Radiographic method, radiographic apparatus, radiographic system, and radiographic program

Abstract

A radiographic apparatus acquires a first image and a second image, receives a selection input of an area of interest in a region corresponding to an article in the first image or the second image, determines a respective first pixel value of a plurality of first pixels and a respective second pixel value of a plurality of second pixels corresponding to the plurality of first pixels, calculates a thickness correction function by approximating a relationship between the first pixel value and the second pixel value corresponding to the first pixel value, respectively calculates a plurality of representative data as a combination of a first representative value and a second representative value based on the respective first pixel value of the plurality of first pixels and the respective second pixel value of the plurality of second pixels corresponding to the plurality of first pixels, and calculates an evaluation coefficient based on a correlation of the thickness correction function and the calculated plurality of representative data.

Description

Technical Field

[0001] This disclosure relates to radiation examination methods, radiation examination apparatus, radiation examination systems, and radiation examination procedures. Background Technology

[0002] Techniques for examining samples from X-ray images using energy subtraction have been employed for some time. For example, Patent Document 1 describes an X-ray imaging method in which, in either of two images of a sample taken using energy subtraction, a region corresponding to the component to be removed is selected, and corrections are performed to eliminate the influence of the sample thickness, thereby effectively removing the component to be removed from the captured image.

[0003] Existing technical documents

[0004] Patent documents

[0005] Patent Document 1: Japanese Patent Application Publication No. 2000-121579 Summary of the Invention

[0006] The technical problem that the invention aims to solve

[0007] The method described in Patent Document 1 is based on the premise that the region corresponding to the component to be removed, i.e., the region of interest, can be appropriately selected in the image. However, depending on the image, there are cases where it is difficult to identify the boundaries between the multiple components constituting the article. In such cases, depending on the region selection state, there are situations where regions other than the region corresponding to the substance to be removed are selectively input into the image, and the influence of the substance's thickness cannot be properly eliminated, thus failing to properly remove the image of the substance from the image.

[0008] Therefore, this disclosure was made in view of this technical problem, with the technical problem being to provide a radiological examination method, radiological examination apparatus, radiological examination procedure, and radiological examination system capable of evaluating whether the area of ​​interest has been appropriately selected.

[0009] Technical means for solving technical problems

[0010] One aspect of the radiation inspection method of this disclosure includes: a first step of acquiring a first image of an article photographed under conditions of irradiation with radiation having a first energy distribution, and a second image of the article photographed under conditions of irradiation with a second energy distribution different from the first energy distribution; a second step of accepting a selection input of a region of interest, said region of interest being a region in the first image or the second image corresponding to the article that corresponds to a substance for eliminating the effect of thickness; and a third step of determining the respective first pixel values ​​of a plurality of first pixels in the region corresponding to the region of interest in the first image, and the respective second pixel values ​​of a plurality of second pixels corresponding to the plurality of first pixels in the second image. The thickness correction function is calculated to eliminate the influence of the material's thickness by approximating the relationship between the first pixel value and the corresponding second pixel value. The thickness correction function is a function representing the relationship between the first pixel value and the second pixel value. The fourth step involves calculating multiple representative data based on the respective first pixel values ​​of multiple first pixels and the respective second pixel values ​​of multiple second pixels corresponding to the multiple first pixels. Each of the multiple representative data is a combination of a first representative value representing the pixel value of the first image and a second representative value representing the pixel value of the second image. The fifth step involves calculating an evaluation coefficient based on the correlation between the thickness correction function and the calculated multiple representative data.

[0011] Alternatively, other aspects of the radiation inspection apparatus of this disclosure include at least one processor that acquires a first image of an article photographed under conditions of irradiation with radiation having a first energy distribution, and a second image of the article photographed under conditions of irradiation with radiation having a second energy distribution different from the first energy distribution; accepts selection input of a region of interest, which is a region in the first or second image corresponding to the article that corresponds to a substance for eliminating the effect of thickness; determines the respective first pixel values ​​of a plurality of first pixels in the region corresponding to the region of interest in the first image, and a plurality of second pixels in the second image corresponding to the plurality of first pixels. The thickness correction function, which eliminates the influence of the material's thickness, is calculated by approximating the relationship between the first pixel value and the corresponding second pixel value. The thickness correction function is a function representing the relationship between the first pixel value and the second pixel value. Multiple representative data are calculated based on the respective first pixel values ​​of multiple first pixels and the respective second pixel values ​​of multiple second pixels corresponding to multiple first pixels. Each of the multiple representative data is a combination of the first representative value of the pixel value representing the first image and the second representative value of the pixel value representing the second image. An evaluation coefficient based on the correlation between the thickness correction function and the calculated multiple representative data is calculated.

[0012] Alternatively, other aspects of the radiation inspection system disclosed herein may include: the aforementioned radiation inspection apparatus; a radiation source that irradiates an article with radiation having a first energy distribution and radiation having a second energy distribution; and a detector that detects radiation transmitted through the article from the radiation source and having the first energy distribution, and radiation transmitted through the article from the radiation source and having the second energy distribution.

[0013] Alternatively, the radiation inspection procedure of other aspects of this disclosure enables the computer to function as follows: Step 1, acquiring a first image of an article photographed under conditions of irradiation with radiation having a first energy distribution, and a second image of the article photographed under conditions of irradiation with radiation having a second energy distribution different from the first energy distribution; Step 2, accepting input for selection of a region of interest, said region of interest being the region in the first or second image corresponding to the article that corresponds to the material for eliminating the effects of thickness; Step 3, determining the respective first pixel values ​​of a plurality of first pixels in the region corresponding to the region of interest in the first image and a plurality of second pixels corresponding to the plurality of first pixels in the second image. Each of the second pixel values ​​is used to calculate a thickness correction function to eliminate the influence of the material's thickness by approximating the relationship between the first pixel value and the second pixel value corresponding to the first pixel value. The thickness correction function is a function representing the relationship between the first pixel value and the second pixel value. In the fourth step, multiple representative data are calculated based on the respective first pixel values ​​of multiple first pixels and the respective second pixel values ​​of multiple second pixels corresponding to multiple first pixels. Each of the multiple representative data is a combination of a first representative value representing the pixel value of the first image and a second representative value representing the pixel value of the second image. In the fifth step, an evaluation coefficient is calculated based on the correlation between the thickness correction function and the calculated multiple representative data.

[0014] According to the method, apparatus, system, and procedure described above, a selection input of a region of interest is received in a first image or a second image. The first pixel values ​​of multiple first pixels in the region corresponding to the region of interest in the first image and the second pixel values ​​of multiple second pixels corresponding to the multiple first pixels in the second image are determined. Then, a thickness correction function is calculated by approximating the relationship between the first pixel values ​​and the corresponding second pixel values, and multiple representative data are calculated, each of which is a combination of a first representative value representing the first pixel value and a second representative value representing the second pixel value. Then, an evaluation coefficient is calculated based on the correlation between the thickness correction function and the multiple representative data. Therefore, based on the evaluation coefficient, it is possible to evaluate whether the region of interest corresponds to a single substance. Thus, according to the method, apparatus, system, and procedure described above, it is possible to evaluate whether the region of interest has been appropriately selected.

[0015] Invention Effects

[0016] According to this disclosure, it is possible to evaluate whether the area of ​​interest has been appropriately selected. Attached Figure Description

[0017] Figure 1 This is a general configuration diagram of the radiation examination system according to the first embodiment.

[0018] Figure 2 This is a perspective view of the radiation examination system according to the first embodiment.

[0019] Figure 3 This is a block diagram illustrating the functional configuration of the radiation examination apparatus according to the first embodiment.

[0020] Figure 4 It means to include Figure 1 A block diagram of the hardware configuration of the computer system for a radiographic examination device.

[0021] Figure 5 It means by Figure 3 The selection process for the receiving department includes an example of the input area of ​​interest.

[0022] Figure 6 It means by Figure 3 The selection process for the receiving department includes an example of the input area of ​​interest.

[0023] Figure 7 It indicates that it displays the same as Figure 5 An example of a graph showing the coordinates of sample points and approximate curves corresponding to the region of interest.

[0024] Figure 8 It indicates that it displays the same as Figure 6 An example of a graph showing the coordinates of sample points and approximate curves corresponding to the region of interest.

[0025] Figure 9 It indicates that it displays the same as Figure 5 This is an example of a coordinate graph showing the representative data and approximate curve corresponding to the region of interest.

[0026] Figure 10 It indicates that it displays the same as Figure 6 This is an example of a coordinate graph showing the representative data and approximate curve corresponding to the region of interest.

[0027] Figure 11 It means and Figure 5 An example of a difference image corresponding to the region of interest.

[0028] Figure 12 It means and Figure 6 An example of a difference image corresponding to the region of interest.

[0029] Figure 13 This is a flowchart illustrating the processing steps of the radiation examination method according to the first embodiment.

[0030] Figure 14 This is a diagram illustrating an example of the image information output in the second embodiment.

[0031] Figure 15 This is a flowchart illustrating the processing steps of the radiation examination method according to the second embodiment.

[0032] Figure 16 This is a flowchart illustrating the processing steps of the radiation examination method according to the third embodiment.

[0033] Figure 17 This is a diagram showing an example of a first image obtained by the radiation examination apparatus of the first embodiment.

[0034] Figure 18 It is a coordinate graph of an approximate curve calculated by using a radiographic examination device.

