Image processing apparatus, image processing method, and image processing program

Abstract

An image processing apparatus includes: an acquisition unit that acquires a transmission image of a pipe as a processing target and an image of the pipe without a defect corresponding to the transmission image, i.e., a defect-free pipe image; a generation unit that generates a difference image representing a difference between the acquired transmission image and the defect-free pipe image; and a visualization unit that visualizes a level of a defect in the pipe based on a pixel value of the difference image.

Description

Technical Field

[0001] This invention relates to an image processing apparatus, an image processing method, and an image processing program. Background Technology

[0002] In the past, measurements of transmission information obtained through radiography were sometimes used to inspect piping. In particular, X-ray and other radiographic inspections have been highly regarded as efficient inspection methods compared to ultrasound inspections due to their high transmittance, eliminating the need to remove protective materials around the piping, and providing clear images of the subject.

[0003] The primary focus of piping inspection is wall thickness reduction. Here, "wall thickness reduction" refers to the phenomenon where the internal thickness of the piping is reduced due to some reason. Excessive wall thickness reduction can damage the piping and cause significant harm to the facilities containing it; therefore, early detection and countermeasures, as well as regular inspections, are essential.

[0004] Piping is typically a large structure along its axis, thus requiring efficient inspection. The following techniques are applicable to this task.

[0005] Japanese Patent Application Publication No. 2008-256603 discloses a non-destructive inspection device, the purpose of which is to non-destructively detect the deterioration or damage of a tubular test object before leakage occurs.

[0006] The non-destructive testing device comprises: a radiation source that irradiates a tubular subject from the side; an image sensor that detects the radiation transmitted through the tubular subject to obtain a transmission image; a tube wall information generation mechanism that generates tube wall information of the tubular subject based on the transmission image obtained by the image sensor; and an evaluation mechanism that evaluates the integrity of the tubular subject based on the tube wall information generated by the tube wall information generation mechanism.

[0007] Furthermore, Japanese Patent Application Publication No. 2021-148755 discloses a method for inspecting structures, the purpose of which is to reliably detect defects existing in structures.

[0008] The inspection method for this structure includes: Step 1, irradiating the structure to be inspected with X-rays to obtain a digital X-ray transmission image; Step 2, identifying defective and intact parts based on an image processed from the digital X-ray transmission image and the digital X-ray transmission image; Step 3, storing the location information of the defective and intact parts; Step 4, calculating the average brightness within a specified area of ​​the defective and intact parts in the digital X-ray transmission image before image processing based on the location information, and then calculating the difference in average brightness between the defective and intact parts; and Step 5, calculating the wall thickness of the defective and intact parts based on the average brightness of the defective and intact parts and their brightness difference. Summary of the Invention

[0009] -The technical problem that the invention aims to solve-

[0010] In transmission images obtained by radiography of the piping under inspection, the areas on the sides of the piping appear nearly white due to their low transmittance, while the interior areas appear nearly black due to their higher transmittance closer to the center. Conversely, the background areas in the aforementioned transmission images appear closest to black because they have the highest transmittance.

[0011] Thus, in the transmission image, the side areas of the pipe are close to white, while the background area is closest to black. Therefore, it is easy to detect wall thinning in this side area. However, the internal areas of the pipe become closer to black towards the center, making it difficult to detect wall thinning. Furthermore, this is not limited to detecting wall thinning; it also occurs when detecting other defects such as cracks and blockages.

[0012] This invention provides an image processing device, image processing method, and image processing program that can detect internal defects in piping with higher accuracy compared to existing technologies.

[0013] -Means used to solve technical problems-

[0014] A first aspect of the present invention is an image processing apparatus comprising at least one processor, the processor performing the following processing: acquiring a transmission image of a pipe as the object of processing and an image of the defect-free pipe corresponding to the transmission image, namely a defect-free pipe image; generating a difference image representing the difference between the transmission image and the defect-free pipe image; and visualizing the level of defects in the pipe based on the pixel values ​​of the difference image.

[0015] In the second aspect of the present invention, the defect-free piping image can be a simulated image, as in the first aspect.

[0016] In the third aspect of the present invention, in the second aspect, the defect-free piping image can be an image generated by averaging the pixel values ​​of the transmission image along the axial direction of the piping.

[0017] In the second aspect, the defect-free piping image can be an image generated based on at least one of the location and region on a transmission image received from the user.

