JP2006329917A - Internal defect inspection method and internal defect inspection device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To further accurately inspect an internal defect based on space discrete data, such as X-ray CT data. <P>SOLUTION: In an internal defect inspection method, the internal defect in an object is inspected, based on the space discrete data describing spatial shape and structure of the object with space discrete elements. The inspection method comprises a process 102 of setting a first detection region of a size, including a plurality of elements in the space discrete data; a process 104 of extracting the first detection region as a defect-detected object; a process 105 of measuring the feature amount of the extracted first detection region as the defect-detected object; a process 106 of setting, in the first detection region, a second detection region obtained, by dividing the first detection region as the defect-detected object; a process 107 of measuring the feature amount of the second detection region; a process 198 of setting a defect determining reference for determining the actual defect and a false defect using the feature amounts of the first and second detection regions; and a process 109 of detecting the defect based on the defect determining reference. <P>COPYRIGHT: (C)2007,JPO&INPIT

Description

  The present invention relates to a technique for inspecting a hollow internal defect such as a cast hole in a casting, for example, with respect to an object, and in particular, spatial discrete such as X-ray CT data obtained by imaging an object with an X-ray CT apparatus. The present invention relates to a technique for inspecting internal defects based on data.

  In industrial products, unintentional cavities may be formed inside the solid portion or may be intentionally included, and the former cavities may become internal defects that degrade the performance of the product. A typical example of such an internal defect is a void in a casting. The cast hole may adversely affect the strength of the casting. For this reason, measures are taken such as adjusting the casting conditions to reduce the occurrence of casting holes, or changing the shape of the casting method of the mold to control the location of the casting holes. These measures are generally determined by repeating trial manufacture and inspection at the design stage and searching for optimum conditions. For this reason, in the field of casting production, it is important to be able to accurately inspect a minute casting hole.

  As a method for inspecting internal defects such as a cast hole, destructive inspection has been used for a long time. In the destructive inspection, the subject is cut and the cross section thereof is observed, or the cut pieces obtained by cutting the subject into dice are weighed. Such destructive inspection requires a lot of labor and accuracy is not so high. In addition, other tests such as a strength test cannot be performed on the specimen subjected to the destructive inspection, and the influence of the internal defects on the performance cannot be accurately evaluated.

  On the other hand, in recent years, nondestructive inspection of internal defects using spatial discrete data has been performed. Spatial discrete data is data that describes the spatial shape and structure of a subject with spatial discrete elements, and one representative example is X-ray CT data. X-ray CT data is 2-dimensional or 3-dimensional bitmap data in a data format in which elements are called “cells (or voxels)” and the cells are spatially discretely arranged. The cell in the X-ray CT data has a value (generally called “cell value” or “CT value”) corresponding to the density at the corresponding portion of the subject. That is, X-ray CT data embodies the density state of the subject. Therefore, by using the X-ray CT data, it is possible to detect internal defects such as casting voids that exist as cavities inside the solid part as the difference between the density of the solid part and the density of the cavity, and nondestructive inspection of internal defects Is possible. Examples of such nondestructive inspection techniques for internal defects based on spatial discrete data are disclosed in, for example, Patent Documents 1 to 6.

JP-A-7-12759 JP 2004-34144 A Japanese Patent Laid-Open No. 2004-12407 Special table 2003-530546 gazette JP 2005-77324 A JP 2005-56087 A

  In order to detect an internal defect from the density state of the object represented by spatial discrete data, for example, X-ray CT data, an appropriate threshold is set, and the defect is detected from the relationship between the CT value and the density value with respect to the threshold. The method is fundamental. However, these methods have problems with artifacts.

  That is, in the X-ray CT data, artifacts generated near the limit of the X-ray transmission capability, artifacts due to partial volume effects resulting from the combination of the X-ray beam width and the shape of the subject, and ring-shaped results resulting from variations in detector characteristics Various artifacts such as artifacts are generated, and these artifacts are accompanied by a phenomenon in which the CT value in the X-ray CT data is locally deviated from the actual one. In X-ray CT data in which there is a local divergence phenomenon of CT values due to artifacts, there is a possibility that portions where CT values are small due to artifacts may be erroneously detected as defects. If there is, the accuracy of the internal defect inspection is lowered.