[0035] Figure 19 This is a diagram showing the subtraction image generated by the X-ray inspection device and the foreign object detection result image obtained by thresholding the subtraction image.

[0036] Figure 20 It is a coordinate graph of the approximate curve calculated by using a second approximation with a radiographic examination device.

[0037] Figure 21 This is a diagram showing the subtraction image generated by the X-ray inspection device and the foreign object detection result image obtained by thresholding the subtraction image.

[0038] Figure 22 This is a diagram showing an example of a first image obtained by the radiation examination apparatus of the first embodiment.

[0039] Figure 23 It is a coordinate graph of the approximate curve calculated by using a second approximation with a radiographic examination device.

[0040] Figure 24 This is a diagram showing the subtraction image generated by the X-ray inspection device and the foreign object detection result image obtained by thresholding the subtraction image.

[0041] Symbol Explanation

[0042] 1…Radiation inspection system, 2…Irradiator (radiation source), 3…Image acquisition device (detector), 5…Radiation inspection device, 20…Computer system, 101…CPU (processor), C1, C2…Approximate curves, D1, D2…Sample points, E…Non-corresponding area, I…Image information (error information), P1…First image, second image, R…Region of interest, S…Item, F1, F2…Representative data. Detailed Implementation

[0043] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the accompanying drawings. Furthermore, identical or corresponding parts in each drawing will be labeled with the same symbols, and repeated descriptions will be omitted.

[0044] [First Implementation]

[0045] [Composition of a Radiation Examination System]

[0046] like Figure 1 As shown, the radiation inspection system 1 includes an irradiator (radiation source) 2, an image acquisition device (detector) 3, and a radiation inspection device 5. For example... Figure 2 As shown, the radiation inspection system 1 is an apparatus that performs the following operations: irradiating an article S with multiple radiations having energy distributions along the irradiation direction Z, acquiring multiple images of the article S under the condition of being irradiated with radiations having various energy distributions, and performing foreign object inspection, etc., based on these multiple images. The radiation inspection system 1 performs foreign object inspection, weight inspection, and inspection of the article S. Examples of its applications include food inspection, baggage inspection, substrate inspection, battery inspection, and material inspection. In this embodiment, the radiation inspection system 1 irradiates the article S with X-rays from an X-ray source. The article S is conveyed in the conveying direction Y at a predetermined conveying speed while being placed on the conveyor belt B1 of the belt conveyor device B. Examples of articles S include various items such as meat, seafood, agricultural products, desserts and other food products, rubber products such as tires, resin products, metal products, mineral resources and other materials, waste, and electronic components, electronic substrates, etc.

[0047] The irradiator 2 is positioned above the conveyor belt section B1 at a predetermined interval. The irradiator 2 functions as an X-ray source, irradiating the article S with X-rays in the irradiation direction Z. The irradiator 2 irradiates the article S with radiation having a first energy distribution and radiation having a second energy distribution different from the first energy distribution. The first energy distribution is, for example, an energy band lower than the second energy distribution. The irradiator 2 is a point source, irradiating the article S in a manner that allows X-rays to diffuse within a predetermined angular range in the inspection direction X. The inspection direction X is orthogonal to the irradiation direction Z and the transport direction Y. The irradiator 2 is configured to irradiate the entire article S with X-rays in the inspection direction X (the width direction of the article S). Additionally, the irradiator 2 is used to irradiate a segment of the article S in the transport direction Y that is smaller than the overall length of the article S. The irradiator 2 is configured to irradiate the entire article S with X-rays in the transport direction Y by transporting the article S using the belt conveyor B.

[0048] The image acquisition device 3 is positioned below the portion of the conveyor belt section B1 carrying the article S and the irradiator 2 in the irradiation direction Z. The image acquisition device 3 detects radiation that has a first energy distribution and is transmitted through the article S from a radiation source, and radiation that has a second energy distribution and is transmitted through the article S from a radiation source. Then, the image acquisition device 3 acquires an image of the article S obtained by photographing it while it is irradiated with radiation having a predetermined energy distribution. The image acquisition device 3 includes a first image acquisition unit 31, a second image acquisition unit 32, and a control unit 33.

[0049] The first image acquisition unit 31 acquires a first image by photographing an object S under conditions where it has been irradiated with radiation having a first energy distribution. The first image acquisition unit 31 includes a first detection unit 311 and a first image correction unit 312.

[0050] The first detection unit 311 is located upstream in the X-ray irradiation direction Z. The first detection unit 311 detects the range of a first energy distribution in the X-rays irradiated from the irradiator 2 and transmitted through the article S, thereby generating image data. Here, the method for generating image data using the first detection unit 311 will be described. The first detection unit 311 includes a scintillator layer (not shown) corresponding to the first energy distribution and a line sensor (not shown) corresponding to the first energy distribution. The scintillator layer corresponding to the first energy distribution extends along the inspection direction X, converting the image of the X-rays having the first energy distribution into a light image. The line sensor corresponding to the first energy distribution has multiple pixels arranged along the inspection direction X, generating image data obtained from the light image converted by the scintillator layer. The image data acquired using the line sensor is composed of a collection of brightness data acquired by the line sensor for each pixel.

[0051] The first image correction unit 312 amplifies and corrects the brightness data generated by the first detection unit 311 for each pixel, thereby acquiring amplified and corrected image data. The first image correction unit 312 includes an amplifier 31a, an A / D converter 31b, a correction circuit 31c, and an output interface 31d. The amplifier 31a amplifies the brightness data of the X-ray image having a first energy distribution. The A / D converter 31b performs A / D conversion on the brightness data of the X-ray image having the first energy distribution amplified by the amplifier 31a. The correction circuit 31c performs a prescribed correction process on the brightness data converted by the A / D converter 31b. The output interface 31d outputs the image data corrected by the correction circuit 31c as a first image to the outside.

[0052] The second image acquisition unit 32 acquires a second image of an object S obtained by photographing it while it is irradiated with radiation having a second energy distribution. The second image acquisition unit 32 includes a second detection unit 321 and a second image correction unit 322.

[0053] The second detection unit 321 is positioned downstream of the first detection unit 311 in the X-ray irradiation direction Z. The second detection unit 321 detects the range of a second energy distribution in the X-rays irradiated from the irradiator 2 and transmitted through the article S and the first detection unit 311, and generates image data. Here, the method for generating image data using the second detection unit 321 will be described. The second detection unit 321 includes a scintillator layer (not shown) corresponding to the second energy distribution and a line sensor (not shown) corresponding to the second energy distribution. The scintillator layer corresponding to the second energy distribution extends along the inspection direction X, converting the image of the X-rays having the second energy distribution into an optical image. The line sensor corresponding to the second energy distribution has multiple pixels arranged along the inspection direction X, acquiring image data obtained from the optical image converted by the scintillator layer. The image data acquired by the line sensor is a collection of brightness data acquired by the line sensor for each pixel. Furthermore, the first detection unit 311 and the second detection unit 321 can also be line sensors comprising the first detection unit 311 and the second detection unit 321 in a single sensor configuration. The line sensor can be a multi-line sensor with multiple pixels in the transport direction, a TDI (Time Delay Integration) scanning X-ray camera, or a two-dimensional X-ray camera. Alternatively, a sensor using a direct conversion method without a scintillator, or a camera using an optical lens coupling a scintillator as a lens, can also be used. A structure with multiple sensors corresponding to multiple radiation sources is also possible. Additionally, the range of the first energy distribution detected by the first detection unit 311 and the range of the second energy distribution detected by the second detection unit 321 may partially overlap. Furthermore, the second detection unit 321 has been described as being positioned downstream of the first detection unit 311 in the X-ray irradiation direction Z, but the configuration of the first detection unit 311 and the second detection unit 321 is not limited to this embodiment. For example, the first detection unit 311 and the second detection unit 321 may be arranged side by side (in such a way that they are located upstream and downstream in the conveying direction Y).

[0054] The second image correction unit 322 amplifies and corrects the brightness data generated by the second detection unit 321 for each pixel, acquiring amplified and corrected image data. The second image correction unit 322 includes an amplifier 32a, an A / D converter 32b, a correction circuit 32c, and an output interface 32d. The amplifier 32a amplifies the brightness data of the X-ray image with a second energy distribution. The A / D converter 32b performs A / D conversion on the brightness data of the X-ray image with the second energy distribution amplified by the amplifier 32a. The correction circuit 32c performs a prescribed correction process on the brightness data converted by the A / D converter 32b. The output interface 32d outputs the brightness data corrected by the correction circuit 32c as a second image to the outside.

[0055] The control unit 33 controls the detection timing of X-rays in the first detection unit 311 and the second detection unit 321. Specifically, the control unit 33 controls the detection timing of the first detection unit 311 and the second detection unit 321 so that X-rays that have transmitted through a segment of the article S can be detected by both the first image acquisition unit 31 and the second image acquisition unit 32. By controlling the detection timing using the control unit 33, the image deviation generated in the first image and the second image can be reduced in the subtraction processing described later.

[0056] Furthermore, the control unit 33 uses a known calibration component to control the process so that each pixel of the first detection unit 311 corresponds to each pixel of the second detection unit 321 in the inspection direction X of the article S. Because the irradiator 2 is a point light source and the X-rays diffuse radially, the positions of the corresponding pixels of the first detection unit 311 and the second detection unit 321 deviate further from each other as they approach the two ends in the inspection direction X. Here, through the calibration control of the control unit 33, the positional deviation between the pixels of the first detection unit 311 and the second detection unit 321 is corrected, reducing image deviations in the first and second images. Through the above processing, the first and second images are acquired such that each pixel of the first image corresponds to each pixel of the second image.

