dimensional border, and an recognized object type information.

13. The apparatus according to claim 12, wherein the user interface is a cloud visual tool.

14. The apparatus according to claim 12, wherein label options for labeling object types are presented on the user interface.

15. A non-transitory storage medium storing one or more programs, the one or more programs when executed by an apparatus, causing the apparatus to perform a method for processing a point cloud frame of point cloud data, comprising:

recognizing an object in the point cloud frame by using an obstacle recognition algorithm;

presenting an initial labeling result of the point cloud frame on a user interface of an electronic device, wherein the initial labeling result comprises the recognized object and a three-dimensional border added to the recognized object; and

updating the initial labeling result in response to a correction operation by a user on the initial labeling result, wherein the correction operation comprises;

a border adjusting operation, the border adjusting operation including a translation operation, zooming operation, and rotation operation on only the three-dimensional border to fit the three-dimensional border to the recognized object with the recognized object remaining unadjusted by the border adjusting operation.

16. The non-transitory storage medium of claim 15, wherein presenting the initial labeling result of the point cloud frame further comprises:

synchronously and nonoverlappingly presenting, on the user interface, the point cloud frame, the recognized object labeled with the three-dimensional border, and an recognized object type information.

17. The method according to claim 16, wherein the user interface is a cloud visual tool.

18. The method according to claim 16, wherein label options for labeling object types are presented on the user interface.