  The present invention has been made in the background as described above, and about detecting internal defects using spatial discrete data such as X-ray CT data obtained by performing tomographic imaging on a subject. The first object is to provide an internal defect inspection method that can effectively prevent false defect detection by distinguishing pseudo defects caused by artifacts from actual defects, and enables more accurate defect inspection. The second object of the internal defect inspection apparatus for that purpose.

  To achieve the first object, the present invention provides an internal defect inspection method for inspecting internal defects in a subject based on spatial discrete data that describes the spatial shape of the subject with spatial discrete elements. A process for setting a first detection area having a size including the element in the spatial discrete data, a process for extracting a first detection area as a defect detection target from the spatial discrete data, and a first of the extracted defect detection target. A process of measuring a feature amount of one detection area, a second detection area having a size including a smaller number of the elements than the first detection area by subdividing the first detection area of the defect detection target A process for setting the first detection area, a process for measuring the feature quantity of the second detection area, an actual defect using the feature quantity of the first detection area and the feature quantity of the second detection area And pseudo defects It is characterized in that it comprises because of process of setting the defect determination criteria, and the set defect determination criteria for detecting a distinguished while defects actual defects and pseudo defects based on the processing.

  According to the present invention, in the internal defect inspection method as described above, as the feature amount used for setting the defect determination reference, the first value regarding the value of each element in each of the first and second detection regions is used. Standard deviations in the first and second detection areas are used.

  In the present invention, in the internal defect inspection method as described above, the defect determination criterion is set based on a difference between the standard deviation in the first detection region and the standard deviation in the second detection region. Like to do.

  In the present invention, a computer program in which a procedure for executing the internal defect inspection method as described above is executed is interposed.

  In the present invention, for the second object, in the internal defect inspection apparatus for inspecting an internal defect in the subject based on spatial discrete data describing a spatial shape of the subject with spatial discrete elements. A detection area setting means for setting a first detection area having a size including a plurality of the elements in the spatial discrete data; and subdividing the first detection area to reduce the number of the elements to be smaller than the first detection area Detection region setting means for setting a second detection region having a size including the first detection region, a feature amount measurement unit for measuring feature amounts of the first and second detection regions, the first detection region Defect determination reference setting means for setting a defect determination reference for determining an actual defect and a pseudo defect using the feature amount of the detection region and the feature amount of the second detection region, and the set defect determination reference Based on actual That it comprises a defect detecting means for detecting a defect while distinguishing a pseudo defect and Recessed are characterized.

  According to the present invention, in the internal defect inspection apparatus as described above, as the feature amount used for setting the defect determination reference, the first of the values of the elements in the first and second detection regions is used. Standard deviations in the first and second detection areas are used.

  In the present invention, for the internal defect inspection apparatus as described above, the defect determination reference is set based on a difference between the standard deviation in the first detection region and the standard deviation in the second detection region. Like to do.

  In the present invention, a first detection area having a predetermined size is set as spatial discrete data, and a second detection area obtained by subdividing the first detection area is set as the first detection area. Using the feature values of each detection area, set a defect discrimination standard for distinguishing between actual defects and pseudo defects, and distinguish actual defects from pseudo defects based on these defect discrimination standards. Defects can be detected. For this reason, according to the present invention, it is possible to effectively prevent false detection by accurately discriminating between a pseudo defect and an actual defect caused by, for example, an artifact in X-ray CT data, and the higher accuracy. Defect inspection is possible.

  Hereinafter, preferred embodiments for carrying out the present invention will be described. Each of the embodiments described below is an example in which a specimen is a cast, and a casting defect such as a casting hole in the casting is used as an internal defect and inspected using X-ray CT data by an X-ray CT apparatus.

  FIG. 1 schematically shows the configuration of the processing function of the internal defect inspection apparatus according to the first embodiment. The internal defect inspection apparatus of the present embodiment is configured by mounting internal defect inspection means 1 configured as a computer program as a software element on a hardware element such as a computer (not shown). The internal defect inspection unit 1 includes a spatial discrete data input unit 2, a detection area setting unit 3, a density distribution data creation unit 4, a feature amount measurement unit 5, a defect discrimination reference setting unit 6, a defect detection unit 7, and a detection result output unit 8. Is included.