[0057] The radiation inspection apparatus 5 is a data processing device that generates a differential image of the region corresponding to the material to be removed. The radiation inspection apparatus 5 can also be a computing device such as a personal computer, microcomputer, cloud server, or smart device. The radiation inspection apparatus 5 is connected to the image acquisition apparatus 3 in a manner that enables data communication between them.

[0058] In either the first or second image, the radiation inspection device 5 accepts a selection input for the region of interest corresponding to the material to be removed, and performs correction processing to eliminate the influence of the thickness of the article S. Then, the radiation inspection device 5 performs subtraction processing to obtain the difference between the pixel values ​​of the logarithmically transformed first image and the pixel values ​​of the logarithmically transformed second image, thereby generating a differential image with the region corresponding to the material to be removed removed. The radiation inspection device 5 then outputs the differential image to a display (not shown) included in the output device 105 described later.

[0059] Furthermore, before generating the differential image, the X-ray inspection apparatus 5 performs a process to evaluate whether the region of interest has been appropriately selected. Based on the acquired first or second image, there may be cases where it is difficult to identify the boundaries between the multiple components constituting the article S. In such cases, depending on the region selection state, regions other than those corresponding to the removed substance may be selectively input into the first or second image, resulting in the inability to properly eliminate the influence of the substance's thickness, and thus the inability to properly remove the image of the substance from the differential image. Therefore, by performing a process to evaluate whether the region of interest has been appropriately selected before generating the differential image in the X-ray inspection apparatus 5, the image of the substance to be removed can be properly removed from the differential image. Furthermore, in the following description, "eliminating the influence of the thickness of the removed substance in the image" will sometimes be abbreviated as "eliminating the thickness of the substance."

[0060] [Composition of a Radiation Examination Device]

[0061] Next, the detailed structure of the radiation inspection device 5 will be explained. For example... Figure 3 As shown, the radiographic examination apparatus 5 includes, as functional components, an acquisition unit 51, a selection and acceptance unit 52, a thickness correction function calculation unit 53, a representative data calculation unit 54, an evaluation coefficient calculation unit 55, a notification unit 56, a differential image generation unit 57, and a storage unit 58. Figure 4As shown, the computer system 20 including the radiation inspection device 5 physically includes a CPU (Central Processing Unit) 101 as a processor, RAM (Random Access Memory) 102 as a storage medium, ROM (Read Only Memory) 103, a communication module 104, an output device 105, and an input device 106. The functional units of the radiation inspection device 5 are implemented by loading the radiation inspection program of this embodiment into the CPU 101, RAM 102, and other hardware. Under the control of the CPU 101, the communication module 104, output device 105, and input device 106 operate, and data is read from and written to the RAM 102 and read from the ROM 103. That is, the radiation inspection procedure of this embodiment enables the computer system 20 to function as an acquisition unit 51, a selection and acceptance unit 52, a thickness correction function calculation unit 53, a representative data calculation unit 54, an evaluation coefficient calculation unit 55, a notification unit 56, a differential image generation unit 57, and a storage unit 58. Furthermore, the CPU can be a single piece of hardware, or it can be installed in a programmable logic circuit such as an FPGA, like a soft-core processor. Regarding RAM and ROM, they can be single pieces of hardware, or they can be built into a programmable logic circuit such as an FPGA.

[0062] The following details the functions of each part of the radiation inspection device 5. Figure 5 and Figure 6 The specific examples of the images shown (image 1 and image 2) P1 will be explained.

[0063] The acquisition unit 51 acquires images, namely the first image and the second image, of the X-rays transmitted through the entire object S from the image acquisition device 3. Figure 5 and Figure 6 The image P1 shown is an example of a first image or a second image acquired by the first image acquisition unit 31. Image P1 shows an article S irradiated with radiation having a first energy distribution or a second energy distribution. In this specific example, the article S is a general-purpose connector having a housing S1 and multiple terminals S2. The housing S1 is made of resin and contains multiple cavities. Each terminal S2 is disposed in a cavity of the housing S1 and is made of metal. Each terminal S2 is disposed in (or occupies) a portion of a cavity.

[0064] The selection and acceptance unit 52 accepts the selection input of the area of ​​interest. The area of ​​interest is the area in the first or second image that corresponds to the item S and the area that corresponds to the material that eliminates the effect of thickness. The selection input of the area of ​​interest is performed, for example, by displaying the first or second image on the display included in the output device 105, and by the user of the radiography device 5 selecting the area of ​​interest on the display via a mouse or the like included in the input device 106.

[0065] exist Figure 5 In the example shown, only the region in image P1 corresponding to the resin of the shell S1 is selected as the region of interest R. That is, in Figure 5 In the example shown, only the material in the image represented by image P1 that eliminates the effect of thickness (including the resin excluding the cavity portion of terminal S2) is correctly selected as the region of interest R. In this case, it is preferable that the region of interest contains information on a wide variety of thicknesses. On the other hand, in Figure 6 In the example shown, the area in image P1 corresponding to the resin of housing S1 and terminal S2 are selected as the area of ​​interest R. That is, in Figure 6 In the example shown, the material that eliminates the effect of thickness and other materials (metal) in the image represented by image P1 are selected as the region of interest R, but the region of interest R is not selected correctly.

[0066] The thickness correction function calculation unit 53, the representative data calculation unit 54, the evaluation coefficient calculation unit 55, and the notification unit 56 are the constituent elements of the process performed by the radiographic inspection device 5 to evaluate whether the area of ​​interest has been appropriately selected. Hereinafter, along with explanations of each constituent element, a summary of the process for evaluating whether the area of ​​interest has been appropriately selected will be provided.

[0067] The thickness correction function calculation unit 53 calculates a thickness correction function to eliminate the influence of the material's thickness. The thickness correction function is a function representing the relationship between the first pixel value and the second pixel value. Here, the calculation method for the thickness correction function will be explained. First, the thickness correction function calculation unit 53 performs a logarithmic transformation on the first image and the second image acquired by the acquisition unit 51. While the logarithmic transformation is not strictly necessary, it is preferable to perform it in order to construct an approximate function.

[0068] Then, the thickness correction function calculation unit 53 determines the first pixel value and the second pixel value. The first pixel value is the individual pixel value of a plurality of first pixels in the region corresponding to the region of interest in the first image after logarithmic transformation. The second pixel value is the individual pixel value of a plurality of second pixels corresponding to a plurality of first pixels in the second image after logarithmic transformation. The plurality of second pixels corresponding to a plurality of first pixels referred to here refers to each pixel in the region corresponding to the region of interest in the second image. That is, the thickness correction function calculation unit 53 performs logarithmic transformation on the first image and the second image, and determines the pixel value in the region corresponding to the region of interest in each of the logarithmically transformed first image and second image. Furthermore, the timing of the logarithmic transformation of the first image and the second image is not particularly limited. For example, it can be performed immediately after acquisition by the acquisition unit 51, or it can be performed after selection of the region of interest by the selection and acceptance unit 52.

[0069] Then, the thickness correction function calculation unit 53 calculates the thickness correction function by approximating the relationship between the first pixel value and the second pixel value corresponding to the first pixel value. The relationship between the first pixel value and the second pixel value corresponding to the first pixel value referred to here means the relationship between the pixel value of a single first pixel and the pixel value of the second pixel corresponding to that single first pixel. Specifically, for example, if the pixel value of the pixel at the top left of the region of interest in the first image is set as the first pixel value, then the pixel value of the pixel at the top left of the region of interest in the second image is the second pixel value corresponding to the first pixel value.

[0070] In this embodiment, the thickness correction function calculation unit 53 uses, for example, the least squares method to approximate the coefficients of the approximate curve by using an Nth-order approximation curve (N being a natural number greater than or equal to 1) on a two-dimensional coordinate system for multiple sample points. Therefore, the thickness correction function calculation unit 53 calculates the function representing the approximate curve, i.e., the thickness correction function. The multiple sample points are points representing the first pixel value and the second pixel value; for example, they are points drawn on a two-dimensional coordinate system with the first pixel value as the X-axis and the second pixel value as the Y-axis.

[0071] Figure 7 The coordinate graph G1 shown is a two-dimensional graph displaying multiple sample points D1 and the approximate curve C1. The horizontal axis (X-axis) of coordinate graph G1 represents the first pixel value, and the vertical axis (Y-axis) represents the second pixel value. Each sample point D1 is... Figure 5The sample points calculated corresponding to the region of interest R shown represent the first pixel value and the second pixel value corresponding to the first pixel value. The number of each sample point D1 is equal to the number of pixels in the first image (and the number of pixels in the second image) corresponding to the region of interest R. The approximation curve C1 is a quadratic thickness correction function (y = ax) calculated by performing a quadratic approximation on multiple sample points D1. 2 The curve is (bx + c).

