Three-dimensional modeling apparatus, three-dimensional modeling method and computer-readable recording medium storing three-dimensional modeling program

• US 8,928,736 B2
• Filed: 04/05/2012
• Issued: 01/06/2015
• Est. Priority Date: 04/06/2011
• Status: Active Grant

• Current Assignee: Casio Computer Company Limited
• Original Assignee: Casio Computer Company Limited
• Inventors: Nakajima, Mitsuyasu, Yamaya, Takashi
• Primary Examiner(s): Perungavoor, Sath V
• Assistant Examiner(s): Le, Peter D
• Application Number:
  US13/440,909
• Publication Number:
  US 20120257016A1
• Time in Patent Office: 1,006 Days
• US Class Current: 348/46
• CPC Class Codes: G06T 17/20 Finite element generation, e.g. wire-frame surface description, tesselation
• CPC Class Codes: G06T 2200/08 involving all processing steps from image acquisition to 3D model generation
• CPC Class Codes: G06T 7/593 from stereo images

Abstract

In three-dimensional modeling apparatus, an image obtaining section obtains image sets picked up by stereoscopic camera. A generating section generates three-dimensional models. A three-dimensional model selecting section selects a first three-dimensional model and a second three-dimensional model to be superimposed on the first three-dimensional model among generated three-dimensional models. A extracting section extracts first and second feature points from the selected first and second three-dimensional model. A feature-point selecting section selects feature points having a closer distance to stereoscopic camera from the extracted first and second feature points. A parameter obtaining section obtains a transformation parameter for transforming a coordinate of the second three-dimensional model into a coordinate system of the first three-dimensional model. A transforming section transforms the coordinate of the second three-dimensional model into the coordinate system of the first three-dimensional model. And a superimposing section superimposes the second three-dimensional model on the first three-dimensional model.

Claims

1. A three-dimensional modeling apparatus comprising:

an image obtaining section that obtains image sets picked up by a stereoscopic camera;

a generating section that generates three-dimensional models of an object based on each of the obtained image sets;

a three-dimensional model selecting section that selects a first three-dimensional model and a second three-dimensional model to be superimposed on the first three-dimensional model among the generated three-dimensional models;

an extracting section which extracts first feature points from the selected first three-dimensional model and second feature points from the selected second three-dimensional model;

a feature-point selecting section that selects feature points having a closer distance to the stereoscopic camera from the extracted first feature points and the extracted second feature points, respectively;

a parameter obtaining section that obtains a transformation parameter for transforming a coordinate of the second three-dimensional model into a coordinate in a coordinate system of the first three-dimensional model based on the selected first and second feature points;

a transforming section that transforms the coordinate of the second three-dimensional model into the coordinate in the coordinate system of the first three-dimensional model using the transformation parameter; and

a superimposing section that superimposes the second three-dimensional model having a transformed coordinate on the first three-dimensional model,

wherein the feature-point selecting section (i) performs clustering on the extracted first and second feature points to create two clusters based on a distance from the stereoscopic camera, and (ii) selects the first feature point and the second feature point belonging to respective clusters having the closer distance to the stereoscopic camera.

2. The three-dimensional modeling apparatus according to claim 1, wherein the feature-point selecting section (i) applies a robust estimation technique, which uses a predetermined evaluation formula so that the closer a distance to the stereoscopic camera is, the higher an evaluation value becomes, to the extracted first and second feature points, and (ii) selects an estimated feature point.

3. A three-dimensional modeling apparatus comprising:

an obtaining section that obtains a first image and a second image obtained by imaging an object at different positions;

an extracting section which extracts first feature points on the first image and second feature points on the second image corresponding to the first feature points, respectively;

a selecting section that selects a combination of feature points having a closer distance to a camera at a time of imaging among the first feature points and the second feature points corresponding to the first feature points;

a parameter obtaining section that obtains a transformation parameter indicating a positional relationship of a position at which the first image is imaged and a position at which the second image is imaged based on the selected combination of the first and second feature points;

an information obtaining section that obtains pieces of three-dimensional information of the first feature points and the second feature points extracted by the extracting section based on the transformation parameter; and

a generating section that generates a three-dimensional model based on the three-dimensional information,

wherein the selecting section (i) performs clustering on the extracted first and second feature points to create two clusters based on a distance from the camera at the time of imaging, and (ii) selects the combination of feature points belonging to respective clusters having the closer distance to the camera at the time of imaging.