[0018] In the fifth aspect of the present invention, in the second aspect, the defect-free piping image can be an image generated by mimicking a transmission image.

[0019] In the sixth aspect of the present invention, in either the first or second aspect, the processor can remove noise by filtering the differential image.

[0020] In the seventh aspect of the present invention, the defect in the first or second aspect may be a reduction in the wall thickness of the piping.

[0021] In the seventh aspect, the processor can visualize the image by creating a heatmap based on the pixel values ​​of the differential image.

[0022] In the seventh embodiment of the invention, the processor can visualize the wall thickness reduction by adjusting at least one of the brightness and contrast of the differential image.

[0023] In the 7th embodiment, the 10th embodiment of the present invention allows the processor to visualize the wall thickness reduction by binarizing the differential image.

[0024] In the eleventh aspect of the present invention, in either the first or second aspect, the defect-free piping image can be an image with the same viewing angle as the transmission image.

[0025] Furthermore, the 12th aspect of the present invention is an image processing method in which a processor of an image processing apparatus performs the following processing: acquiring a transmission image of a pipe as the object of processing and an image of the defect-free pipe corresponding to the transmission image, i.e., a defect-free pipe image; generating a difference image representing the difference between the transmission image and the defect-free pipe image; and visualizing the level of defects in the pipe based on the pixel values ​​of the difference image.

[0026] Furthermore, the 13th aspect of the present invention is an image processing program that causes a processor of an image processing apparatus to perform the following processing: acquiring a transmission image of a pipe as the object of processing and an image of the defect-free pipe corresponding to the transmission image, i.e., a defect-free pipe image; generating a difference image representing the difference between the transmission image and the defect-free pipe image; and visualizing the level of defects in the pipe based on the pixel values ​​of the difference image.

[0027] -Invention Effects-

[0028] According to the above method, the image processing apparatus, image processing method and image processing program of the present invention can detect the internal defect state of piping with higher accuracy compared with the prior art. Attached Figure Description

[0029] Figure 1 This is a block diagram illustrating the general structure of a radiographic imaging apparatus according to an exemplary embodiment.

[0030] Figure 2 This is a block diagram illustrating an example of the hardware structure of an image processing apparatus according to an exemplary embodiment.

[0031] Figure 3 This is a diagram used to illustrate the prior art; the left diagram shows... Figure 1 The image shown is an example of a transmission image obtained under the indicated conditions. The right image is a graph showing the distribution of brightness values ​​in the left-right direction of the transmission image shown in the left image.

[0032] Figure 4 This is a diagram used to illustrate the prior art, and is a side view showing an example of another photographic state based on a radiographic imaging device.

[0033] Figure 5 This is a diagram used to illustrate the prior art; the left diagram shows... Figure 4 The image shown is an example of a transmission image obtained under the indicated conditions. The right image is a graph showing the distribution of brightness values ​​in the left-right direction of the transmission image shown in the left image.

[0034] Figure 6 This is a block diagram illustrating an example of the functional structure of an image processing apparatus according to an exemplary embodiment.

[0035] Figure 7 This is a diagram illustrating an example of the original image, defect-free piping image, and differential image involved in an exemplary embodiment.

[0036] Figure 8 This is a diagram used to illustrate a method for generating defect-free piping images according to an exemplary embodiment.

[0037] Figure 9 This is a diagram illustrating an example of a visual screen related to an exemplary implementation.

[0038] Figure 10 This is a schematic diagram illustrating an example of the structure of an image information database involved in an exemplary implementation.

[0039] Figure 11This is a flowchart illustrating an example of image processing involved in an exemplary implementation.

[0040] Figure 12 This is a diagram illustrating another method for generating defect-free piping images according to an exemplary embodiment.

[0041] Figure 13 This is a diagram illustrating another method for generating defect-free piping images according to an exemplary embodiment.

[0042] Figure 14 This is a diagram illustrating another method for generating defect-free piping images according to an exemplary embodiment.

[0043] Figure 15 This is a diagram showing another example of a visual screen involved in an exemplary implementation.

[0044] Figure 16 This is a diagram showing another example of a visual screen involved in an exemplary implementation. Detailed Implementation

[0045] Hereinafter, with reference to the accompanying drawings, examples of methods for implementing the technology of the present invention will be described in detail. Furthermore, in this exemplary embodiment, the application of the image processing apparatus and image processing program according to the technology of the present invention to a radiographic imaging apparatus for non-destructive wall thickness reduction inspection (hereinafter referred to as "wall thickness reduction inspection") of a pipe made of metal or the like will be described.