[0072] Figure 8 The coordinate graph G2 shown is a two-dimensional graph displaying multiple sample points D2 and the approximate curve C2. Figure 7 Similarly, the horizontal axis (X-axis) of the coordinate graph represents the value of the first pixel, and the vertical axis (Y-axis) represents the value of the second pixel. Each sample point D2 is related to... Figure 6 The region of interest R shown corresponds to the calculated sample points. For example... Figure 8 As shown, multiple sample points D2 are compared to sample points D1 calculated corresponding to the correctly selected region of interest R (refer to...). Figure 7 The distribution is over a wider range. This is because multiple sample points D2 are based on the fact that image P1 consists not only of pixels equivalent to resin but also of parts outside of resin ( Figure 6 The region of interest R, consisting of pixels corresponding to terminal S2, is calculated. The approximate curve C2 is a quadratic thickness correction function (y = ax) calculated by performing a quadratic approximation on multiple sample points D2. 2 The curve is approximating C2 (+bx+c). The curvature of approximate curve C2 is greater than that of approximate curve C1 (refer to) the thickness correction function calculated corresponding to the correctly selected region of interest R. Figure 7 ).

[0073] The representative data calculation unit 54 calculates multiple representative data based on the respective first pixel values ​​of multiple first pixels and the respective second pixel values ​​of multiple second pixels corresponding to the multiple first pixels. The representative data is a combination of first representative values ​​and second representative values. The first representative value is a representative pixel value of a first image after logarithmic transformation, and the second representative value is a representative pixel value of a second image after logarithmic transformation. In this embodiment, the representative data is calculated by setting one first pixel value as the first representative value and setting the average of one or more second pixel values ​​corresponding to that first pixel value as the second representative value. Specifically, a first pixel value is set as the representative value, and if there is only one second pixel value corresponding to that first pixel, that second pixel value is set as the second representative value. If there are multiple second pixel values ​​corresponding to that first pixel, the average of those second pixel values ​​is set as the second representative value. The representative data calculation unit 54 calculates multiple representative data by repeatedly selecting a first pixel value (first representative value) and calculating a second representative value at predetermined intervals, such as every 0.01.

[0074] Figure 9 The coordinate graph G3 shown is related to Figure 7 The coordinate graph G1 shown is a two-dimensional graph that displays the approximate curve C1 and multiple representative data points F1. The horizontal axis (X-axis) of coordinate graph G3 represents the first pixel value, and the vertical axis (Y-axis) of coordinate graph G3 represents the second pixel value. Each representative data point F1 corresponds to a correctly selected region of interest R (refer to...). Figure 5 The combination of the first pixel value (i.e., the first representative value) and the average of the second pixel values ​​corresponding to each first pixel value (i.e., the second representative value).

[0075] Figure 10 The coordinate graph G4 shown is related to Figure 8 The coordinate graph G2 shown is a two-dimensional graph that displays the approximate curve C2 and multiple representative data points F2. The horizontal axis (X-axis) of coordinate graph G4 represents the first pixel value, and the vertical axis (Y-axis) of coordinate graph G4 represents the second pixel value. Representative data F2 represents the incorrectly selected region of interest R (see reference). Figure 6 The first pixel value (i.e., the first representative value) is the combination of the first pixel value and the average of the second pixel values ​​corresponding to each first pixel value (i.e., the second representative value).

[0076] The evaluation coefficient calculation unit 55 calculates an evaluation coefficient based on the correlation between the thickness correction function and multiple representative data points. In this embodiment, the evaluation coefficient is a determination coefficient calculated based on the thickness correction function and the representative data F1. The determination coefficient is calculated by subtracting the value obtained by dividing the residual variation of all representative data F1 by the total variation from 1. Therefore, the closer the determination coefficient is to 1, the higher the consistency. Figure 5In the example shown, the region of interest was correctly selected based on the thickness correction function and representative data F1 (refer to...). Figure 9 The coefficient of determination calculated is 0.988. On the other hand, in Figure 6 In the example shown, where the region of interest was not selected correctly, based on the thickness correction factor and representative data F2 (refer to...) Figure 10 The coefficient of determination calculated is 0.889. This means that the consistency between representative data F1 and approximate curve C1 is higher than the consistency between representative data F2 and approximate curve C2. When the region of interest is correctly selected, a higher evaluation coefficient can be calculated compared to when the region of interest is not correctly selected.

[0077] If the evaluation coefficient is below a predetermined evaluation coefficient threshold, the notification unit 56 outputs an error message related to the selection input of the area of ​​interest. The evaluation coefficient threshold is the lower limit of the determination coefficient output when the area of ​​interest is correctly selected, and is a preset value. For example, the evaluation coefficient threshold is set to 0.9. As an example, the notification unit 56 outputs an error message to the display included in the output device 105 containing at least one of text and graphics indicating that the selected area has not been correctly selected. Figure 5 In the example where the area of ​​interest was correctly selected, the calculated determination coefficient was 0.988. Therefore, the notification unit 56 determined that the evaluation coefficient was not below the evaluation coefficient threshold and did not output an error message to the display. On the other hand, in Figure 6 In the example shown where the area of ​​interest was not selected correctly, the calculated determination coefficient was 0.889. Therefore, the notification unit 56 determined that the evaluation coefficient was below the evaluation coefficient threshold and output an error message related to the selection input of the area of ​​interest to the display.

[0078] Furthermore, if the evaluation coefficient is below the prescribed evaluation coefficient threshold, the selection receiving unit 52 accepts the selection input of a region of interest that is different from the region of interest. Thus, for example, the user of the radiation inspection device 5 can select the region of interest again on the display via a mouse or the like included in the input device 106.

[0079] The differential image generation unit 57 is a component that performs subtraction processing to eliminate the influence of the thickness of the article S in the processing performed by the radiographic examination apparatus 5. Hereinafter, along with an explanation of the function of the differential image generation unit 57, an overview of the subtraction processing to eliminate the influence of the thickness of the article S will be described.

[0080] If the notification unit 56 determines that the evaluation coefficient is not below the evaluation coefficient threshold, the difference image generation unit 57 generates a difference image based on the first image and the second image, which have undergone logarithmic transformation. The difference image is an image with the region corresponding to the substance to be removed removed removed. The difference image generation unit 57 generates the difference image by performing a subtraction process. Specifically, as a subtraction process, the difference image generation unit 57 subtracts the values ​​calculated by applying a thickness correction function to the pixel values ​​of the first image, which has undergone logarithmic transformation, from the pixel values ​​of the second image, which has undergone logarithmic transformation. Alternatively, the difference image generation unit 57 subtracts the values ​​calculated by applying a thickness correction function to the pixel values ​​of the second image, which has undergone logarithmic transformation, from the pixel values ​​of the first image, which has undergone logarithmic transformation. Alternatively, the difference image generation unit 57 subtracts the pixel values ​​of the second image, which has undergone logarithmic transformation, from the values ​​calculated by applying a thickness correction function to the pixel values ​​of the first image, which has undergone logarithmic transformation. Alternatively, the differential image generation unit 57 subtracts the pixel values ​​of the logarithmically transformed first image from the values ​​calculated by applying a thickness correction function to the pixel values ​​of the logarithmically transformed second image. This generates a differential image, in which the material corresponding to the region of interest has been removed from either the first or second image. The differential image generation unit 57 then outputs the differential image to a display (not shown) included in the output device 105.

[0081] Figure 11 The difference image P2 shown is an example where the region of interest R is correctly selected (see reference). Figure 5 , Figure 7 and Figure 9 The differential image P2 is generated by the differential image generation unit 57. In the differential image P2, the resin portion of the housing S1 is appropriately removed. On the other hand, Figure 12 The difference image P3 shown is an example where the region of interest R was not selected correctly (see reference). Figure 6 , Figure 8 and Figure 10 In this case, even if the evaluation coefficient is below the evaluation coefficient threshold, the differential image generated by the differential image generation unit 57 is still performing differential image generation processing. In the differential image P3, the resin portion of the housing S1 is not properly removed.

[0082] The storage unit 58 stores the first and second images acquired by the acquisition unit 51, the data representing the region of interest input to the selection and acceptance unit 52, the thickness correction function calculated by the thickness correction function calculation unit 53, the representative data calculated by the representative data calculation unit 54, the evaluation coefficient calculated by the evaluation coefficient calculation unit 55, and the difference image generated by the difference image generation unit 57.

[0083] [Handling of X-ray examination equipment]

[0084] Next, refer to Figure 13 The flowchart describes the operation method of the radiation inspection apparatus 5 of this embodiment (the processes performed by the radiation inspection apparatus 5), i.e., the radiation inspection method, by each process. Figure 13 This is a flowchart illustrating the radiation inspection method of this embodiment. The processing performed by the radiation inspection apparatus 5 is carried out, for example, each time the article S is inspected by the image acquisition device 3.

[0085] First, in S11, the first image and the second image can be acquired from the image acquisition device 3 using the acquisition unit 51 (first step).

[0086] Next, in S12, the selection input of the area of ​​interest can be received using the selection acceptance unit 52 (step 2).

[0087] Next, in S13, the thickness correction function is calculated using the thickness correction function calculation unit 53 (step 3). Specifically, firstly, the first pixel values ​​of multiple first pixels in the region corresponding to the region of interest in the first image after logarithmic transformation are determined, and the second pixel values ​​of multiple second pixels corresponding to the multiple first pixels in the second image after logarithmic transformation are determined. Then, using least squares or similar methods, multiple sample points representing the first pixel values ​​of the multiple first pixels and the second pixel values ​​of the multiple second pixels corresponding to the multiple first pixels are approximated on a two-dimensional coordinate system using a quadratic approximation curve, thereby calculating the coefficients of the approximation curve. Finally, the function representing this approximation curve, i.e., the thickness correction function, is calculated.