  The spatial discrete data input means 2 inputs spatial discrete data, specifically X-ray CT data. The density distribution data creation means 4 creates density distribution data for the input X-ray CT data. The detection area setting means 3 sets a first detection area for the X-ray CT data, and sets a second detection area obtained by subdividing the first detection area. The feature amount measuring means 5 measures the feature amount for each of the first and second detection regions. The defect determination reference setting unit 6 obtains and sets a reference for determining an actual defect and a pseudo defect due to an artifact from the feature amount measured by the feature amount measurement unit 5. The defect detection unit 7 detects a defect by making a determination based on the defect determination criterion set by the defect determination criterion setting unit 6. The detection result output means 8 outputs a defect detection result.

  Below, the content of the inspection process in an internal defect inspection apparatus is demonstrated. FIG. 2 shows the flow of the inspection process. The inspection processing includes data input processing 101, first detection region setting processing 102, density distribution data creation processing 103, target detection region extraction processing 104, first detection region feature amount measurement processing 105, and second detection region setting processing. 106, a second detection area feature amount measurement process 107, a defect discrimination reference setting process 108, a defect detection process 109, and a defect detection result output process 110.

  In the data input process 101, spatial discrete data, specifically X-ray CT data is input by the spatial discrete data input means 2.

  In the first detection area setting process 102, the first detection area is set by the detection area setting means 3 for the input X-ray CT data. The first detection area is a unit for proceeding the defect detection process, and is set as an area having a predetermined size (including a predetermined number of cells). The predetermined size is a trade-off between the condition of increasing the size, which reduces the burden of density distribution data creation processing as much as possible, and the condition of decreasing the size, so as to obtain an appropriate density distribution state. It is determined from the off relationship and is usually about several hundred cells in size.

  In the density distribution data creation processing 103, density distribution data is created by the density distribution data creation means 4 based on the first detection area for the X-ray CT data in which the first detection area is set. A well-known method can be used for creating density distribution data by the density distribution data creating means 4.

  In the target detection area extraction process 104, one of the first detection areas in which the presence of a defect is estimated from the density distribution data is extracted as a process target from the process 105 to the process 109.

  In the first detection area feature amount measurement process 105, the feature amount measurement unit 5 measures the feature amount of the extracted first detection area. There are various feature quantities to be measured. The feature quantity necessary for the purpose of detecting a defect while distinguishing an actual defect from a pseudo defect due to an artifact is a CT value (or a value in the first detection region). Density) mean and standard deviation. These are obtained by appropriately sampling the CT value of each cell included in the region.

  In the second detection area setting process 106, the detection area setting means 3 sets a second detection area for the first detection area. The second detection area is a unit for advancing defect detection while distinguishing between actual defects and pseudo defects due to artifacts, and is set by subdividing the first detection area as an area having a predetermined size. The The predetermined size can appropriately determine the standard deviation of the CT value in the region, and the condition in the direction of increasing the size, and the influence of the artifact appropriately appears in the standard deviation of the CT value in the region. The size is determined from the trade-off relationship of conditions in the direction of decreasing the size, and is usually about several to several tens of cells. FIG. 3 schematically shows the relationship between the first detection region and the second detection region. The first detection region E is set by dividing the X-ray CT data D by a predetermined size so as to fill the X-ray CT data D. The second detection area SE is set by subdividing the first detection area Ex extracted by the target detection area extraction processing 104.

  In the second detection area feature quantity measurement processing 107, the feature quantity measurement unit 5 measures the feature quantity of the second detection area. The feature quantity to be measured is the average value and standard deviation of the CT values in the second detection area, as in the case of the first detection area.