[0046] First, refer to Figure 1 The structure of the radiographic imaging apparatus 1 according to this exemplary embodiment will be described. Figure 1 This is a block diagram showing the general structure of the radiographic imaging apparatus 1 according to the exemplary embodiment of this example.

[0047] like Figure 1 As shown, the radiation imaging apparatus 1 includes an image processing unit 10, a radiation source 12, and a radiation detector 14. The image processing unit 10, the radiation source 12, and the radiation detector 14 are connected in a manner that enables them to communicate with each other. The image processing unit 10 is, for example, a computer such as a personal computer or a server computer.

[0048] Radiation source 12, for example, irradiates pipe O with radiation R such as X-rays. In this exemplary embodiment, radiation source 12 irradiates pipe O with a cone-shaped beam of radiation R.

[0049] The radiation detector 14 includes a scintillator as an example of a light-emitting layer that emits light when irradiated by radiation R, and a TFT (Thin Film Transistor) substrate. The scintillator and the TFT substrate are stacked. The TFT substrate has a plurality of pixels arranged in a two-dimensional shape, and each pixel has a sensor section as an example of a conversion element that generates an increase in charge as the amount of irradiated radiation increases, and a field-effect thin-film transistor. The sensor section absorbs the light emitted by the scintillator to generate a charge and accumulates the generated charge. The field-effect thin-film transistor converts the charge accumulated in the sensor section into an electrical signal and outputs it. With the above structure, the radiation detector 14 generates a two-dimensional transmission image corresponding to the radiation R irradiated from the radiation source 12 onto the pipe O, and outputs the generated transmission image to the image processing device 10.

[0050] Thus, in the radiographic imaging apparatus 1, the transmitted image captured by irradiating the pipe O with radiation R from the radiation source 12 is stored in the image processing apparatus 10.

[0051] In this exemplary embodiment, the case of a radiation detector 14 using an indirect conversion method is described, but it is not limited to this method; a radiation detector 14 using a direct conversion method may also be used.

[0052] Next, refer to Figure 2 The hardware structure of the image processing apparatus 10 according to this exemplary embodiment will be described. Figure 2 This is a block diagram illustrating an example of the hardware structure of the image processing apparatus 10 according to this exemplary embodiment.

[0053] like Figure 2 As shown, the image processing apparatus 10 includes a CPU (Central Processing Unit) 20, a memory 21 serving as a temporary storage area, and a non-volatile storage unit 22. Furthermore, the image processing apparatus 10 includes a display 23 such as a liquid crystal display, an input device 24 such as a keyboard and mouse, and a network I / F (Interface) 25 connected to a network. The image processing apparatus 10 also includes an external I / F 26 connected to a radiation source 12 and a radiation detector 14. The CPU 20, memory 21, storage unit 22, display 23, input device 24, network I / F 25, and external I / F 26 are connected to a bus 27. The CPU 20 is an example of a processor according to the disclosed technology.

[0054] The storage unit 22 is implemented using an HDD (Hard Disk Drive), SSD (Solid State Drive), or flash memory, etc. The image processing program 30 is stored in the storage unit 22 as the storage medium. After the CPU 20 reads the image processing program 30 from the storage unit 22, it loads it into the memory 21 and executes the loaded image processing program 30.

[0055] Furthermore, an image information database 32 is stored in the storage unit 22. Details regarding the image information database 32 will be described later.

[0056] As an example, when inspecting pipe wall thickness reduction, such as... Figure 1 As shown, in the case of a photographic system that captures images of piping O from a distance, as an example, the following can be obtained: Figure 3 The left image shows a transmission image, and the distribution of brightness values ​​in this transmission image in the left-right direction is as follows: Figure 3 As shown in the right figure. Figure 3 This is a diagram used to illustrate the prior art; the left diagram shows... Figure 1 The image shown is an example of a transmission image obtained under the indicated conditions. The right image is a graph showing the distribution of brightness values ​​in the left-right direction of the transmission image shown in the left image.

[0057] As an example, such as Figure 3 As shown, in this imaging system, the areas on the sides of the pipe appear nearly white due to their low transmittance, while the areas inside the pipe appear nearly black because the transmittance increases closer to the center. On the other hand, the background area in the aforementioned transmission image appears closest to black because it has the highest transmittance.