4. A three-dimensional modeling method comprising:

obtaining image sets picked up by a stereoscopic camera;

generating three-dimensional models of an object based on each of the obtained image sets;

selecting a first three-dimensional model and a second three-dimensional model to be superimposed on the first three-dimensional model among the generated three-dimensional models;

extracting first feature points from the selected first three-dimensional model and second feature points from the selected second three-dimensional model;

selecting feature points having a closer distance to the stereoscopic camera from the extracted first feature points and the extracted second feature points, respectively;

obtaining a transformation parameter for transforming a coordinate of the second three-dimensional model into a coordinate in a coordinate system of the first three-dimensional model based on the selected first and second feature points;

transforming the coordinate of the second three-dimensional model into the coordinate in the coordinate system of the first three-dimensional model using the transformation parameter; and

superimposing the second three-dimensional model having a transformed coordinate on the first three-dimensional model,

wherein selecting the feature points having the closer distance to the stereoscopic camera comprises;

clustering the extracted first and second feature points to create two clusters based on a distance from the stereoscopic camera; and

selecting the first feature point and the second feature point belonging to respective clusters having the closer distance to the stereoscopic camera.

5. A three-dimensional modeling method comprising:

obtaining a first image and a second image obtained by imaging an object at different positions;

extracting first feature points on the first image and second feature points on the second image corresponding to the first feature points, respectively;

selecting a combination of feature points having a closer distance to a camera at a time of imaging among the first feature points and the second feature points corresponding to the first feature points;

obtaining a transformation parameter indicating a positional relationship of a position at which the first image is imaged and a position at which the second image is imaged based on the selected combination of the first and second feature points;

obtaining pieces of three-dimensional information of the extracted first feature points and the extracted second feature points based on the transformation parameter; and

generating a three-dimensional model based on the three-dimensional information,

wherein selecting the combination of feature points having the closer distance to the camera at the time of imaging comprises;

clustering the extracted first and second feature points to create two clusters based on a distance from the camera at the time of imaging; and

selecting the combination of feature points belonging to respective clusters having the closer distance to the camera at the time of imaging.

6. A non-transitory computer-readable recording medium having stored therein a program executable by a computer, the program causing the computer to realize functions of:

obtaining image sets picked up by a stereoscopic camera;

generating three-dimensional models of an object based on each of the obtained image sets;

selecting a first three-dimensional model and a second three-dimensional model to be superimposed on the first three-dimensional model among the generated three-dimensional models;

extracting first feature points from the selected first three-dimensional model and second feature points from the selected second three-dimensional model;

selecting feature points having a closer distance to the stereoscopic camera from the extracted first feature points and the extracted second feature points, respectively;

obtaining a transformation parameter for transforming a coordinate of the second three-dimensional model into a coordinate in a coordinate system of the first three-dimensional model based on the selected first and second feature points;

transforming the coordinate of the second three-dimensional model into the coordinate in the coordinate system of the first three-dimensional model using the transformation parameter; and

superimposing the second three-dimensional model having a transformed coordinate on the first three-dimensional model,

wherein selecting the feature points having the closer distance to the stereoscopic camera comprises;

clustering the extracted first and second feature points to create two clusters based on a distance from the stereoscopic camera; and

selecting the first feature point and the second feature point belonging to respective clusters having the closer distance to the stereoscopic camera.

7. A non-transitory computer-readable recording medium having stored therein a program executable by a computer, the program causing the computer to realize functions of:

obtaining a first image and a second image obtained by imaging an object at different positions;

extracting first feature points on the first image and second feature points on the second image corresponding to the first feature points, respectively;

selecting a combination of feature points having a closer distance to a camera at a time of imaging among the first feature points and the second feature points corresponding to the first feature points;

obtaining a transformation parameter indicating a positional relationship of a position at which the first image is imaged and a position at which the second image is imaged based on the selected combination of the first and second feature points;

obtaining pieces of three-dimensional information of the extracted first feature points and the extracted second feature points based on the transformation parameter; and

generating a three-dimensional model based on the three-dimensional information,

wherein selecting the combination of feature points having the closer distance to the camera at the time of imaging comprises;

clustering the extracted first and second feature points to create two clusters based on a distance from the camera at the time of imaging; and

selecting the combination of feature points belonging to respective clusters having the closer distance to the camera at the time of imaging.