[0088] Next, in S14, the representative data calculation unit 54 calculates multiple representative data based on the respective first pixel values ​​of the multiple first pixels and the respective second pixel values ​​of the multiple second pixels corresponding to the multiple first pixels (step 4). The representative data is a combination of a first representative value and a second representative value. The representative data is calculated by the representative data calculation unit 54 by setting a first pixel value as the first representative value and setting the average of one or more second pixel values ​​corresponding to that first pixel value as the second representative value.

[0089] Next, in S15, the evaluation coefficient calculation unit 55 calculates the evaluation coefficient based on the correlation between the thickness correction function and the calculated multiple representative data (step 5). In this embodiment, the evaluation coefficient is a determination coefficient calculated based on the thickness correction function and the calculated multiple representative data.

[0090] Next, in S16, the notification unit 56 determines whether the evaluation coefficient is below the evaluation coefficient threshold. If it is determined that the evaluation coefficient is below the evaluation coefficient threshold (S16: Yes), in S17, the notification unit 56 outputs error information related to the selection input of the area of ​​interest to the display of the output device 105 (step 6), causing the process to return to S12. Then, in S12, the selection acceptance unit 52 accepts the selection input of an area of ​​interest that is different from the area of ​​interest selected in the previous process S12 (step 7).

[0091] On the other hand, when it is determined that the evaluation coefficient is not below the evaluation coefficient threshold (S16: No), in S18, the differential image generation unit 57 generates a differential image in the first or second image from which the substance corresponding to the region of interest has been removed, and displays the differential image on the display of the output device 105. Thus, the radiographic inspection process for the item S ends.

[0092] [Functions and Effects]

[0093] According to the above-described radiation inspection apparatus 5, in a first image or a second image, a selection input of a region of interest is received. The apparatus determines the respective first pixel values ​​of multiple first pixels in the region corresponding to the region of interest in the first image (which has undergone logarithmic transformation), and the respective second pixel values ​​of multiple second pixels corresponding to the multiple first pixels in the second image (which has undergone logarithmic transformation). Then, an approximation is made of the relationship between the first pixel values ​​and the corresponding second pixel values, thereby calculating a thickness correction function and calculating multiple representative data, each of which is a combination of a first representative value representing the first pixel value and a second representative value representing the second pixel value. Then, an evaluation coefficient is calculated based on the correlation between the thickness correction function and the multiple representative data. Therefore, based on the evaluation coefficient, it is possible to evaluate whether the region of interest corresponds to a single substance. Thus, according to this embodiment, it is possible to evaluate whether the region of interest has been appropriately selected.

[0094] Specifically, in the X-ray inspection apparatus 5, a thickness correction function calculation unit 53 performs a logarithmic transformation on the first image and the second image, and a representative data calculation unit 54 calculates representative data, which is a combination of the representative pixel value (first representative value) of the logarithmically transformed first image and the representative pixel value (second representative value) of the logarithmically transformed second image. This allows for a more appropriate calculation of the thickness correction function representing the relationship between the first and second pixel values.

[0095] In the first embodiment, in the third step, the thickness correction function calculation unit 53 approximates multiple sample points representing the respective first pixel values ​​of multiple first pixels and the respective second pixel values ​​of multiple second pixels corresponding to the multiple first pixels on a two-dimensional coordinate system, thereby calculating the thickness correction function. This allows for reliable calculation of the thickness correction function.

[0096] In the first embodiment, in the third step, the thickness correction function calculation unit 53 approximates multiple sample points on a two-dimensional coordinate system using an Nth-order approximation curve, thereby calculating the coefficients of the approximation curve. This allows for the calculation of a thickness correction function that accurately represents the correspondence between the first pixel value and the corresponding second pixel value, thus enabling the calculation of highly reliable evaluation coefficients.

[0097] In the first embodiment, in the fourth step, the representative data calculation unit 54 calculates representative data by setting a first pixel value as a first representative value and setting the average of one or more second pixel values ​​corresponding to that first pixel value as a second representative value. The relationship between the first pixel value and the second pixel value can exist not only as a pixel value corresponding to a pixel of a material whose thickness influence is to be eliminated, but also as a pixel value corresponding to a pixel containing noise contained in the first and second images. In the first embodiment, by calculating the average of one or more first pixel values ​​or second pixel values, the influence of noise on the calculation of the evaluation coefficient can be suppressed, thus enabling high-precision calculation of the evaluation coefficient.

[0098] In the first embodiment, the evaluation coefficient is the determination coefficient. Therefore, it is possible to calculate an evaluation coefficient with high reliability.

[0099] In the first embodiment, in step 6, the notification unit 56 outputs error information related to the selection input of the area of ​​interest when the evaluation coefficient is below a predetermined evaluation coefficient threshold. This reminds the user to select an appropriate area of ​​interest again.

[0100] In the first embodiment, when the evaluation coefficient is below a predetermined evaluation coefficient threshold, the selection acceptance unit 52 accepts the selection input of a region of interest that is different from the region of interest selected in the previous selection acceptance process in step 7. Therefore, the user can reselect the region of interest corresponding to a single substance, and thus obtain an image with the image of the substance corresponding to the region of interest appropriately removed.

[0101] [Second Implementation]

[0102] [Composition of a Radiation Examination Device]

[0103] Next, the radiation inspection apparatus 5 according to the second embodiment will be described. In the second embodiment, as a means of processing error information output to the display included in the output device 105, the notification unit 56 of the radiation inspection apparatus 5 performs the following processing: When the evaluation coefficient is below a predetermined evaluation coefficient threshold, as error information, the notification unit 56 outputs a non-corresponding area overlapping the first image or the second image. A non-corresponding area refers to a region in the first image or the second image that does not correspond to the material used to eliminate the effect of thickness. The notification unit 56 displays, for example, a non-corresponding area selected in the area of ​​interest due to user error in the first image or the second image.

[0104] The following explanation of the functionality of the notification unit 56 will use an example of approximation using a quadratic approximation curve on a two-dimensional coordinate system with the first pixel value as the X-axis and the second pixel value as the Y-axis. First, when the evaluation coefficient is below the evaluation coefficient threshold, the notification unit 56 outputs an error message. Then, the notification unit 56 calculates the distance between the pixel value calculated based on the approximation curve and the actual pixel value (the pixel value in each sample point), i.e., the error distance.

[0105] Here, the method for calculating the error distance is explained. First, the notification unit 56 uses a thickness correction function to determine the difference between the second pixel value corresponding to the first pixel value and the actual second pixel value corresponding to the first pixel value in each of the plurality of first pixel values. Specifically, the notification unit 56 calculates the second pixel value derived from an approximate curve (the formula for the approximate curve is y = ax) for each first pixel value. 2 The difference between the value of y in +bx+c and the actual second pixel value corresponding to a first pixel value. Then, the notification unit 56 squares the difference value and takes the square root to calculate the error distance between the second pixel value calculated based on the approximate curve and the actual second pixel value.

[0106] Then, the notification unit 56 determines, based on the error distance, non-corresponding regions in the first or second image that do not correspond to the material that eliminates the effect of thickness. Specifically, the notification unit 56 determines whether a pixel corresponding to a first pixel value (and a pixel corresponding to a second pixel value) corresponds to the material that eliminates the effect of thickness by comparing the error distance and an error threshold. The notification unit 56 performs processing on all first pixel values ​​from calculating the difference value to determining whether they correspond to the material that eliminates the effect of thickness. Furthermore, when the second pixel value is on the X-axis and the first pixel value is on the Y-axis, the error distance between the first pixel value calculated based on the approximate curve and the actual first pixel value can also be calculated for each second pixel value.

[0107] The smaller the difference between the calculated second pixel value obtained using the approximate curve and the actual second pixel value, the higher the accuracy of the pixel corresponding to a first pixel value (and the second pixel value corresponding to that first pixel value) in representing the substance that eliminates the effect of thickness. Conversely, the larger the difference, the lower the accuracy of the pixel in representing the substance that eliminates the effect of thickness. The error threshold is a threshold used to determine whether the difference value represents the value of a pixel corresponding to the substance that eliminates the effect of thickness. The error threshold is a preset value, which can be calculated, for example, by the following method. First, when the radiation inspection device 5 is started, before the radiation inspection of the article S begins, the acquisition unit 51 captures a first image or a second image of the article S. Then, the selection input of the region of interest is received by the selection receiving unit 52. At this time, the user selects a region of interest in the first or second image that intentionally includes the region corresponding to the substance that eliminates the effect of thickness and the region that does not correspond to it. Then, the user determines the error threshold so that only the error distance in the pixels corresponding to the non-corresponding region can be determined. The method for determining the error threshold is as follows: First, the user inputs an arbitrary candidate threshold as a candidate error threshold into the computer system 20 via the input device 106. Then, image information (details described later) based on the comparison of the error distance in each first pixel value with the candidate threshold is displayed on the screen. If the image information includes not only non-corresponding areas but also areas corresponding to the material that eliminates the effect of thickness, the user inputs the candidate threshold again and repeats the above process. If the image information only represents areas that do not correspond to the material that eliminates the effect of thickness as non-corresponding areas, the user determines the candidate threshold as the error threshold. Thus, an appropriate candidate threshold is determined as the error threshold.