[0058] Therefore, in this photographic system, the wall thickness reduction state of the side of the pipe O can be detected or measured by focusing on the thickness of the white pixels. Furthermore, within the internal region of the pipe O, the wall thickness reduction state can also be detected by the change in pixel value compared to surrounding pixels.

[0059] On the other hand, in situations where the space available for shooting is relatively narrow, for example, such as Figure 4 The image shown is a photographic system that takes pictures from the vicinity of pipe O. In this photographic system, as an example, the image can be obtained... Figure 5 The left image shows a transmission image, and the distribution of brightness values ​​in this transmission image in the left-right direction is as follows: Figure 5 As shown in the right figure. Figure 4 This is a diagram used to illustrate the prior art, a side view showing an example of another photographic state of the radiographic imaging apparatus 1. Furthermore, Figure 5 This is a diagram used to illustrate the prior art; the left diagram shows... Figure 4The image shown is an example of a transmission image obtained under the indicated conditions. The right image is a graph showing the distribution of brightness values ​​in the left-right direction of the transmission image shown in the left image.

[0060] As an example, such as Figure 5 As shown, in this photographic system, the wall thickness reduction state of the internal region of the pipe O can be detected by focusing on the change in pixel value compared with the surrounding pixels.

[0061] Thus, in the transmission image, the area on the side of pipe O is close to white, while the background area is closest to black, making it easy to detect wall thinning in this area. However, in the transmission image, the internal area of ​​pipe O becomes closer to black towards the center, creating a gradient effect, making it difficult to detect wall thinning.

[0062] The image processing apparatus 10 according to this exemplary embodiment performs image processing to solve the above-mentioned problems. Next, refer to Figure 6 The functional structure of the image processing apparatus 10 according to this exemplary embodiment will be described. Figure 6 This is a block diagram illustrating an example of the functional structure of the image processing apparatus 10 according to this exemplary embodiment.

[0063] like Figure 6 As shown, the image processing apparatus 10 includes an acquisition unit 20A, a generation unit 20B, a noise removal unit 20C, and a visualization unit 20D. The image processing program 30 is executed by the CPU 20, and the apparatus functions as the acquisition unit 20A, the generation unit 20B, the noise removal unit 20C, and the visualization unit 20D.

[0064] The acquisition unit 20A in this exemplary embodiment acquires a transmission image of the pipe O to be processed and an image of the defect-free pipe O corresponding to the transmission image, which is the same as the perspective of the transmission image, i.e., the defect-free pipe image.

[0065] Furthermore, in this exemplary embodiment, the image is obtained by reading the transmission image of the piping O from the storage unit 22, but this method is not limited to it. It is also possible to obtain the transmission image of the piping O by reading it from an external device connected to the network I / F 25. Moreover, in this exemplary embodiment, the image of the defect-free piping is obtained by simulation generation, but this method is not limited to it. For example, it is also possible to obtain the image of the defect-free piping by taking a transmission image of the piping O in an unused state using the radiographic imaging device 1, storing it in the storage unit 22, and then reading the transmission image from the storage unit 22.

[0066] Then, the generation unit 20B according to this exemplary embodiment generates a difference image representing the difference between the transmission image (hereinafter also referred to as the "original image") acquired by the acquisition unit 20A and the defect-free piping image.

[0067] In this exemplary embodiment, as an example, such as Figure 7 As shown, the difference image 44 is generated by dividing the original image 40 by the defect-free piping image 42 for each corresponding pixel, but this method is not limited to it. For example, it can also be configured to generate the difference image 44 by dividing the original image 40 by the defect-free piping image 42 for each corresponding pixel, or it can be configured to generate the difference image 44 by calculating the difference between the original image 40 and the defect-free piping image 42 for each corresponding pixel (subtraction operation). Figure 7 This is a diagram illustrating an example of the original image 40, the defect-free piping image 42, and the differential image 44 involved in this exemplary embodiment.

[0068] Here, in this exemplary embodiment, the acquisition unit 20A generates a defect-free piping image 42 by averaging the pixel values ​​of the transmission image (original image 40) along the axial direction of the piping O. More specifically, as an example, such as Figure 8 As shown, the region of pipe O in the original image 40 is divided into multiple parts along the axial direction of pipe O. For each part obtained in this way, the average value of the pixel value is calculated along the axial direction. Then, the calculated average value is assigned as the pixel value of each pixel in the axial direction of the corresponding part image. By stitching together the parts obtained in this way, a defect-free pipe image 42 with the same size as the original transmission image is generated. Figure 8 This is a diagram used to illustrate the method for generating a defect-free piping image 42 according to this exemplary embodiment.

