Object detection apparatus, object detection method, object detection program, and distance sensor

• US 7,463,184 B2
• Filed: 04/29/2005
• Issued: 12/09/2008
• Est. Priority Date: 05/13/2003
• Status: Expired due to Fees

• Current Assignee: Fujitsu Limited
• Original Assignee: Fujitsu Limited
• Inventors: Segawa, Eigo
• Primary Examiner(s): Tarcza; Thomas H.
• Assistant Examiner(s): LIU, HARRY K
• Application Number:
  US11/117,678
• Publication Number:
  US 20050206553A1
• Time in Patent Office: 1,320 Days
• Field of Search: 250/221, 342/52, 375/240, 705/4
• US Class Current: 342/52
• CPC Class Codes: G01S 13/89 for mapping or imaging
• CPC Class Codes: G01S 13/931 of land vehicles
• CPC Class Codes: G01V 3/12 operating with electromagnetic waves operating with millimetre waves G01V8/005

Abstract

The present invention provides an object detection apparatus for detecting at least one object existing in an observation subject space. The object detection apparatus comprises an object model recording unit 3 that calculates and records an object model, which is an image expressing the form of a detection subject, a millimeter wave distance sensor 1 that calculates the distance to an object candidate by emitting millimeter waves so as to cover the entirety of the observation subject space at one time, a stereo image sensor 2 that calculates a characteristic image and a distance image having different viewpoints, a control unit 7 that selects an object candidate, a model conversion unit 4 that calculates a comparison model from the object model, a comparison image extraction unit 5 that extracts a comparison image from the characteristic image and distance image, and a similarity comparison unit 6 that calculates the similarity between the comparison model and comparison image, and when a maximum similarity is within a predetermined range, calculates the three-dimensional position of the object from the coordinates of the comparison image and the distance to the object candidate, and outputs the three-dimensional position of the object to the outside.

Claims

1. An object detection apparatus for detecting a detection subject existing in a predetermined space, comprising:

an object model recording unit that records an object model, the object model being an image expressing a form of at least one detection subject to be detected;

a distance sensor that detects at least one object candidate and calculates the distance to said object candidate by emitting electromagnetic waves in a radial form so as to cover the entirety of said predetermined space and detecting reflected waves of said electromagnetic waves;

a stereo image sensor that captures two images of said object candidate from different viewpoints and obtains image information from said two captured images;

a model conversion unit that calculates a comparison model, which is an image in which the size of said object model is converted to a size corresponding to said calculated distance to said object candidate; and

a determination unit that extracts a comparison image from said image information and determines a similarity between said comparison model and said comparison image,

wherein said determination unit determines whether or not said object candidate is said detection subject on the basis of said similarity between said comparison model and said comparison image.

2. The object detection apparatus as set forth in claim 1, wherein said image information obtained by said stereo image sensor is information of a distance image relating to the distance of said object candidate.

3. The object detection apparatus as set forth in claim 1, wherein said image information obtained by said stereo image sensor is information of a characteristic image relating to the form of said object candidate.

4. The object detection apparatus as set forth in claim 1, wherein said image information obtained by said stereo image sensor is information of a distance image relating to the distance of said object candidate and information of a characteristic image relating to the form of said object candidate.

5. The object detection apparatus as set forth in claim 1, wherein said object model recorded in said object model recording unit takes a form obtained by capturing an image of said detection subject at a predetermined distance.

6. The object detection apparatus as set forth in claim 1, further comprising:

a position calculation unit which, when said object candidate is determined to be said detection subject by said determination unit, calculates a three-dimensional position of said object candidate on the basis of said image information and the distance to said object candidate, assuming that said object candidate is said detection subject.

7. The object detection apparatus as set forth in claim 1, wherein said distance sensor calculates the distance to said object candidate from said reflected waves exceeding a predetermined threshold in intensity.

8. The object detection apparatus as set forth in claim 1, wherein said electromagnetic waves are millimeter waves.

9. An object detection method for detecting a detection subject existing in a predetermined space, comprising:

recording a model form of at least one detection subject to be detected as an object model;

detecting at least one object candidate and calculating the distance to said object candidate by emitting electromagnetic waves in a radial form so as to cover the entirety of said predetermined space and detecting reflected waves of said electromagnetic waves;

capturing two images of said object candidate from different viewpoints and obtaining image information from said two captured images;

calculating a comparison model, which is an image in which the size of said object model is converted to a size corresponding to said calculated distance to said object candidate;

extracting a comparison image from said image information;

determining a similarity between said comparison model and said comparison image; and

determining whether or not said object candidate is said detection subject on the basis of said similarity between said comparison model and said comparison image.

10. The object detection method as set forth in claim 9, further comprising, when said object candidate is determined to be said detection subject in the determination step, calculating a three-dimensional position of said object candidate on the basis of said image information and the distance to said object candidate, assuming that said object candidate is said detection subject.

11. An object detection program stored on a computer-readable medium, for causing a computer to execute an object detection method for detecting a detection subject existing in a predetermined space, said program comprising:

recording a model form of at least one detection subject to be detected as an object model;

detecting at least one object candidate and calculating the distance to said object candidate by emitting electromagnetic waves in a radial form so as to cover the entirety of said predetermined space and detecting reflected waves of said electromagnetic waves;

capturing two images of said object candidate from different viewpoints and obtaining image information from said two captured images; and

calculating a comparison model, which is an image in which the size of said object model is converted to a size corresponding to said calculated distance to said object candidate;

extracting a comparison image from said image information;

determining a similarity between said comparison model and said comparison image; and

determining whether or not said object candidate is said detection subject on the basis of said similarity between said comparison model and said comparison image.

12. The object detection program as set forth in claim 11, further comprising, when said object candidate is determined to be said detection subject in the determination step, calculating a three-dimensional position of said object candidate on the basis of said image information and the distance to said object candidate, assuming that said object candidate is said detection subject.