Three dimensional modeling apparatus

• US 20030038801A1
• Filed: 08/21/2002
• Published: 02/27/2003
• Est. Priority Date: 08/24/2001
• Status: Active Grant

• Current Assignee: Mind Fusion, LLC (Ascend Innovation Management, LLC)
• Original Assignee: Sanyo Electric Company Limited (Panasonic Holdings Corporation)
• Inventors: Fujimura, Kouta, Hojo, Mikio, Terauchi, Tomoya
• Granted Patent: US 6,809,728 B2
• Time in Patent Office: Days
• US Class Current: 345/420
• CPC Class Codes: G06T 1/0007 Image acquisition
• CPC Class Codes: G06T 17/00 Three dimensional [3D] modelling, e.g. data description of 3D objects
• CPC Class Codes: G06T 2207/10012 Stereo images
• CPC Class Codes: G06T 7/85 Stereo camera calibration
• CPC Class Codes: H04N 13/221 using the relative movement between cameras and objects
• CPC Class Codes: H04N 13/239 using two 2D image sensors having a relative position equal to or related to the interocular distance H04N13/243 takes precedence
• CPC Class Codes: H04N 13/246 Calibration of cameras
• CPC Class Codes: H04N 2013/0081 Depth or disparity estimation from stereoscopic image signals

Abstract

Images of a reference object for detecting each camera position are captured by cameras 14, 16 and the relative position between the cameras 14 and 16 is obtained by using the captured images. Then, a target 18 is placed on a turntable 10 onto which reference marks 10a for detecting position are formed, and while the turntable 10 is rotated, images are captured by the cameras 14 and 16. Based on the reference marks 10a on the turntable 10, the positional relationship between the camera 14 and the target 18 is determined, and, based on the relative position of the cameras 14 and 16, the positional relationship between the target 18 and the camera 16 is determined. A three-dimensional model of the target is produced based on the images captured by the cameras 14 and 16. In this manner, a target of relatively large size is easily three-dimensionally modeled.

Claims

1. A three-dimensional modeling apparatus comprising:

means for calculating relative positions among a plurality of cameras by using captured images of a reference object for detecting each camera position, the images being obtained from the plurality of cameras; and

means for producing a three-dimensional model of a target object from captured images, obtained from the plurality of cameras, of the target object on a turntable having reference marks for detecting position provided thereon, the modeling target rotating with the turntable, wherein

the means for producing a three-dimensional model calculates, based on images captured by a camera which captures images of the turntable, position information with respect to the turntable and the camera;

the means for producing a three-dimensional model further calculates, using the position information and the relative position obtained by the means for calculating relative positions, position information with respect to the turntable and a camera which does not capture images of the turntable; and

a three-dimensional model of the modeling target is produced based on the obtained position information of each camera and images captured by each camera.

2. A three-dimensional modeling apparatus according to claim 1, wherein

the plurality of cameras are placed at positions separated in the vertical direction so that images of the modeling target on the turntable are captured from different vertical positions.

3. A three-dimensional modeling apparatus according to claim 1, wherein, while the turntable is continuously rotated, a plurality of images are captured by each of the plurality of cameras which are synchronized with each other.

4. A three-dimensional modeling apparatus according to claim 1, wherein a voting process is separately applied to the images captured by each camera to generate voxel data, and the voxel data are then synthesized.

5. A three-dimensional modeling apparatus according to claim 1, wherein images captured by the cameras are synthesized based on the relative positions of the cameras, and a voting process is then applied based on the synthesized data to produce voxels.

6. A method for producing a three-dimensional model, comprising the steps of:

calculating relative positions among a plurality of cameras by capturing images of a reference object for detecting position by the plurality of cameras; and

capturing images, with the plurality of cameras, of a modeling target on a turntable having reference marks for detecting position provided thereon, the modeling target rotating with the turntable, and producing a three-dimensional model of the modeling target from the obtained captured images, wherein

in the step for capturing images and producing a three-dimensional model, position information with respect to the turntable and a camera which captures images of the turntable is calculated based on the images captured by the camera, and, using the position information and the relative position obtained in the step for calculating the relative positions, position information with respect to the turntable and a camera which does not capture images of the turntable is calculated; and

a three-dimensional model of the modeling target is produced based on the obtained position information of each camera and images captured by each camera.

7. A three-dimensional modeling method according to claim 6, wherein while the turntable is continuously rotated, a plurality of images are captured by each of the plurality of cameras which are synchronized with each other.

8. A three-dimensional modeling method according to claim 6, wherein a voting process is separately applied to the images captured by each camera to generate voxel data, and the voxel data are then synthesized.

9. A three-dimensional modeling method according to claim 6, wherein images captured by the cameras are synthesized based on the relative positions of the cameras, and a voting process is then applied based on the synthesized data to generate voxel data.

10. A three-dimensional modeling program for producing a three-dimensional model of a modeling target from a plurality of images of the modeling target captured by a plurality of cameras, which causes the execution of the steps of:

calculating relative positions among a plurality of cameras from captured images of a reference object for detecting position, the images obtained from a plurality of cameras; and

producing a three-dimensional model of a modeling target from captured images, obtained from the plurality of cameras, of the modeling target on a turntable having reference marks for detecting position provided thereon, the modeling target rotating with the turntable, wherein

in the step for producing three-dimensional model, position information with respect to the turntable and a camera which captures images of the turntable is calculated based on the images captured by the camera, and, using the position information and relative position information obtained in the step of calculating the relative positions, position information with respect to the turntable and a camera which does not capture images of the turntable is calculated, and

a three-dimensional model of the modeling target is produced based on the obtained position information of each camera and images captured by each camera.

11. A three-dimensional modeling program product according to claim 10, wherein a voting process is separately applied to the images captured by each camera to generate voxel data, and the voxel data are then synthesized.

12. A three-dimensional modeling program product according to claim 10, wherein images captured by the cameras are synthesized based on the relative positions of the cameras, and a voting process is then applied based on the synthesized data to generate voxel data.