[0069] On the other hand, the noise removal unit 20C according to this exemplary embodiment removes noise by filtering the difference image 44. In this exemplary embodiment, a Gaussian filter is applied as the filter used in the filtering process, but it is not limited to this method. For example, a medium filter may also be used as the filter used in the filtering process.

[0070] Then, the visualization unit 20D according to this exemplary embodiment visualizes the level of defects (wall thinning in this exemplary embodiment) in the piping O based on the pixel values ​​of the differential image 44. In the visualization unit 20D according to this exemplary embodiment, visualization is performed by creating a heat map based on the pixel values ​​of the differential image 44. More specifically, the heat map is created in such a way that the larger the pixel value of each pixel in the differential image 44, that is, the larger the difference, the closer it is to red. Hereinafter, the visualized image will be referred to as a "visualized image", and the screen on the display 23 that displays the visualized image will be referred to as a "visualized screen".

[0071] exist Figure 9 The diagram shows an example of a visual screen representing the exemplary embodiment described herein. For example... Figure 9 As shown, in this visualization, the differential image 44 and the visualization image 46, which is a heat map corresponding to the differential image 44, are displayed side by side on the display 23. Therefore, by referring to this visualization, the user can intuitively grasp the wall thickness reduction status of the pipe O.

[0072] Next, refer to Figure 10 The image information database 32 involved in this exemplary embodiment will be described. Figure 10 This is a schematic diagram illustrating an example of the structure of the image information database 32 involved in this exemplary embodiment.

[0073] The image information database 32 involved in this exemplary embodiment is a database that registers information representing the aforementioned transmission image. For example... Figure 10 As shown, the image information database 32 involved in this exemplary embodiment stores various information such as piping ID (Identification: identification code) and image information.

[0074] The aforementioned piping ID is information pre-assigned to each piping O for the purpose of individually identifying the piping O corresponding to the radiographic imaging device 1. Furthermore, the aforementioned image information is information representing the image information itself of the aforementioned transmission image.

[0075] Next, refer to Figure 11 The operation of the image processing apparatus 10 according to this exemplary embodiment will be explained. The image processing program 30 is executed by the CPU 20 of the image processing apparatus 10. Figure 11 The image processing shown. Figure 11 The image processing shown is performed, for example, when a user of the image processing device 10 inputs a command indicating the intended action. Furthermore, to avoid complexity, the case where the image information database 32 has already been constructed will be described here. Also, the case where the pipe ID of the pipe O to be processed is pre-specified will be described here.

[0076] exist Figure 11 In step S100, the CPU 20 obtains the image information corresponding to the pre-specified piping ID by reading it from the image information database 32. In step S102, the CPU 20 uses the read image information to generate a defect-free piping image 42 as described above.

[0077] In step S104, the CPU 20 uses the acquired image information, namely, the image information representing the original image 40 of the piping O to be processed and the image information representing the acquired defect-free piping image 42, to generate image information representing the differential image 44 as described above. In step S106, the CPU 20 performs the above-described filtering process on the image information representing the generated differential image 44.

[0078] In step S108, the CPU 20 creates a visualization image 46 as described above. Using the visualization image 46 and the filtered difference image 44, it performs the visualization process described above, displaying the visualization on the display 23, and then ends the image processing. As an example, through this visualization process, as described above... Figure 9 The visualization shown is displayed on monitor 23.

[0079] As explained above, the image processing apparatus according to this exemplary embodiment acquires a transmission image of the piping to be processed and an image of the defect-free piping corresponding to the transmission image, i.e., a defect-free piping image, and generates a difference image representing the difference between the transmission image and the defect-free piping image. Therefore, by using the generated difference image, the wall thinning state inside the piping can be detected with higher accuracy compared to the prior art.

[0080] Furthermore, according to the image processing apparatus of this exemplary embodiment, the defect-free piping image is set as a simulated image. Therefore, compared to the case where an actual image is used as the defect-free piping image, a defect-free piping image can be obtained more easily.

