Systems and methods for generating three dimensional, textured models

• US 6,201,546 B1
• Filed: 05/29/1998
• Issued: 03/13/2001
• Est. Priority Date: 05/29/1998
• Status: Expired due to Fees

• Current Assignee: Pointcloud Inc.
• Original Assignee: Pointcloud Inc.
• Inventors: Bodor, Robert, Myers, Thomas B.
• Primary Examiner(s): Powell, Mark R.
• Assistant Examiner(s): HAVAN, THU THAO
• Application Number:
  US09/087,077
• Time in Patent Office: 1,019 Days
• Field of Search: 345/430, 345/419, 345/423, 345/425, 345/427, 345/428, 345/420, 345/139, 345/150, 345/429, 345/418, 703/6, 703/1, 382/154, 382/441, 382/285, 382/276
• US Class Current: 345/620
• CPC Class Codes: G06T 15/10 Geometric effects
• CPC Class Codes: G06T 2200/08 involving all processing steps from image acquisition to 3D model generation
• CPC Class Codes: G06T 7/40 Analysis of texture depth or shape recovery from texture G06T7/529
• CPC Class Codes: G06T 7/97 Determining parameters from multiple pictures depth or shape recovery from multiple images G06T7/55; stereo camera calibration G06T7/85

Abstract

A method for creating a three dimensional, textured model of at least one object is provided. The method includes identifying regions of the at least one object in a number of two dimensional images of the at least one object. The method further interrelates the regions identified from the two dimensional images and stores data concerning the region, including its texture, in a data structure that represents the model of the object.

Claims

1. A method for creating a three dimensional, textured model of at least one object from a number of two dimensional images, the method comprising:

loading an initial two dimensional image;

identifying a region of the at least one object in the loaded, two dimensional image;

storing data concerning the region, including its texture, in a data structure;

sequentially loading the remaining two-dimensional images; and

for each of the remaining two dimensional images;

identifying a relationship between the two dimensional image and at least one other, loaded two dimensional image,

identifying a region of the at least one object in the two dimensional image, and

storing data concerning the region, including information on its position with respect to other regions, and the texture of the region, in a data structure.

2. The method of claim 1, and further comprising preprocessing the images to identify interrelationships between the images by viewing thumbnails of each image and identifying the interrelationships.

3. The method of claim 1, and further comprising identifying a dimension of a region in an image to provide a basis for correlating the relative size of the images.

4. The method of claim 1, and further comprising acquiring the images with a wide angle lens.

5. The method of claim 4, and further comprising removing distortions in the image created by the wide angle lens.

6. The method of claim 1, and further comprising converting each image into a plane pixel array prior to processing.

7. The method of claim 1, and further comprising transforming a region to align edges of connecting regions from first and second images.

8. The method of claim 1, wherein identifying a region comprises identifying the vertices of the region using a clipping tool.

9. The method of claim 8, wherein identifying the vertices comprises identifying vertices of a four sided polygon.

10. The method of claim 1, and further providing real time display of a top view of the model.

11. The method of claim 1, and further providing a three dimensional view of the model that is updated as regions are added to the data structure.

12. The method of claim 1, wherein storing data in a data structure comprises storing data in a data structure that includes a WorldObject with data on the overall model, including a linked list of all regions, and a number of SurfaceObjects that contain data on each of the regions, including a pixel array with the texture for the region.

13. The method of claim 1, and further including defining openings in a region.

14. The method of claim 13, and further automatically aligning openings in back to back regions.

15. The method of claim 1, and further defining regions that intersect another region at a line other than an edge of the region.

16. A method for creating a three dimensional, textured model of at least one object, the method comprising:

identifying regions of the at least one object in a number of two dimensional images of the at least one object;

interrelating the regions identified from the two dimensional images; and

storing data concerning the region, including its texture, in a data structure that represents the model of the object.

17. The method of claim 16, and further comprising preprocessing the two dimensional images to identify interrelationships between the two dimensional images by viewing thumbnails of each two dimensional image and identifying the interrelationships.

18. The method of claim 16, and further comprising identifying a dimension of a region in an image to provide a basis for correlating the relative size of the images.

19. The method of claim 16, and further comprising transforming a region to align edges of connecting regions from first and second images.

20. The method of claim 16, wherein identifying a region comprises identifying the vertices of the region using a clipping tool.

21. The method of claim 20, wherein identifying the vertices comprises identifying vertices of a four sided polygon.

22. The method of claim 16, and further providing real time display of a top view of the model.

23. The method of claim 16, and further providing a three dimensional view of the model that is updated as regions are added to the data structure.

24. The method of claim 16, wherein storing data in a data structure comprises storing data in a data structure that includes a WorldObject with data on the overall model, including a linked list of all regions, and a number of SurfaceObjects that contain data on each of the regions, including a pixel array with the texture for the region.

25. The method of claim 16, and further including defining openings in a region.

26. The method of claim 25, and further automatically aligning openings in back to back regions.

27. The method of claim 16, and further defining regions that intersect another region at a line other than an edge of the region.

28. A method for generating a three dimensional, textured model of at least one object, the method comprising:

determining the dimensions, in pixels, of at least one source image;

performing a series of transformations on target polygons of a selected model to resize the model so that its dimensions match those of the at least one source image; and

mapping the texture from the at least one source image onto the correctly-sized polygons without having to distort or transform the image.

29. The method of claim 28, and further comprising selecting a model from a library of models.

30. The method of claim 28, and further comprising posting the completed image to an Internet site.

31. The method of claim 28, and further comprising receiving the source image at an Internet site over the Internet.

32. The method of claim 28, wherein the source image comprises images for a packaging label.

33. The method of claim 28, wherein the source images comprises images for printed media.

34. The method of claim 28, and further comprising sub-dividing the image into a number of images for mapping to the polygons of the model.

35. The method of claim 34, wherein sub-dividing the image includes detecting the edges of the image.

36. The method of claim 28, and further comprising generating the source image from a data file.

37. The method of claim 28, and further comprising generating the source image from a format defining data file for a printed media.

38. A method for generating a three dimensional, textured model of at least one object from a single texture image, the method comprising:

selecting a wire-frame base model;

loading the single texture image;

detecting the edges of the image to find an outline;

subdividing the image into multiple images based on the selected model and the outline;

determining image sizes;

adjusting the dimensions of the model based on the dimensions of the images;

applying the subdivided images to the adjusted model; and

outputting the model into a selected format.

39. The method of claim 38, and further comprising posting the completed image to an Internet site in VRML format.

40. The method of claim 38, and further comprising receiving the source image at an Internet site over the Internet.

41. The method of claim 38, wherein the source image comprises an image for a packaging label.

42. A machine readable medium having instructions stored thereon for causing execution of a method comprising:

identifying regions of the at least one object in a number of two dimensional images of the at least one object;

interrelating the regions identified from the two dimensional images; and

storing data concerning the region, including its texture, in a data structure that represents the model of the object.

43. A machine readable medium having instructions stored thereon for causing execution of a method comprising:

determining the dimensions, in pixels, of at least one source image;

performing a series of transformations on target polygons of a selected model to resize the model so that its dimensions match those of the at least one source image; and

mapping the texture from the at least one source image onto the correctly-sized polygons without having to distort or transform the image.