[0108] Then, the notification unit 56, overlapping with the first image or the second image, outputs image information representing the region in the first image or the second image that does not correspond to the material that eliminates the effect of thickness, i.e., the non-corresponding region, as error information. Specifically, the notification unit 56 displays the pixel corresponding to the error distance overlapping with the first image or the second image. Figure 14 The image information (error message) I shown is from image P1 (reference) which is the first image. Figure 5 and Figure 6 (Example of overlapping non-corresponding region E) Figure 14 In the example shown, the regions in image P1 corresponding to the resin (the substance that eliminates the effect of thickness) of the shell S1 and the regions not corresponding to the resin (equivalent to...) are... Figure 14The area of ​​terminal S2 in the image is selected as the area of ​​interest. Therefore, the area of ​​terminal S2 that does not correspond to the material that eliminates the effect of thickness is displayed as non-corresponding area E in image information I of image P1 on the display. Examples of displaying non-corresponding area E include areas drawn with diagonal lines in red boxes, areas filled with red, etc.

[0109] [Handling of X-ray examination equipment]

[0110] Next, refer to Figure 15 The flowchart describes the operation method (processes performed in the radiation inspection apparatus 5) of this embodiment, i.e., the radiation inspection method, by each process. Figure 15 This is a flowchart illustrating the radiation examination method of this embodiment.

[0111] The processing of S21 to S27 and Figure 13 The processing of S11 to S17 is the same. After the processing of S21 to S27, in S28, using the notification unit 56, the error distance between the second pixel value calculated using the thickness correction function (in other words, calculated from the approximate curve) and the actual second pixel value is calculated in each first pixel value (step 6). Next, in S29, using the notification unit 56, the error distance in each first pixel value is compared with the error threshold. Therefore, pixels with error distances larger than the error threshold are identified as non-corresponding regions that do not correspond to the material that eliminates the effect of thickness (step 6). Next, in S30, using the notification unit 56, image information showing non-corresponding regions overlapping in the first or second image is displayed (output) as error information, and the processing returns to S22 (step 6). The processing of S31 is the same as... Figure 13 The processing in S18 shown is the same. Furthermore, the processing of notifying the user of an error (S27) and the processing of displaying image information (S30) can be performed simultaneously. That is, in the radiographic examination method, the notification of an error and the display of image information can be performed simultaneously.

[0112] [Functions and Effects]

[0113] In the radiation inspection apparatus 5 of the second embodiment described above, the same effects as in the first embodiment are achieved. Furthermore, in the sixth step, in the radiation inspection apparatus 5 of the second embodiment, a thickness correction function is used to calculate the distance between the second pixel value corresponding to the first pixel value and the actual second pixel value corresponding to the first pixel value, i.e., the error distance, among the respective first pixel values ​​of the plurality of first pixels. Based on the error distance, a region in the first or second image that does not correspond to the material that eliminates the effect of thickness, i.e., a non-corresponding region, is identified as error information and is output as an overlapping non-corresponding region in the first or second image. This allows the user to easily select an appropriate area of ​​interest. In other words, in the radiation inspection apparatus 5 of the second embodiment, the display shows not only the error message but also the aforementioned image information, thereby assisting the user in reselecting an appropriate area of ​​interest.

[0114] [Third Implementation]

[0115] [Composition of a Radiation Examination Device]

[0116] Next, the radiation inspection apparatus 5 according to the third embodiment will be described. In the third embodiment, the thickness correction function calculation unit 53 of the radiation inspection apparatus 5 performs the following processing: The thickness correction function calculation unit 53 suppresses the influence of outliers in multiple sample points representing the respective first pixel values ​​of multiple first pixels and the respective second pixel values ​​of multiple second pixels corresponding to the multiple first pixels, and approximates the multiple sample points on a two-dimensional coordinate system, thereby calculating the thickness correction function.

[0117] Specifically, in this embodiment, the thickness correction function calculation unit 53 calculates the thickness correction function by implementing the M-estimation method in the robust estimation method. First, the thickness correction function calculation unit 53 approximates multiple sample points on a two-dimensional coordinate system using methods such as least squares, thereby deriving an approximate curve. Then, for each first pixel value, the thickness correction function calculation unit 53 weights the approximate curve again based on the distance between the second pixel value calculated using the approximate curve and the actual second pixel value. Specifically, the thickness correction function calculation unit 53 reduces the weight as the distance between the second pixel value calculated using the approximate curve and the actual second pixel value increases. The thickness correction function calculation unit 53 repeatedly performs this process from weighting to recalculating the approximate curve. Thus, the influence of outliers among the multiple sample points is suppressed, the approximate curve is determined, and the function representing the approximate curve, i.e., the thickness correction function, is calculated.

[0118] [Handling of X-ray examination equipment]

[0119] Next, refer to Figure 16 The flowchart describes the operation method of the radiation inspection apparatus 5 of this embodiment (the processes performed by the radiation inspection apparatus 5), i.e., the radiation inspection method, by each process. Figure 16 This is a flowchart illustrating the radiation examination method of this embodiment.

[0120] The processing of S41 and S42 and Figure 13 The processing in S11 and S12 is the same. After the processing in S41 and S42, in S43, the thickness correction function calculation unit 53 is used to suppress the influence of outliers in multiple sample points, and the multiple sample points are approximated on two-dimensional coordinates, thereby calculating the thickness correction function (step 3). Specifically, the thickness correction function calculation unit 53 uses the M-estimation method in the robust estimation method described above, thereby repeatedly calculating the approximate curve using the weighted least squares method, and thus determining the approximate curve. Therefore, the function representing the approximate curve, i.e., the thickness correction function, is calculated. The processing in S44 to S48 is the same as... Figure 13 The processing of S14 to S18 shown is the same.

[0121] [Functions and Effects]

[0122] In the radiographic inspection apparatus 5 of the third embodiment described above, the same effects as in the first embodiment are achieved. Furthermore, in the third step of the radiographic inspection apparatus 5 of the third embodiment, the influence of outliers in multiple sample points representing the respective first pixel values ​​of multiple first pixels and the respective second pixel values ​​of multiple second pixels corresponding to the multiple first pixels is suppressed, and the multiple sample points are approximated on a two-dimensional coordinate system, thereby calculating the thickness correction function. Therefore, the approximation of the multiple sample points can be performed with high precision, and thus the evaluation coefficient can be calculated with high precision.

[0123] In particular, in the radiation inspection apparatus 5 of the third embodiment, M-estimation in robust estimation is used as a method to suppress the influence of outliers among multiple sample points. This enables a reliable and highly accurate approximation of multiple sample points.

[0124] [Variation Example]

[0125] This disclosure is not limited to the embodiments described above. For example, the approximation curve used to approximate multiple sample points is not limited to quadratic. For example, the thickness correction function calculation unit 53 can approximate multiple sample points on a two-dimensional coordinate system using an Nth-order approximation curve (N is a natural number greater than or equal to 1). That is, the approximation curve used to approximate multiple sample points can also be of tertiary order or higher. As a result, a thickness correction function that accurately represents the correspondence between the first pixel value and the corresponding second pixel value can be calculated, thus enabling the calculation of a highly reliable evaluation coefficient. Furthermore, the approximation of multiple sample points can also be implemented using a single-order approximation straight line.

[0126] Here, an example is given of the result of the correction process on an article S, which is a food product, using the radiation inspection device 5 of the first embodiment.

[0127] exist Figure 17 The image shown is an example of the first image of item S. Item S, the object being processed, is an item containing multiple aluminum sheets S4 within a portion of a plurality of arranged dumplings S3. Figure 18 The graph in the image represents the coordinates of the approximate curve C3. The approximate curve C3 is... Figure 17 The curve was calculated using an approximation by the X-ray inspection device 5, where a region of interest R was selected in the area containing dumplings S3 (excluding aluminum sheet S4) on the first image shown. Specifically, in Figure 18 In the diagram, section (a) shows the sample point D3 and coordinate graph G5 representing the approximate curve C3, and section (b) shows the representative data F3 and coordinate graph G6 representing the approximate curve C3. In this case, the coefficient of determination is calculated to be 0.990, which is a relatively high value. Furthermore, in... Figure 19 The image shows a differential image P4 generated by the X-ray inspection device 5 in this case, and a foreign object detection result image P5 after thresholding the differential image. Thus, in the differential image P4 and the result image P5, the influence of the thickness of the portion of the dumpling S3 can be eliminated to some extent, allowing the detection of part H1 of the aluminum sheet S4. However, false detections of part H2 and parts H3 that should have been detected but were not will occur. In particular, the thicker portion of the dumpling S3 is imaged as white in the differential image P4, and there is a tendency that this portion cannot be completely eliminated.

[0128] exist Figure 20 The image shows a coordinate graph of the approximate curve C4. The approximate curve C4 is... Figure 17 The curve shown in the first image is calculated using a second approximation by the X-ray inspection device 5, with the region of interest R selected in the area containing the dumpling S3 (excluding the aluminum sheet S4). Specifically, in Figure 20In the diagram, section (a) shows the sample point D3 and the coordinate graph G7 representing the approximate curve C4, and section (b) shows the representative data F3 and the coordinate graph G8 representing the approximate curve C4. In this case, the coefficient of determination is calculated to be 0.998, which is higher than that of the first approximation. Furthermore, in... Figure 21 The image shows a differential image P6 generated by the X-ray inspection device 5 and a foreign object detection result image P7 after thresholding the differential image. Thus, in both the differential image P6 and the result image P7, the influence of the thickness of the portion of dumpling S3 can be eliminated entirely at any brightness, and the entire portion H1 of aluminum sheet S4 can be detected.