[0081] Furthermore, according to the image processing apparatus of this exemplary embodiment, a defect-free piping image is generated by averaging the pixel values ​​of the transmitted image along the axial direction of the piping. Therefore, compared to the case where the user specifies the defect-free portion of the piping, it is easier to generate a defect-free piping image.

[0082] Furthermore, according to the image processing apparatus of this exemplary embodiment, noise is removed by filtering the differential image. Therefore, compared to the case where the differential image is not filtered, the wall thickness reduction state can be detected with higher accuracy.

[0083] Furthermore, according to the image processing apparatus of this exemplary embodiment, the level of defects in the piping (wall thickness reduction in this exemplary embodiment) is visualized based on the pixel values ​​of the differential image. Therefore, by referring to the visualized level of wall thickness reduction, the state of wall thickness reduction can be grasped more intuitively.

[0084] In particular, the image processing apparatus according to this exemplary embodiment visualizes the image by creating a heatmap based on the pixel values ​​of the differential image. Therefore, the state of wall thickness reduction can be assessed using the heatmap.

[0085] Furthermore, in the exemplary embodiment described above, the case where a defect-free piping image 42 is generated by averaging the pixel values ​​of the transmission image (original image 40) along the axial direction of the piping O has been explained, but this method is not limited to it. For example, the defect-free piping image 42 may be generated based on at least one of the positions and regions on the transmission image received from the user. More specifically, as an example, such as Figure 12 As shown, a transmission image (original image 40) of the object is displayed on the monitor 23, and the user specifies at least one of the locations and areas in the displayed transmission image where wall thickness reduction is not considered to have occurred (in... Figure 12 (The example shown is a region). Next, the region specified by the user is used as a reference region, and the average pixel value is calculated in the axial direction for this reference region. Then, the calculated average value is assigned as the pixel value of each pixel in the axial direction of the reference region. By connecting the image of the reference region thus obtained in the axial direction to the same size as the original transmission image, a defect-free piping image 42 of the same size as the original transmission image is generated. Figure 12 This is a diagram illustrating another method for generating the defect-free piping image 42 according to the above exemplary embodiment.

[0086] According to this method, a defect-free piping image 42 can be generated more accurately than the case where a defect-free piping image 42 is generated by using calculations of a transmission image.

[0087] Furthermore, for example, it can be configured to generate a defect-free piping image 42 by mimicking a transmission image. As examples of this method, a method of generating a defect-free piping image 42 using three-dimensional information representing the shooting environment and information representing the size of the piping O is illustrated (hereinafter referred to as "the first method"). A method of generating a defect-free piping image 42 by means of machine learning is also illustrated (hereinafter referred to as "the second method").

[0088] exist Figure 13 The figure shown illustrates the method for generating the defect-free piping image 42 involved in the first method.

[0089] In this method, the positions and orientations of the radiation source 12, the piping O, and the radiation detector 14, as well as the dimensions of the piping O, are used as an example. Figure 13 The simulated photographic environment shown is used to derive the distance of the portion of the wall through which radiation penetrates the piping O. The longer this distance, the closer the pixels in the defect-free piping image 42 are to white. Therefore, the defect-free piping image 42 is generated by converting this distance into pixel values ​​for each pixel.

[0090] In addition, as a variation of this method, it can be configured to generate a defect-free piping image 42 based on the radiation source 12, piping O and radiation detector 14 positions and orientations and the size of piping O, in addition to the above-mentioned radiation source 12, piping O and radiation detector 14 positions and orientations and the size of piping O.

[0091] exist Figure 14 The figure shown illustrates the method for generating the defect-free piping image 42 involved in the second method.

[0092] In this approach, as an example, such as Figure 14 As shown, an existing transmission image (original image 40) is used as input information, and a defect-free transmission image within that transmission image is used as output information to perform machine learning, thereby pre-creating an AI (Artificial Intelligence) model. Then, in this method, a defect-free piping image 42 is generated by inputting the transmission image as the processing object into the AI ​​model. Furthermore, examples of machine learning models that can be applied at this time include autoencoders, GANs (Generative Adversarial Networks), etc.

[0093] Based on these methods, it is easier to generate a defect-free piping image 42 compared to having the user specify a defect-free portion of the piping O.

[0094] Furthermore, in the exemplary embodiment described above, visualization was performed by creating a heatmap based on the pixel values ​​of the difference image 44, but this method is not the only limitation. For example, as an example, Figure 15 As shown, it can be configured to emphasize the wall thickness reduction by adjusting at least one of the brightness and contrast of the differential image 44, thereby making it visual. Figure 15 This is a diagram showing another example of a visual screen involved in this exemplary embodiment.

