Method and arrangement for detecting free fibre ends in paper

Patent metadata

Field	Value
Publication number	US-9816977-B2
Application number	US-201414888989-A
Country	US
Kind code	B2
Filing date	May 9, 2014
Priority date	May 10, 2013
Publication date	Nov 14, 2017
Grant date	Nov 14, 2017

Abstract

The invention relates to a method and arrangement for detecting free fiber ends in a paper surface. The method comprises illuminating a target sample ( 6 ) surface, which comprises free fiber ends, from at least two directions one at the time, with at least one light source ( 1 ). Original reflectance images are obtained for the target sample ( 6 ) surface with an imaging device ( 4 ), and a surface normal is estimated for each image pixel of the original reflectance image. Thus it is possible to reconstruct a reconstructed reflectance image from the estimated surface normals, and to compare the reconstructed reflectance image and the corresponding original reflectance image and to construct a difference image, where the differences represent shadow objects of the free fiber ends in a paper surface.

First claim

The invention claimed is:

1. A method for detecting free fibre ends in a paper surface, the method comprising
illuminating a target sample surface, which comprises free fibre ends, from at least two directions one direction at a time, with at least one light source,
obtaining for each light source direction an original reflectance image for the target sample surface with an imaging device,
estimating a surface normal for each image pixel of the original reflectance image,
reconstructing a reconstructed reflectance image from the estimated surface normals, and
comparing the reconstructed reflectance image and the corresponding original reflectance image and constructing a difference image, where the differences represent shadow objects of the free fibre ends in a paper surface.

2. The method according to claim 1 , characterised in reconstructing a plurality of reconstructed reflectance images from the estimated surface normals,
comparing each reconstructed reflectance image and the corresponding original reflectance image and constructing difference images, where the differences represent shadow objects of the free fibre ends in a paper surface, and
calculating an average value of a number of the shadow objects in the difference images.

3. The method according to claim 1 , characterised in that a unit normal of the surface for each image pixel is estimated by minimizing a square of error with pseudoinverse as

n
=

(

L
T

⁢
L

)

-
1

⁢

L
T

⁢
i

ρ
⁢

⁢
E

where n is the unit normal of the surface, ρis a surface albedo describing reflectivity of the surface, E is an intensity of a light source, L is a matrix consisting of 1×3 unit vectors toward each light source, and i is an intensity vector of an image pixel for each light source direction.

4. The method according to claim 3 , characterised in that the reconstructed reflectance image is reconstructed by calculating a reflectance intensity for each image pixel from an equation
i=ρEL 1×3 n 3×1 .

5. The method according to claim 1 , characterised in compensating a beam pattern of the light source on the target sample surface by using a 2D second order polynomial fitted on the original reflectance image.

6. The method according to claim 1 , characterised in
compensating pixel intensity values by computing distances between each image pixel and the light source in order to obtain a matrix of pixel intensity compensation results, and
multiplying the original reflectance image pointwise with the matrix of pixel intensity compensation results and dividing the original reflectance image pointwise with a 2D polynomial.

7. The method according to claim 1 , characterised in detecting the shadow objects in the difference image by using filtering and/or processing methods which enforce the shadow objects of the free fibre ends in a paper surface.

8. The method according to claim 7 , characterised in
computing a histogram which shows a distribution of pixel values of the filtered/processed difference image in which the shadow objects are enforced,
setting a threshold limit to a desired value and obtaining a thresholded difference image,
removing circular objects from the thresholded difference image by using ellipse fitting algorithms, and
accepting from the thresholded difference image shadow objects whose length is larger than an acceptance limit, and/or objects having eccentricity exceeding a predetermined value, and/or objects which major axis deviates at the most 30 degrees, 45 degrees or 90 degrees from the direction of the light source.

9. The method according to claim 1 , characterised in illuminating the target sample surface, which comprises free fibre ends from at least 6 directions.

10. The method according to claim 1 , characterised in that the target sample is moving paper web and that the method is performed on-line.

11. The method according to claim 1 , characterised in counting a number of shadow objects and changing process parameters according to an obtained result.

12. The method according to claim 1 , characterised in
arranging a first linear polarizer in front of the imaging device and a second linear polarizer in front of the light source, the first linear polarizer and the second linear polarizer being at 90 degrees angle in relation to each other, and
keeping the orientation between the first linear polarizer and the second linear polarizer constant during a measurement.

13. The method according to claim 1 , characterised in illuminating the target sample surface, which comprises free fibre ends from at least 8 directions.

14. The method according to claim 1 , characterised in illuminating the target sample surface, which comprises free fibre ends from at least 10 directions.

15. The method according to claim 1 , characterised in illuminating the target sample surface, which comprises free fibre ends from at least 12 directions.

16. An arrangement for detecting free fibre ends in paper, the arrangement comprising
an imaging device such as a digital systems camera, arranged to a distance from a sample holder, for obtaining original reflectance images of a target sample surface,
at least two light sources attached around the imaging device, or one light source, which is attached to a supporting arm, which is arranged to rotate around the imaging device, the at least two light sources or the one light source being arranged to illuminate the target sample surface from at least two directions one direction at a time,
a data processing unit, which is arranged
to receive original reflectance images obtained for each light source direction from the imaging device,
to estimate a surface normal for each image pixel of the original reflectance image,
to reconstruct a reconstructed reflectance image from the estimated surface normals, and

to compare the reconstructed reflectance image and the corresponding original reflectance image and to construct a difference image, where the differences represent shadow objects of the free fibre ends in a paper surface.

17. The arrangement according to claim 16 , wherein the paper is a tissue paper grade.

18. The arrangement according to claim 16 , wherein the at least two light sources are LED light sources.

Assignees

• Kemira Oyj

Inventors

• Raunio Jukka-Pekka
• Makinen Mikko
• Skoog Henry
• Campbell Clayton

Classifications

• G01N33/34

  Paper · CPC title

• G01N21/8914

  characterised by the material examined · CPC title

• G01N2021/8924

  Dents; Relief flaws · CPC title

• G01N21/4738

  Diffuse reflection (precedence is given to G01N21/55 - G01N21/57 if specular component is taken into consideration), e.g. also for testing fluids, fibrous materials · CPC title

• G06T7/0004

  Industrial image inspection · CPC title