Method and system for producing classifier for recognizing obstacle

• US 9,207,320 B2
• Filed: 12/13/2012
• Issued: 12/08/2015
• Est. Priority Date: 08/06/2012
• Status: Active Grant

• Current Assignee: Hyundai Motor Company (Hyundai Motor Group)
• Original Assignee: Hyundai Motor Company (Hyundai Motor Group)
• Inventors: Kim, Jin Hak, Yoo, Kyung Ho, Jang, Yoon Ho, Im, Seong Su, Lee, Hui Sung, Min, Kyoung Moo, Choi, Eun Jin
• Primary Examiner(s): Bythrow, Peter
• Application Number:
  US13/713,820
• Publication Number:
  US 20140035777A1
• Time in Patent Office: 1,090 Days
• Field of Search: 342 70- 72, 342 90- 97, 342/175, 342/189, 382/104
• US Class Current: 1/1
• CPC Class Codes: G01S 13/86 Combinations of radar systems with non-radar systems, e.g. sonar, direction finder
• CPC Class Codes: G01S 13/91 for traffic control G01S13/93 takes precedence
• CPC Class Codes: G01S 13/93 for anti-collision purposes
• CPC Class Codes: G01S 13/931 of land vehicles
• CPC Class Codes: G01S 7/292 Extracting wanted echo-signals
• CPC Class Codes: G01S 7/41 using analysis of echo signal for target characterisation; Target signature; Target cross-section
• CPC Class Codes: G08G 1/16 Anti-collision systems road vehicle drive control systems for predicting or avoiding probable or impending collision otherwise than by control of a particular sub-unit B60W30/08

Abstract

The present invention provides a method and a system for producing a classifier for recognizing an obstacle, including a processor configured to: display surface data of a plurality of obstacles measured by a distance measurement sensor in a two-dimensional (2D) coordinate system; group and classify the surface data displayed in the 2D coordinate system for each obstacle; setting a plurality of feature references to analyze region based features displayed for each obstacle in the 2D coordinate system and calculate the respective feature references for each obstacle grouping; and producing the classifier by applying a weight to each of the feature references.

Claims

1. A system for producing a classifier for recognizing an obstacle, comprising:

a processor configured to;

display surface data of a plurality of obstacles measured by a distance measurement sensor in a two-dimensional (2D) coordinate system;

group and classify the surface data displayed in the 2D coordinate system for each obstacle;

project the surface data of each obstacle vertically to a ground surface coordinate system having an X-Y axis parallel to a ground surface;

set a plurality of feature references to analyze region based features displayed for each obstacle in the 2D coordinate system and calculate the respective feature references for each obstacle; and

produce the classifier by applying a weight to each of the feature references.

2. The system of claim 1, wherein the processor is further configured to control the weight when a result value is not matched with a predetermined value by deducing a result value depending on the produced classifier and comparing the result value with the predetermined value to display the surface data.

3. The system of claim 1, wherein the distance measurement sensor may be selected from a group consisting of:

a radar sensor and a laser sensor.

4. The system of claim 1, wherein the feature references further comprise:

an area reference, wherein the area reference is a space occupied by the obstacle grouping in the 2D coordinate system;

a centroid reference, wherein the centroid reference is a coordinate value of a centroid of the obstacle grouping displayed in the 2D coordinate system;

a spacing reference, wherein the spacing reference is a level in which the surface data corresponding to each obstacle displayed in the 2D coordinate system is spaced apart from the coordinate value of the centroid of each corresponding obstacle;

a perimeter reference, wherein the perimeter reference is a perimeter length value of the obstacle grouping; and

a roundness reference, wherein the roundness reference is a roundness level calculated by the area value and the perimeter length value of the obstacle grouping.

5. A non-transitory computer medium containing program instructions executed by a processor, the computer readable medium comprising:

program instructions that display surface data of a plurality of obstacles measured by a distance measurement sensor in a two-dimensional (2D) coordinate system;

program instructions that group and classify the surface data displayed in the 2D coordinate system for each obstacle;

program instructions that project the surface data of each obstacle vertically to a ground surface coordinate system having an X-Y axis parallel to a ground surface;

program instructions that set a plurality of feature references to analyze region based features displayed for each obstacle in the 2D coordinate system and calculate the respective feature references for each obstacle; and

program instructions that produce the classifier by applying a weight to each of the feature references.

6. The non-transitory computer medium of claim 5, further comprising program instructions that control the weight when a result value is not matched with a predetermined value by deducing a result value depending on the produced classifier and comparing the result value with the predetermined value to display the surface data.

7. The non-transitory computer medium of claim 5, wherein the distance measurement sensor may be selected from a group consisting of:

a radar sensor and a laser sensor.

8. The non-transitory computer medium of claim 5, wherein the feature references further comprise:

an area reference, wherein the area reference is a space occupied by the obstacle grouping in the 2D coordinate system;

a centroid reference, wherein the centroid reference is a coordinate value of a centroid of the obstacle grouping displayed in the 2D coordinate system;

a spacing reference, wherein the spacing reference is a level in which the surface data corresponding to each obstacle displayed in the 2D coordinate system is spaced apart from the coordinate value of the centroid of each corresponding obstacle;

a perimeter reference, wherein the perimeter reference is a perimeter length value of the obstacle grouping; and

a roundness reference, wherein the roundness reference is a roundness level calculated by the area value and the perimeter length value of the obstacle grouping.