[0095] According to this method, visualization is easier to perform compared to visualization using heatmaps.

[0096] And, for example, as an example, such as Figure 16As shown, it can be visualized by binarizing the difference image 44 to emphasize the wall thickness reduction. Figure 16 This is a diagram showing another example of a visual screen involved in this exemplary embodiment.

[0097] According to this method, visualization is easier compared to adjusting at least one of the brightness and contrast of the differential image 44.

[0098] Furthermore, while the exemplary embodiments described above illustrate the application of wall thickness reduction as a defect in the technology of the present invention, this approach is not limited to this method. For example, other defects such as cracks or blockages in the piping O may be considered as defects in the technology of the present invention.

[0099] Furthermore, in the exemplary embodiments described above, for example, the hardware structure of the processing unit that performs various processes such as the acquisition unit 20A, the generation unit 20B, the noise removal unit 20C, and the visualization unit 20D can use various processors as shown below. As mentioned above, among the various processors described above, in addition to the general-purpose processor, i.e., the CPU, which executes software (program) and functions as various processing units, there are also processors whose circuit structure can be changed after manufacturing FPGAs (Field Programmable Gate Arrays), i.e., programmable logic devices (PLDs), ASICs (Application Specific Integrated Circuits), and other processors with specially designed circuit structures for performing specific processes, i.e., dedicated circuits.

[0100] A processing unit can consist of one of these various processors, or it can consist of a combination of two or more processors of the same or different types (e.g., a combination of multiple FPGAs or a combination of a CPU and an FPGA). Furthermore, multiple processing units can also be composed of a single processor.

[0101] As examples of a single processor comprising multiple processing units, firstly, in client and server computers, a combination of one or more CPUs and software constitutes a single processor, which performs the functions of multiple processing units. Secondly, in systems-on-a-chip (SoCs), a processor that implements the functions of the entire system, including multiple processing units, is used on a single integrated circuit (IC) chip. Thus, various processing units are constructed using one or more of these different processors as their hardware architecture.

[0102] Furthermore, as the hardware structure of these various processors, more specifically, circuits composed of combined semiconductor elements and other circuitry can be used.

[0103] Furthermore, in the exemplary embodiment described above, the image processing program 30 was pre-stored (installed) in the storage unit 22 of the image processing apparatus 10, but this is not a limitation. The image processing program 30 may also be provided on a recording medium such as a CD-ROM (Compact Disc Read Only Memory), DVD-ROM (Digital Versatile Disc Read Only Memory), or USB (Universal Serial Bus) memory. Additionally, the image processing program 30 may be downloaded from an external device via a network.

[0104] Based on the above records, the inventions described in the following notes can be understood.

[0105] [Postscript 1]

[0106] An image processing apparatus comprising at least one processor,

[0107] The processor performs the following processing:

[0108] Acquire a transmission image of the piping to be processed and an image of the defect-free piping corresponding to the transmission image, i.e., a defect-free piping image.

[0109] Generate a difference image representing the difference between the transmission image and the defect-free piping image; and

[0110] The severity of defects in the piping is visualized based on the pixel values ​​of the differential image.

[0111] [Postscript 2]

[0112] According to the image processing apparatus described in Appendix 1, wherein,

[0113] The defect-free piping image is a simulated image.

[0114] [Postscript 3]

[0115] According to the image processing apparatus described in Appendix 2, wherein,

[0116] The defect-free piping image is generated by averaging the pixel values ​​of the transmission image along the axial direction of the piping.

[0117] [Postscript 4]

[0118] According to the image processing apparatus described in Appendix 2, wherein,

[0119] The defect-free piping image is generated based on at least one of the location and region on the transmission image received from the user.

[0120] [Postscript 5]

[0121] According to the image processing apparatus described in Appendix 2, wherein,

[0122] The defect-free piping image is generated by mimicking the transmission image.

[0123] [Postscript 6]

[0124] The image processing apparatus according to any one of Appendices 1 to 5, wherein,

[0125] The processor removes noise by filtering the differential image.

[0126] [Postscript 7]

[0127] The image processing apparatus according to any one of Appendices 1 to 6, wherein,

[0128] The defect is a reduction in the wall thickness of the piping.