In the defect determination reference setting process 108, a reference for determining an actual defect and a pseudo defect due to an artifact is obtained by the defect determination reference setting means 6 and set. FIG. 4 shows an image of the contents of processing by the defect discrimination reference setting means 6. In the defect discrimination reference setting process, the standard deviation σ obtained for the first detection area E in the first detection area feature quantity measurement process 105 and the second detection area SE in the second detection area feature quantity measurement process 107 are displayed. _A threshold S _ij is individually set as a defect determination criterion for each of the second detection areas SE _ij using the standard deviation σ _ij obtained for each _ij . Specifically, a threshold value S _ij is obtained as S _ij = σ−σ _ij , and this is set in each of the second detection areas SE _ij . Here, the second detection areas SE are arranged in a two-dimensional direction and are actually the second detection areas SE _ij , but the second detection areas SE _i ( Second detection areas SE ₁ , SE ₂ ,... SE ₇ ) and standard deviations σ _i (σ ₁ , σ ₂ ,... Σ ₇ ) are shown.

By setting the threshold value S _ij as S _ij = σ−σ _ij for each second detection region SE _{ij obtained} by subdividing the first detection region E in this way, the actual defect and the artificial defect due to the artifact are detected. Can be accurately distinguished, and false detection of artifacts due to artifacts can be effectively prevented. This will be described in comparison with the conventional method.

  As in the example of FIG. 4, the conventional detection method is used for the first detection region E in which casting defects P are scattered and the line-shaped artifacts F are generated along the irradiation direction of the X-ray R. In the conventional method, since the presence or absence of a defect is determined based on the average CT value and the standard deviation σ for the first detection region E, a pseudo defect caused by the artifact F is actually detected. It cannot be distinguished from a defect, and it is also detected as a defect.

In contrast, in the present invention, the presence of sigma _ij in the second detection region SE _ij each casting defects P and artifacts F, sigma shown on the lower side of FIG. 4 for line A-A 'in FIG. 4 _i ((σ ₁ , σ ₂ ,..., σ ₇ ), and the fluctuation is particularly large in the second detection area SE in which the artifact F is generated. _The threshold S _ij individually set as S _ij = σ−σ _ij for each _ij is smaller in the second detection area SE _ij affected by the artifact F than in the other second detection areas SE _ij. For this reason, even if the average CT value is small due to the influence of the artifact F, the threshold S _ij is small, so that the decrease in the average CT value in the second detection region SE _ij is not sufficient. It can be determined that it is not caused by a defect, and it is possible to avoid detecting a pseudo defect due to the artifact F as an actual defect.

Here, it is also a preferable mode that the threshold value S _ij is calculated as S _ij = k ₁ σ−k ₂ σ _ij . That is, X-ray CT data cannot avoid a certain amount of noise, and σ and σ _ij are affected according to the level of the noise. Therefore, the defect determination can be performed with higher accuracy by multiplying the measurement σ or σ _ij by the constants k ₁ and k ₂ corresponding to the actual noise in the X-ray CT data so that the influence of the noise can be canceled. . Further, it is preferable that the first detection area E in which no artifact is generated is configured such that the processes from the second detection area setting process 106 to the defect detection process 108 can be omitted. That is, for the first detection region E in which no artifact is generated, the artifact may be detected even if the defect is detected based on the average CT value and the standard deviation for the first detection region E, as in the conventional method. Therefore, the processing efficiency can be improved by omitting the processing for defect determination.

  In the defect detection process 109, a defect is detected while distinguishing an actual defect and a pseudo defect due to an artifact based on the defect determination standard set in the defect determination standard setting process 108. However, when the processes 106 to 108 are omitted for the first detection area where no artifact is generated, the defect is detected based on the average CT value and the standard deviation for the first detection area. .

  The processes from the object detection area extraction process 104 to the defect detection process 109 described above are repeated for all the first detection areas to be subjected to defect detection, and when the process is completed, the defect detection result output process 110 is performed. The defect detection result is output in a predetermined format. FIG. 5 shows an example of an output display of the result of defect detection for each of the conventional method and the method according to the present invention for X-ray CT data having defects and artifacts. FIG. 5A is an X-ray CT image having defects and artifacts, FIG. 5B is an output display of defect detection results by a conventional method, and FIG. It is an output display of the defect detection result by invention. The black part in the center part of (a) of FIG. 5 is a casting defect, and the lines that appear to be black and white stripe patterns above and below are artifacts. In this example, artifacts are generated over the entire surface. In FIG. 5B and FIG. 5C, the portion detected as a defect is displayed in color (gray in the figure). In FIG. 5 (b), the artifact portion is also detected as a defect, whereas in FIG. 5 (c), only the actual casting defect portion is detected as a defect.

