Method, system and computer program product for producing a raised relief map from images of an object

Patent metadata

Field	Value
Publication number	US-10007995-B2
Application number	US-201414892826-A
Country	US
Kind code	B2
Filing date	May 22, 2014
Priority date	May 23, 2013
Publication date	Jun 26, 2018
Grant date	Jun 26, 2018

Abstract

A method for obtaining a model of an object surface from a plurality of images of said object, wherein the object is illuminated with one or more illumination sources, the method comprising: •—obtaining a set of images each comprising an array of pixels and representing similar views of an object, wherein the similar views are obtained under different illumination conditions; •—defining a model function that expresses the known pixel intensity values of the images in terms of the following unknown model components: •—a first model component (A) representing the albedo of the object at each pixel location and being the same for the plurality of images and being an albedo value; •—a second model component (L) representing an intensity of illumination for each image and being an illumination source intensity value being the same for all pixels of each image; •—a third model component (V) representing a specific illumination direction and being different for each image and being an illumination vector and being the same for all pixels of each image; •—a fourth model component (N) representing surface normal directions of the object surface at each pixel position and being the same for all images and being a normal vector; •—performing one or more sequences of minimization operations to minimize a difference function between the pixel values obtained from the set of images and pixel values calculated using said model function, each minimization operation being performed by allowing one of said model components (A, L, V, N) to vary while the others remain unchanged; •—outputting the fourth model component (N) as said model of the object surface.

First claim

The invention claimed is:

1. A method for obtaining a model of an object surface from a plurality of images of said object ( 107 ), wherein the object ( 107 ) is illuminated with one or more illumination sources ( 104 , 105 , 106 ), the method comprising:
obtaining a set of images each comprising an array of pixels and representing similar views of an object, wherein the similar views are obtained under different illumination conditions;
defining a model function (f) that expresses the known pixel intensity values of the images in terms of the following unknown model components:
a first model component (A) representing the albedo of the object ( 107 ) at each pixel location, the first model component being an albedo value which is the same for the plurality of images;
a second model component (L) representing an intensity of illumination for each image, the second model component being an illumination source intensity value which is the same for all pixels of each image;
a third model component (V) representing a specific illumination direction, the third model component being an illumination vector which is different for each image and the same for all pixels of each image;
a fourth model component (N) representing surface normal directions of the object surface at each pixel position, the fourth model component being a normal vector which is the same for all images;

performing one or more sequences of minimization operations to minimize a difference function between the pixel values obtained from the set of images and pixel values calculated using said model function (f), each minimization operation being performed by allowing one of said model components (A, L, V, N) to vary while the others remain unchanged;
outputting the fourth model component (N) as said model of the object surface, wherein model function (f) comprises a modified model function (g), wherein the model function (f) related to the first, second, third and fourth model components (A, L, V, N) is equal to the modified model function (g) related to a first, second and third model sub-components (A′, L′, V′) and the fourth model component (N), wherein
the first model sub-component (A′) is adapted by a first structuring component (W A ), to reproduce the surface properties of the object ( 107 );
the second model sub-component (L′) is adapted by a second structuring component (W L ), to reproduce the illumination conditions of the illumination sources ( 104 , 105 , 106 ); and

the third model sub-component (V′) is adapted by a third structuring component (W V ), to reproduce the illumination conditions of the illumination sources ( 104 , 105 , 106 ).

2. The method as claimed in claim 1 , wherein said sequence of minimization operations comprises an iterative loop in which each of said model components (A, L, V, N) is varied in turn at least once before a subsequent iteration, said iterative loop being repeated until a termination condition is met.

3. The method as claimed in claim 2 , wherein said iterative loop comprises one or more inner loops wherein a subset of said model components (V, N) are varied a number of times before another subset of said model components (A, L) is varied.

4. The method as claimed in claim 3 , wherein one inner loop comprises varying said third model component (V) and said fourth model component (N) a number of times.

5. The method as claimed in claim 3 , wherein one inner loop comprises varying said first model component (A) and said second model component (L).

6. The method as claimed in claim 1 , wherein minimization operations are performed, varying said third model component (V) and said fourth model component (N) a number of times before minimization is performed, varying said first model component (A) and said second model component (L).

7. The method as claimed in claim 1 , wherein the first model component (A) comprises first model sub-component (A′) and a first structuring component (W A ), where the first structuring component (W A ) adapts the first model sub-component (A′) to reproduce the surface properties of the object ( 107 ).

8. The method as claimed in claim 1 , wherein the second model component (L) comprises the second model sub-component (L′) and a second structuring component (W L ), wherein the second structuring component (W L ) adapts the second sub-component (L′) to reproduce the illumination conditions of the illumination sources ( 104 , 105 , 106 ).

9. The method as claimed in claim 1 , wherein the third model component (V) comprises the third model sub-component (V′) and a third structuring component (W v ), wherein the third structuring component (W v ) adapts the third model sub-component (V′) to reproduce the illumination conditions of the illumination sources ( 104 , 105 , 106 ).

10. The method as claimed in claim 7 , wherein the first structuring component (W A ) represents a same coefficient for the plurality of images comprising the same albedo value for each pixel for the plurality of images.

11. The method as claimed in claim 8 , wherein the second structuring component (W L ) represents a same intensity coefficient for all pixels of an image from the plurality of images.

12. The method as claimed in claim 9 , wherein the third structuring component (W v ) represents the direction of illumination of one illumination source ( 104 , 105 , 106 ) associated with one image of said plurality of images, this being the same for all pixels of an image from the plurality of images.

13. The method as claimed in claim 1 , wherein the one or more illumination sources ( 104 , 105 , 106 ) comprise Red, Green and Blue color channels.

14. The method as claimed in claim 13 , wherein the first model component (A) is the same for each pixel and each color channel for the plurality of images and the second model component (L) is the same for each color channel for all pixels of each image.

15. The method as claimed in claim 13 , wherein the first structuring component (W A ) comprises the same albedo value for each pixel for each color channel for the plurality of images.

16. The method as claimed in claim 13 , wherein the second structuring component (W L ) represents a same intensity coefficient for each color channel for all pixels of each image.

17. A computer program product comprising instructions which when executed cause a programmable data processing apparatus to perform steps of the method according to claim 1 .

18. An imaging system ( 100 ) for obtaining a model of an object surface from a plurality of images of said object ( 107 ), the imaging system ( 100 ) comprising:
one or more illumination sources ( 104 , 105 , 106 ) to illuminate the object ( 107 ) with different directions of illumination, wherein each image is associated with a determined direction of illumination;
an image capture device ( 108 ) for detecting the pixel values of each pixel in each image of the plurality of images;
a data processor ( 114 ) comprising a first memory component ( 116 ) for storing the following unknown model components:
a first model component (A) representing the albedo of the object ( 107 ) at each pixel location, the first model component being an albedo value which is the same for the plurality of images;
a second model component (L) representing an intensity of illumination for each image, the second model component being an illumination source intensity value which is the same for all pixels of each image;
a third model component (V) representing a specific illumination direction, the third model component being an ill

Assignees

• Biomerieux Sa
• Commissariat Energie Atomique

Inventors

• Allano Lorene
• Kyrgyzova Khrystyna

Classifications

• G06T2207/10152

  Varying illumination · CPC title

• G06T2207/10024

  Color image · CPC title

• G06T7/586 Primary

  from multiple light sources, e.g. photometric stereo · CPC title

• G06T15/506

  Illumination models · CPC title

• G06T5/50

  using two or more images, e.g. averaging or subtraction · CPC title