[0129] [Postscript 8]

[0130] According to the image processing apparatus described in Appendix 7, wherein,

[0131] The processor performs the visualization by creating a heatmap based on the pixel values ​​of the differential image.

[0132] [Postscript 9]

[0133] According to the image processing apparatus described in Appendix 7, wherein,

[0134] The processor performs the visualization by adjusting at least one of the brightness and contrast of the differential image to emphasize the wall thickness reduction.

[0135] [Postscript 10]

[0136] According to the image processing apparatus described in Appendix 7, wherein,

[0137] The processor performs the visualization by binarizing the differential image to emphasize the wall thickness reduction.

[0138] [Postscript 11]

[0139] The image processing apparatus according to any one of Appendices 1 to 10, wherein,

[0140] The defect-free piping image is an image with the same viewing angle as the transmission image.

[0141] [Postscript 12]

[0142] An image processing method, wherein the following processing is performed by a processor included in an image processing device:

[0143] Acquire a transmission image of the piping to be processed and an image of the defect-free piping corresponding to the transmission image, i.e., a defect-free piping image.

[0144] Generate a difference image representing the difference between the transmission image and the defect-free piping image; and

[0145] The severity of defects in the piping is visualized based on the pixel values ​​of the differential image.

[0146] [Postscript 13]

[0147] An image processing program that causes a processor in an image processing device to perform the following processing:

[0148] Acquire a transmission image of the piping to be processed and an image of the defect-free piping corresponding to the transmission image, i.e., a defect-free piping image.

[0149] Generate a difference image representing the difference between the transmission image and the defect-free piping image; and

[0150] The severity of defects in the piping is visualized based on the pixel values ​​of the differential image.

[0151] The full text of the Japanese application dated March 25, 2024, Japanese Special Application 2024-047956, is incorporated herein by reference.

[0152] All documents, patent applications and technical standards described in this specification, as well as the individual documents, patent applications and technical standards specifically described and incorporated herein by reference, are incorporated herein by reference.

Claims

1. An image processing apparatus comprising at least one processor, The processor performs the following processing: Acquire a transmission image of the piping to be processed and an image of the defect-free piping corresponding to the transmission image, i.e., a defect-free piping image. Generate a difference image representing the difference between the transmission image and the defect-free piping image; and The severity of defects in the piping is visualized based on the pixel values ​​of the differential image.

2. The image processing apparatus according to claim 1, wherein, The defect-free piping image is a simulated image.

3. The image processing apparatus according to claim 2, wherein, The defect-free piping image is generated by averaging the pixel values ​​of the transmission image along the axial direction of the piping.

4. The image processing apparatus according to claim 2, wherein, The defect-free piping image is generated based on at least one of the location and region on the transmission image received from the user.

5. The image processing apparatus according to claim 2, wherein, The defect-free piping image is generated by mimicking the transmission image.

6. The image processing apparatus according to claim 1 or 2, wherein, The processor removes noise by filtering the differential image.

7. The image processing apparatus according to claim 1 or 2, wherein, The defect is a reduction in the wall thickness of the piping.

8. The image processing apparatus according to claim 7, wherein, The processor performs the visualization by creating a heatmap based on the pixel values ​​of the differential image.

9. The image processing apparatus according to claim 7, wherein, The processor performs the visualization by adjusting at least one of the brightness and contrast of the differential image to emphasize the wall thickness reduction.

10. The image processing apparatus according to claim 7, wherein, The processor performs the visualization by binarizing the differential image to emphasize the wall thickness reduction.

11. The image processing apparatus according to claim 1 or 2, wherein, The defect-free piping image is an image with the same viewing angle as the transmission image.

12. An image processing method, wherein a processor included in an image processing apparatus performs the following processing: Acquire a transmission image of the piping to be processed and an image of the defect-free piping corresponding to the transmission image, i.e., a defect-free piping image. Generate a difference image representing the difference between the transmission image and the defect-free piping image; and The severity of defects in the piping is visualized based on the pixel values ​​of the differential image.

13. An image processing program that causes a processor in an image processing apparatus to perform the following processing: Acquire a transmission image of the piping to be processed and an image of the defect-free piping corresponding to the transmission image, i.e., a defect-free piping image. Generate a difference image representing the difference between the transmission image and the defect-free piping image; and The severity of defects in the piping is visualized based on the pixel values ​​of the differential image.

Images