  FIG. 6 shows an example in which the result of detecting a defect in the three-dimensional X-ray CT data is displayed as a rendering image R. The rendered image R displays the external shape of the subject and the internal defects represented by dark black. As for the display of defects, it is possible to adopt a form in which display by color or enlarged display is performed in accordance with the feature amount such as the size, shape, or location. It is also possible to display only defects in the area selected by the user, or display only defects larger than the size selected by the user. For such selective display, an operation input field (not shown) for the selection is provided on the display screen.

  Although the case where X-ray CT data is used has been described above, the present invention can also be applied to the case where data such as MRI or PET is used as spatial discrete data.

  INDUSTRIAL APPLICABILITY The present invention makes it possible to perform an internal defect inspection with spatially discrete data such as X-ray CT data with higher accuracy, and can be widely applied, for example, to a field such as a casting hole inspection of a casting. .

It is a figure which shows typically the structure in the processing function of the internal defect inspection apparatus by one Embodiment. It is a figure which shows the flow of a process in the internal defect inspection apparatus of FIG. It is a figure which shows typically the relationship between a 1st detection area and a 2nd detection area. It is a figure showing the contents of the defect discrimination standard setting processing as an image. It is a figure which shows the example of an output display of the defect detection result about the X-ray CT data with a defect and an artifact. It is a figure which shows the example in the case of displaying the defect detection result about three-dimensional X-ray CT data as a rendering image.

Explanation of symbols

DESCRIPTION OF SYMBOLS 3 Detection area | region setting means 5 Feature-value measurement means 6 Defect discrimination reference | standard setting means 7 Defect detection means E 1st detection area SE 2nd detection area P Defect

Claims (7)

In an internal defect inspection method for inspecting an internal defect in the subject based on spatial discrete data describing a spatial shape of the subject with spatial discrete elements,
A process of setting a first detection area having a size including a plurality of the elements in the spatial discrete data, a process of extracting a first detection area as a defect detection target from the spatial discrete data, and an extracted defect detection target A process for measuring the feature amount of the first detection area, a second detection area having a size that includes a smaller number of the elements than the first detection area by subdividing the first detection area of the defect detection target Is set in the first detection area, the process for measuring the feature quantity of the second detection area, the feature quantity of the first detection area and the feature quantity of the second detection area are used to actually Including a process for setting a defect determination standard for determining a defect and a pseudo defect, and a process for detecting a defect while distinguishing an actual defect from a pseudo defect based on the set defect determination standard. Internal defect inspection method   The feature amount used for setting the defect determination standard is a standard deviation in each of the first and second detection regions with respect to a value of each of the elements in the first and second detection regions. The internal defect inspection method according to 1.   The internal defect inspection according to claim 2, wherein the defect determination criterion is set based on a difference between the standard deviation in the first detection region and the standard deviation in the second detection region. Method.   The computer program in which the procedure for performing the internal defect inspection method of any one of Claims 1-3 is described. In an internal defect inspection apparatus that inspects an internal defect in the subject based on spatial discrete data that describes the spatial shape of the subject with spatial discrete elements,
A detection area setting means for setting a first detection area having a size including a plurality of the elements in the spatial discrete data; and subdividing the first detection area into a smaller number of the elements than the first detection area. A detection area setting means for setting a second detection area having a size to be included in the first detection area; a feature quantity measurement means for measuring a feature quantity of each of the first and second detection areas; and the first detection. A defect determination reference setting means for setting a defect determination reference for determining an actual defect and a pseudo defect using the feature amount of the region and the feature amount of the second detection region, and the set defect determination reference An internal defect inspection apparatus comprising defect detection means for detecting a defect while distinguishing an actual defect from a pseudo defect based on the defect.   The feature amount used for setting the defect determination standard is a standard deviation in each of the first and second detection regions with respect to a value of each of the elements in the first and second detection regions. The internal defect inspection apparatus according to 1. The internal defect inspection according to claim 6, wherein the defect determination criterion is set based on a difference between the standard deviation in the first detection region and the standard deviation in the second detection region. apparatus.

Images