[0129] Based on these results, it can be seen that, depending on the type of item S, the second approximation is more effective than the first approximation in completely eliminating the effect of thickness in the difference image.

[0130] In addition, Figure 22 In, it is shown that in relation to Figure 17 In the first image of the same item S, the area containing aluminum sheet S4 is incorrectly selected as the area of ​​interest R. Figure 23 The diagram shows a coordinate graph of the approximate curve C5 calculated using a second approximation with the radiation examination device 5 in this case. Specifically, in Figure 23 In the diagram, section (a) shows the coordinate graph G9 representing sample point D4 and the approximate curve C5, and section (b) shows the coordinate graph G10 representing representative data F4 and the approximate curve C5. In this case, the coefficient of determination is calculated to be 0.988, which is relatively low compared to the coefficient of determination of 0.998 when the region of interest R is not suitable. Furthermore, in... Figure 24 The diagram shows a differential image P8 generated by the radiation inspection device 5 in this case, and a foreign object detection result image P9 after thresholding the differential image. Thus, in the differential image P8 and the result image P9, the dumpling S3 and the aluminum sheet S4 are generated with the same brightness, indicating that the aluminum sheet S4 cannot be detected. The threshold of the determination coefficient varies depending on the object being inspected; preferably, the threshold of the determination coefficient is changed according to the object being inspected. The method for determining the threshold is as described above.

[0131] Furthermore, in the embodiments described above, the thickness correction function calculation unit 53 performs a logarithmic transformation on the first image and the second image, and the representative data calculation unit 54 calculates a combination of the representative pixel value (first representative value) of the logarithmically transformed first image and the representative pixel value (second representative value) of the logarithmically transformed second image, which is the representative data. However, the first image and the second image may also be exempt from logarithmic transformation.

[0132] In addition, in the above embodiments, for example, the relationship between the first pixel value and the second pixel value corresponding to the first pixel value can also be calculated based on a LUT (Look Up Table).

[0133] Alternatively, the evaluation coefficient can also be a correlation coefficient. Therefore, it is also possible to calculate an evaluation coefficient with high reliability.

[0134] Alternatively, for example, representative data can also be calculated by setting a second pixel value as the second representative value and setting the average of one or more first pixel values ​​corresponding to that second pixel value as the first representative value.

[0135] Alternatively, representative data can be calculated, for example, by setting a first pixel value as the first representative value and setting the median value of one or more second pixel values ​​corresponding to that first pixel value as the second representative value; or by setting a second pixel value as the second representative value and setting the median value of one or more first pixel values ​​corresponding to that second pixel value as the first representative value. By obtaining the median value of one or more first pixel values ​​or second pixel values, the influence of noise on the calculation of evaluation coefficients can be suppressed, thus enabling the calculation of evaluation coefficients with high accuracy.

[0136] Furthermore, the error messages displayed by the notification unit 56 are not limited to the forms described in the above embodiments. Additionally, in the radiation inspection apparatus 5, it is also possible to omit the output of error messages related to the selection input of the area of ​​interest.

[0137] Alternatively, other robust estimation methods can be used to implement this process. Examples of other robust estimation methods include RANSAC and the minimum median method.

[0138] In addition, flowcharts for the aforementioned radiation examination methods ( Figure 13 , Figure 15 and Figure 16 The processes shown can also be omitted appropriately. Furthermore, the order of the processes can be rearranged as appropriate.

[0139] In the methods and apparatus of the above embodiments, the first pixel value may be the individual pixel value of a plurality of first pixels in the region corresponding to the region of interest in the first image after logarithmic transformation, and the second pixel value may be the individual pixel value of a plurality of second pixels corresponding to the plurality of first pixels in the second image after logarithmic transformation, and each of the plurality of representative data may be a combination of a representative pixel value (i.e., a first representative value) of the first image after logarithmic transformation and a representative pixel value (i.e., a second representative value) of the second image after logarithmic transformation. This allows for a more appropriate calculation of the thickness correction function representing the relationship between the first pixel value and the second pixel value.

[0140] In the above method and apparatus, the thickness correction function can also be calculated by approximating multiple sample points representing the respective first pixel values ​​of multiple first pixels and the respective second pixel values ​​of multiple second pixels corresponding to the multiple first pixels on a two-dimensional coordinate system. This allows for the reliable calculation of the thickness correction function.

[0141] In the above methods and apparatus, the coefficients of the approximate curve can also be calculated by approximating multiple sample points on a two-dimensional coordinate system using an approximation curve of degree N (N being a natural number greater than 1). This allows for the calculation of a thickness correction function that accurately represents the correspondence between the first pixel value and the corresponding second pixel value, thus enabling the calculation of highly reliable evaluation coefficients.

[0142] In the above-described method and apparatus, the multiple representative data can be calculated by setting a first pixel value as a first representative value and setting the average of one or more second pixel values ​​corresponding to that first pixel value as a second representative value, or by setting a second pixel value as a second representative value and setting the average of one or more first pixel values ​​corresponding to that second pixel value as a first representative value. The relationship between the first pixel value and the second pixel value can exist not only as pixel values ​​corresponding to pixels of the material whose thickness effect is to be eliminated, but also as pixel values ​​corresponding to pixels containing noise contained in the first and second images. In the above-described method and apparatus, by calculating the average of one or more first pixel values ​​or second pixel values, the influence of noise on the calculation of the evaluation coefficient can be suppressed, thus enabling the evaluation coefficient to be calculated with high accuracy.

[0143] In the above-described method and apparatus, each of the multiple representative data is calculated by setting a first pixel value as a first representative value and setting the median value of one or more second pixel values ​​corresponding to that first pixel value as a second representative value, or by setting a second pixel value as a second representative value and setting the median value of one or more first pixel values ​​corresponding to that second pixel value as a first representative value. The relationship between the first pixel value and the second pixel value can exist not only as pixel values ​​corresponding to pixels of the material whose thickness effect is to be eliminated, but also as pixel values ​​corresponding to pixels containing noise contained in the first and second images. In the above-described method and apparatus, by obtaining the median value of one or more first pixel values ​​or second pixel values, the influence of noise on the calculation of the evaluation coefficient can be suppressed, thus enabling high-precision calculation of the evaluation coefficient.

[0144] In the above methods and apparatus, the evaluation coefficient can also be the determining coefficient. Therefore, it is possible to calculate an evaluation coefficient indicating high reliability.

[0145] In the above methods and apparatus, the evaluation coefficient can also be a correlation coefficient. Therefore, it is possible to calculate an evaluation coefficient with high reliability.

[0146] In the above method and apparatus, when the evaluation coefficient is below a predetermined evaluation coefficient threshold, error information related to the selection input of the area of ​​interest can be output. This reminds the user to reselect an appropriate area of ​​interest.

[0147] In the above method and apparatus, a thickness correction function can be used to calculate the distance (error distance) between the second pixel value corresponding to the first pixel value and the actual second pixel value corresponding to the first pixel value. Based on the error distance, regions in the first or second image that do not correspond to the material used to eliminate the effect of thickness (i.e., non-corresponding regions) are identified and output as error information, overlapping the non-corresponding regions in the first or second image. This allows users to easily select appropriate regions of interest.

[0148] In the above method and apparatus, if the evaluation coefficient is below a predetermined evaluation coefficient threshold, the user can accept the selection input of a region of interest that is different from the region of interest. This allows the user to select the region of interest corresponding to a single substance again, thus enabling the acquisition of an image with the image of the substance corresponding to the region of interest appropriately removed.

[0149] In the above-described method and apparatus, the thickness correction function can also be calculated by suppressing the influence of outliers in multiple sample points representing the respective first pixel values ​​of multiple first pixels and the respective second pixel values ​​of multiple second pixels corresponding to the multiple first pixels, and by approximating the multiple sample points in two-dimensional coordinates. This allows for high-precision approximation of the multiple sample points, and thus enables high-precision calculation of the evaluation coefficients.

Claims

1. A method for radiation examination, in, include: Step 1: Obtain a first image and a second image. The first image is obtained by photographing the object under conditions where it is irradiated with radiation having a first energy distribution, and the second image is obtained by photographing the object under conditions where it is irradiated with radiation having a second energy distribution that is different from the first energy distribution. Step 2: Accept the input of the selection of the area of ​​interest, which is the area in the first image or the second image that corresponds to the material that eliminates the effect of thickness in the area corresponding to the item. Step 3: Determine the respective first pixel values ​​of multiple first pixels in the region corresponding to the region of interest in the first image, and the respective second pixel values ​​of multiple second pixels corresponding to the multiple first pixels in the second image. By approximating the relationship between the first pixel values ​​and the second pixel values ​​corresponding to the first pixel values, calculate a thickness correction function to eliminate the influence of the thickness of the material. The thickness correction function is a function that represents the relationship between the first pixel values ​​and the second pixel values. Step 4: Calculate multiple representative data based on the respective first pixel values ​​of the plurality of first pixels and the respective second pixel values ​​of the plurality of second pixels corresponding to the plurality of first pixels. Each of the multiple representative data is a combination of a first representative value representing the pixel value of the first image and a second representative value representing the pixel value of the second image; and Step 5 involves calculating the evaluation coefficient based on the correlation between the thickness correction function and the calculated representative data, and then evaluating whether the region of interest corresponds to a single substance based on the evaluation coefficient. In step 4, each of the plurality of representative data is calculated in the following manner: The calculation is performed by setting a first pixel value as the first representative value and setting the average or median value of one or more second pixel values ​​corresponding to that first pixel value as the second representative value. The calculation is performed by setting a second pixel value as the second representative value and setting the average or median value of one or more first pixel values ​​corresponding to the second pixel value as the first representative value.

