Image analysis apparatus, image analysis method, and program

Patent metadata

Field	Value
Publication number	US-10068349-B2
Application number	US-201515514712-A
Country	US
Kind code	B2
Filing date	Sep 29, 2015
Priority date	Sep 29, 2014
Publication date	Sep 4, 2018
Grant date	Sep 4, 2018

Abstract

An image analysis apparatus, which analyzes orientations of fiber bundles of X-yarns and Y-yarns from a three-dimensional image of a woven fabric made of fiber bundles of the X-yarns, the Y-yarns, and Z-yarns, includes: a binarization unit that binarizes the three-dimensional image; an overlapping area extraction unit that extracts an overlapping area from the binarized image; a reference direction determination unit that averages an overlapping direction of each voxel included in the overlapping area and determines the averaged direction as a reference direction; a Z-yarn removal unit that removes the Z-yarns from the binarized image by applying a directional distance method on a reference plane perpendicular to the reference direction; and a fiber bundle orientation estimation unit that applies the directional distance method on the reference plane and estimates the orientations of the fiber bundles of the X-yarns and the Y-yarns based on a calculated directional distance.

First claim

The invention claimed is:

1. An image analysis apparatus for analyzing orientations of fiber bundles of X-yarns and Y-yarns from a three-dimensional image of a woven fabric made of fiber bundles of the X-yarns, the Y-yarns, and Z-yarns,
the image analysis apparatus comprising:
a binarization unit that binarizes the three-dimensional image;
an overlapping area extraction unit that extracts an overlapping area, in which the X-yarns and the Y-yarns perpendicularly and three-dimensionally intersect with each other, from the binarized image;
a reference direction determination unit that averages an overlapping direction of each voxel included in the overlapping area and determines the averaged direction as a reference direction;
a Z-yarn removal unit that removes the Z-yarns from the binarized image by applying a directional distance method on a reference plane perpendicular to the reference direction; and
a fiber bundle orientation estimation unit that applies the directional distance method again to the image, from which the Z-yarns have been removed, on the reference plane and estimates the orientations of the fiber bundles of the X-yarns and the Y-yarns on the basis of a directional distance calculated upon the application.

2. The image analysis apparatus according to claim 1 ,
wherein the binarization unit:
binarizes the three-dimensional image on the basis of a specified threshold value; and
executes closing processing on the image after the binarization.

3. The image analysis apparatus according to claim 2 ,
wherein the overlapping area extraction unit:
executes opening processing on the image after the closing processing;
executes dilation processing on the image after the opening processing; and
extracts the overlapping area by calculating a product set of the image after the dilation processing and the image after the closing processing.

4. The image analysis apparatus according to claim 1 ,
wherein the reference direction determination unit:
applies the directional distance method to the overlapping area and estimates the overlapping direction of each voxel included in the overlapping area on the basis of the directional distance of each voxel calculated upon the application;
extracts a central area by executing erosion processing on the overlapping area; and
averages the overlapping direction of each voxel included in the central area and determines the averaged direction as the reference direction.

5. The image analysis apparatus according to claim 1 ,
wherein the Z-yarn removal unit:
applies the directional distance method on the reference plane and calculates an eigenvector for a second largest eigen value of a directional tensor on the basis of the directional distance of each voxel which is calculated upon the application;
separates the binarized image into an area of the X-yarns or the Y-yarns and an area of the Z-yarns on the basis of the eigenvector; and
removes the Z-yarns from the binarized image by removing voxels included in the Z-yarns area.

6. The image analysis apparatus according to claim 5 ,
wherein upon calculating a directional distance of a voxel of interest, the Z-yarn removal unit proceeds with voxels in a direction indicated by an eigenvector of the voxel of interest and a direction opposite to the direction indicated by the eigenvector; and when the Z-yarn removal unit reaches a background or an angle formed with a direction indicated by an eigenvector of a next voxel to proceed is larger than a specified threshold value, the Z-yarn removal unit stops proceeding and calculates a distance between respective last reached voxels as a conditional directional distance; and
the Z-yarn removal unit separates the binarized image into the area of the X-yarns or the Y-yarns and the area of the Z-yarns on the basis of the conditional directional distance.

