Computing camera parameters

Patent metadata

Field	Value
Publication number	US-9886530-B2
Application number	US-201414541860-A
Country	US
Kind code	B2
Filing date	Nov 14, 2014
Priority date	Nov 18, 2013
Publication date	Feb 6, 2018
Grant date	Feb 6, 2018

Abstract

The invention notably relates to a computer-implemented method for designing a 3D modeled object representing a real object. The method comprises the steps of providing a 3D representation of the real object, identifying occurrences of a geometric feature at 3D positions of the 3D representation, providing at least one 2D view of the real object, identifying occurrences of a graphic feature at 2D positions of the 2D view, the geometric feature corresponding to the graphic feature, computing camera parameters that minimize a distance between a set of projections of the 3D positions on the 2D view and a set of 2D positions. This provides an improved solution for designing a 3D modeled object representing a real object.

First claim

What is claimed is:

1. A computer-implemented method for designing a 3D modeled object representing a real object, comprising the steps of:
providing a 3D representation of the real object;
identifying occurrences of a geometric feature at 3D positions of the 3D representation, the geometric feature being a 3D curvature above a predetermined 3D curvature threshold;
providing at least one 2D view of the real object, the 2D view being an image of the real object;
identifying occurrences of a graphic feature at 2D positions of the 2D view, the geometric feature corresponding to the graphic feature, the graphic feature relating to pixel gradient;
providing 3D positions at which occurrences of the geometric feature are identified, the 3D representation having a 3D curvature above the predetermined 3D curvature threshold at the 3D positions;
providing 2D positions at which occurrences of the graphic feature are identified; and
computing camera parameters that minimize a distance between a set of projections of the provided 3D positions on the 2D view and a set of 2D positions.

2. The method of claim 1 , wherein the graphic feature is a pixel gradient above a predetermined threshold.

3. The method of claim 2 , wherein the method further comprises determining a texture mapping between the 3D representation and the 2D view based on the computed camera parameters.

4. The method of claim 1 , wherein the identifying of occurrences of the graphic feature includes first applying a Gaussian blur on the 2D view.

5. The method of claim 1 , wherein the method further comprises, prior to the computing, filtering 3D positions of the 3D representation according to criteria evaluated according to a predetermined estimation of the camera parameters.

6. The method of claim 5 , wherein the criteria include visibility on the 2D view and/or distance of the projection with borders of the 2D view.

7. The method of claim 6 , wherein the identifying of occurrences of the graphic feature is iterated, with an increasing number of occurrences of the graphic feature at each iteration, and the computing of camera parameters includes an iterative optimization algorithm.

8. The method of claim 7 , wherein the identifying of occurrences of the graphic feature is iterated over resized versions of the 2D view, that are ordered from the smallest to the largest.

9. The method of claim 1 , wherein the intrinsic camera parameters are predetermined, the computed camera parameters thereby being the extrinsic camera parameters.

10. The method of claim 1 , wherein providing the 3D representation of the real object comprises reconstructing the 3D representation from measurements of at least one depth sensor.

11. The method of claim 10 , wherein the 2D view is an image captured by a high resolution camera and/or by a camera synchronous to the depth sensor.

12. A non-transitory computer readable storage medium comprising:
a computer memory device having recorded thereon a computer program comprising instructions for execution by a computer, the computer when executing the instructions performs design of a 3D modeled object representing a real object by:
providing a 3D representation of the real object;
identifying occurrences of a geometric feature at 3D positions of the 3D representation, the geometric feature being a 3D curvature above a predetermined 3D curvature threshold;
providing at least one 2D view of the real object, the 2D view being an image of the real object;
identifying occurrences of a graphic feature at 2D positions of the 2D view, the geometric feature corresponding to the graphic feature, the graphic feature relating to pixel gradient;
providing 3D positions at which occurrences of the geometric feature are identified, the 3D representation having a 3D curvature above the predetermined 3D curvature threshold at the 3D positions;
providing 2D positions at which occurrences of the graphic feature are identified; and
computing camera parameters that minimize a distance between a set of projections of the provided 3D positions on the 2D view and a set of 2D positions.

13. A system comprising a processor coupled to a memory and a graphical user interface, the system being configured for the design of a 3D modeled object representing a real object by:
providing a 3D representation of the real object;
identifying occurrences of a geometric feature at 3D positions of the 3D representation, the geometric feature being a 3D curvature above a predetermined 3D curvature threshold;
providing at least one 2D view of the real object, the 2D view being an image of the real object;
identifying occurrences of a graphic feature at 2D positions of the 2D view, the geometric feature corresponding to the graphic feature, the graphic feature relating to pixel gradient;
providing 3D positions at which occurrences of the geometric feature are identified, the 3D representation having a 3D curvature above the predetermined 3D curvature threshold at the 3D positions;
providing 2D positions at which occurrences of the graphic feature are identified; and
computing camera parameters that minimize a distance between a set of projections of the provided 3D positions on the 2D view and a set of 2D positions.

14. The system of claim 13 , wherein the system further determines a texture mapping between the 3D representation and the 2D view based on the computed camera parameters.

15. The system of claim 13 , wherein the identifying of occurrences of the graphic feature includes first applying a Gaussian blur on the 2D view.

16. The system of claim 13 , wherein the system, prior to the computing, filters 3D positions of the 3D representation according to criteria evaluated according to a predetermined estimation of the camera parameters.

17. The system of claim 13 , wherein the intrinsic camera parameters are predetermined, the computed camera parameters thereby being the extrinsic camera parameters.

18. The system of claim 13 , wherein providing the 3D representation of the real object comprises reconstructing the 3D representation from measurements of at least one depth sensor, and wherein the 2D view is an image captured by a high resolution camera and/or by a camera synchronous to the depth sensor.

Assignees

• Dassault Systemes
• Dassault Systems

Inventors

• Mehr Eloi
• Guitteny Vincent
• Boufarguine Mourad

Classifications

• G06T2200/08

  involving all processing steps from image acquisition to 3D model generation · CPC title

• G06T2207/10028

  Range image; Depth image; 3D point clouds · CPC title

• G06T17/00

  Three-dimensional [3D] modelling for computer graphics · CPC title

• G06T2200/04

  involving 3D image data · CPC title

• G06T17/10

  Constructive solid geometry [CSG] using solid primitives, e.g. cylinders, cubes · CPC title