2. The radiation examination method as described in claim 1, wherein, The first pixel value is the individual pixel value of the plurality of first pixels in the region corresponding to the region of interest in the first image after logarithmic transformation. The second pixel value is the individual pixel value of the plurality of second pixels in the second image that corresponds to the plurality of first pixels, after logarithmic transformation. Each of the plurality of representative data is a combination of the first representative value and the second representative value, wherein the first representative value is a representative pixel value of the first image after logarithmic transformation, and the second representative value is a representative pixel value of the second image after logarithmic transformation.

3. The radiation examination method as described in claim 1 or 2, wherein, In the third step, the thickness correction function is calculated by approximating multiple sample points on a two-dimensional coordinate system, representing the respective first pixel values ​​of the plurality of first pixels and the respective second pixel values ​​of the plurality of second pixels corresponding to the plurality of first pixels.

4. The radiation examination method as described in claim 3, wherein, In the third step, the coefficients of the approximate curve are calculated by approximating the plurality of sample points on the two-dimensional coordinates using an approximation curve of degree N, where N is a natural number greater than or equal to 1.

5. The radiation examination method as described in claim 1 or 2, wherein, The evaluation coefficient is the coefficient of determination.

6. The radiation examination method as described in claim 1 or 2, wherein, The evaluation coefficient is the correlation coefficient.

7. The radiation examination method as described in claim 1 or 2, wherein, It also includes: Step 6, when the evaluation coefficient is below the specified evaluation coefficient threshold, outputting error information related to the selection input of the region of interest.

8. The radiological examination method as described in claim 7, wherein, In the sixth step, Among the respective first pixel values ​​of the plurality of first pixels, the thickness correction function is used to calculate an error distance, which is the distance between the second pixel value corresponding to the first pixel value and the actual second pixel value corresponding to the first pixel value. The non-corresponding regions are identified based on the error distance. These non-corresponding regions are areas in either the first or second image that do not correspond to the material used to eliminate the effect of thickness. As the error message, the non-corresponding area is output overlapping the first image or the second image.

9. The radiation examination method as described in claim 1 or 2, wherein, It also includes: Step 7, when the evaluation coefficient is below the prescribed evaluation coefficient threshold, accepting the selection input of a region of interest that is different from the region of interest.

10. The radiation examination method as described in claim 1 or 2, wherein, In the third step, the thickness correction function is calculated by suppressing the influence of outliers in a plurality of sample points representing the respective first pixel values ​​of the plurality of first pixels and the respective second pixel values ​​of the plurality of second pixels corresponding to the plurality of first pixels, and by approximating the plurality of sample points on two-dimensional coordinates.

11. A radiation examination device, wherein, Includes at least one processor, The at least one processor performs the following processing: A first image and a second image are acquired. The first image is obtained by photographing the object under conditions where it is irradiated with radiation having a first energy distribution. The second image is obtained by photographing the object under conditions where it is irradiated with radiation having a second energy distribution that is different from the first energy distribution. The input for selecting the area of ​​interest is the area in the first or second image that corresponds to the material used to eliminate the effect of thickness, within the region corresponding to the item. A thickness correction function is calculated to eliminate the influence of the material's thickness by determining the respective first pixel values ​​of multiple first pixels in the region corresponding to the region of interest in the first image, and the respective second pixel values ​​of multiple second pixels corresponding to the multiple first pixels in the second image. This is done by approximating the relationship between the first pixel values ​​and the corresponding second pixel values. Based on the respective first pixel values ​​of the plurality of first pixels and the respective second pixel values ​​of the plurality of second pixels corresponding to the plurality of first pixels, a plurality of representative data are calculated, each of the plurality of representative data being a combination of a first representative value representing the pixel value of the first image and a second representative value representing the pixel value of the second image. Calculate the evaluation coefficient based on the correlation between the thickness correction function and the calculated representative data, and based on the evaluation coefficient, evaluate whether the region of interest corresponds to a single substance. Each of the multiple representative data points is calculated in the following ways: The calculation is performed by setting a first pixel value as the first representative value and setting the average or median value of one or more second pixel values ​​corresponding to that first pixel value as the second representative value. The calculation is performed by setting a second pixel value as the second representative value and setting the average or median value of one or more first pixel values ​​corresponding to the second pixel value as the first representative value.

12. The radiation examination apparatus as claimed in claim 11, wherein, The first pixel value is the individual pixel value of the plurality of first pixels in the region corresponding to the region of interest in the first image after logarithmic transformation. The second pixel value is the individual pixel value of the plurality of second pixels in the second image that corresponds to the plurality of first pixels, after logarithmic transformation. Each of the plurality of representative data is a combination of the first representative value and the second representative value, wherein the first representative value is a representative pixel value of the first image after logarithmic transformation, and the second representative value is a representative pixel value of the second image after logarithmic transformation.

13. The radiation examination apparatus as claimed in claim 11 or 12, wherein, The at least one processor calculates the thickness correction function by approximating multiple sample points on two-dimensional coordinates representing the respective first pixel values ​​of the plurality of first pixels and the respective second pixel values ​​of the plurality of second pixels corresponding to the plurality of first pixels.

14. The radiation examination apparatus as claimed in claim 13, wherein, The at least one processor calculates the coefficients of the approximate curve by approximating the plurality of sample points on the two-dimensional coordinates using an approximation curve of degree N, where N is a natural number greater than or equal to 1.

15. The radiation examination apparatus as claimed in claim 11 or 12, wherein, The evaluation coefficient is the coefficient of determination.

16. The radiation examination apparatus as claimed in claim 11 or 12, wherein, The evaluation coefficient is the correlation coefficient.

17. The radiation examination apparatus as claimed in claim 11 or 12, wherein, When the evaluation coefficient is below a specified evaluation coefficient threshold, the at least one processor outputs error information related to the selection input of the region of interest.

18. The radiation examination apparatus as claimed in claim 17, wherein, The at least one processor performs the following processing: Among the respective first pixel values ​​of the plurality of first pixels, the thickness correction function is used to calculate an error distance, which is the distance between the second pixel value corresponding to the first pixel value and the actual second pixel value corresponding to the first pixel value. The non-corresponding regions are identified based on the error distance. These non-corresponding regions are areas in either the first or second image that do not correspond to the material used to eliminate the effect of thickness. As the error message, the non-corresponding area is output overlapping the first image or the second image.

19. The radiation examination apparatus as claimed in claim 11 or 12, wherein, When the evaluation coefficient is below a specified evaluation coefficient threshold, the at least one processor accepts the selection input of a region of interest that is different from the region of interest.

20. The radiation examination apparatus as claimed in claim 11 or 12, wherein, The at least one processor calculates the thickness correction function by suppressing the influence of outliers in a plurality of sample points representing the respective first pixel values ​​of the plurality of first pixels and the respective second pixel values ​​of the plurality of second pixels corresponding to the plurality of first pixels, and by approximating the plurality of sample points in two-dimensional coordinates.

21. A radiographic examination system, comprising: The radiation examination apparatus according to any one of claims 11 to 20; A radiation source that irradiates the article with radiation having the first energy distribution and radiation having the second energy distribution; and A detector that detects radiation that has been irradiated from the radiation source and has transmitted through the article and has the first energy distribution, and radiation that has been irradiated from the radiation source and has transmitted through the article and has the second energy distribution.

22. A radiation screening program product that enables a computer to function by comprising the following steps: Step 1: Obtain a first image and a second image. The first image is obtained by photographing the object under conditions where it is irradiated with radiation having a first energy distribution, and the second image is obtained by photographing the object under conditions where it is irradiated with radiation having a second energy distribution that is different from the first energy distribution. Step 2: Accept the input of the selection of the area of ​​interest, which is the area in the first image or the second image that corresponds to the material that eliminates the effect of thickness in the area corresponding to the item. Step 3: Determine the respective first pixel values ​​of multiple first pixels in the region corresponding to the region of interest in the first image, and the respective second pixel values ​​of multiple second pixels corresponding to the multiple first pixels in the second image. By approximating the relationship between the first pixel values ​​and the second pixel values ​​corresponding to the first pixel values, calculate a thickness correction function to eliminate the influence of the thickness of the material. The thickness correction function is a function that represents the relationship between the first pixel values ​​and the second pixel values. Step 4: Calculate multiple representative data based on the respective first pixel values ​​of the plurality of first pixels and the respective second pixel values ​​of the plurality of second pixels corresponding to the plurality of first pixels. Each of the multiple representative data is a combination of a first representative value representing the pixel value of the first image and a second representative value representing the pixel value of the second image. and Step 5 involves calculating the evaluation coefficient based on the correlation between the thickness correction function and the calculated representative data, and then evaluating whether the region of interest corresponds to a single substance based on the evaluation coefficient. In step 4, each of the plurality of representative data is calculated in the following manner: The calculation is performed by setting a first pixel value as the first representative value and setting the average or median value of one or more second pixel values ​​corresponding to that first pixel value as the second representative value. The calculation is performed by setting a second pixel value as the second representative value and setting the average or median value of one or more first pixel values ​​corresponding to the second pixel value as the first representative value.

Images