7. The image analysis apparatus according to claim 1 ,
wherein the Z-yarn removal unit:
executes dilation processing after removing the Z-yarns from the binarized image; and
calculates a product set of the image after the dilation processing and the image before the dilation processing.

8. The image analysis apparatus according to claim 1 ,
wherein the fiber bundle orientation estimation unit:
applies the directional distance method again to the image, from which the Z-yarns have been removed, on the reference plane and estimates the orientations of the fiber bundles of the X-yarns and the Y-yarns;
extracts voxels in a vicinity of a central part from among the voxels for which the orientations of the fiber bundles are estimated;
executes clustering processing on the voxels in the vicinity of the central part; and
eliminates noise by deleting a cluster in which a minimum voxel belongs.

9. The image analysis apparatus according to claim 1 , comprising a display unit that displays an image showing the orientations of the fiber bundles of the X-yarns and the Y-yarns.

10. An image analysis method for analyzing orientations of fiber bundles of X-yarns and Y-yarns from a three-dimensional image of a woven fabric made of fiber bundles of the X-yarns, the Y-yarns, and Z-yarns,
the image analysis method comprising the following steps executed by a computer:
a first step of binarizing the three-dimensional image;
a second step of extracting an overlapping area, in which the X-yarns and the Y-yarns perpendicularly and three-dimensionally intersect with each other, from the binarized image;
a third step of averaging an overlapping direction of each voxel included in the overlapping area and dete lining the averaged direction as a reference direction;
a fourth step of removing the Z-yarns from the binarized image by applying a directional distance method on a reference plane perpendicular to the reference direction; and
a fifth step of applying the directional distance method again to the image, from which the Z-yarns have been removed, on the reference plane and estimating the orientations of the fiber bundles of the X-yarns and the Y-yarns on the basis of a directional distance calculated upon the application.

11. A non-transitory computer readable storage medium storing a computer program for analyzing orientations of fiber bundles of X-yarns and Y-yarns from a three-dimensional image of a woven fabric made of fiber bundles of the X-yarns, the Y-yarns, and Z-yarns,
the program causing a computer to execute:
a first step of binarizing the three-dimensional image;
a second step of extracting an overlapping area, in which the X-yarns and the Y-yarns perpendicularly and three-dimensionally intersect with each other, from the binarized image;
a third step of averaging an overlapping direction of each voxel included in the overlapping area and determining the averaged direction as a reference direction;
a fourth step of removing the Z-yarns from the binarized image by applying a directional distance method on a reference plane perpendicular to the reference direction; and
a fifth step of applying the directional distance method again to the image, from which the Z-yarns have been removed, on the reference plane and estimating the orientations of the fiber bundles of the X-yarns and the Y-yarns on the basis of a directional distance calculated upon the application.

12. An image analysis apparatus comprising:
a binarization unit that binarizes a three-dimensional image of a woven fabric made of fiber bundles of X-yarns, Y-yarns, and Z-yarns;
an overlapping area extraction unit that extracts an overlapping area, in which the X-yarns and the Y-yarns perpendicularly and three-dimensionally intersect with each other, from the binarized image; and
an overlapping ar

Assignees

• Ihi Corp
• Univ Tokyo

Inventors

• Hishida Hiroyuki
• Inagaki Koichi
• Nakamura Takeshi
• Hasegawa Yu
• Suzuki Hiromasa
• Michikawa Takashi
• Ohtake Yutaka
• Kondo Suguru

Classifications

• G01N23/04

  and forming images of the material · CPC title

• D06H3/08

  by photo-electric or television means (D06H3/12, D06H3/14, D06H3/16 take precedence) · CPC title

• G01N2021/8887

  based on image processing techniques · CPC title

• G01N21/8851

  Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges (G01N21/8806 and G01N21/93 - G01N21/95692 take precedence; optical measurement of dimensions G01B11/00; optical scanning G02B26/10; image transformation G06T3/00; computerised image enhancement G06T5/00; image processing per se for flaw detection G06T7/0002) · CPC title

• G01N23/046

  using tomography, e.g. computed tomography [CT] · CPC title