State determination assistance device, state determination model generation device, ultrasonic inspection device, state determination assistance method, state determination model generation method, ultrasonic inspection method, state determination assistance program, state determination model generation program, ultrasonic inspection program, and recording medium

Abstract

To provide a state determination assistance device which enables easy determination of the state of a pipe under inspection.SOLUTION: A state determination assistance device disclosed herein comprises an inspection result acquisition unit for acquiring ultrasonic flaw inspection data of a pipe under inspection, and a state determination unit configured to determine the state of the pipe under inspection based on the ultrasonic flaw inspection data.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present disclosure relates to a condition judgment support device, a condition judgment model manufacturing device, an ultrasonic inspection device, a condition judgment support method, a condition judgment model manufacturing method, an ultrasonic inspection method, a condition judgment support program, a condition judgment model manufacturing program, an ultrasonic inspection program, and a recording medium. [Background technology]

[0002] In various plants and factories in the chemical industry, there are many pipes, and non-destructive testing is performed to inspect the pipes for corrosion, etc., using ultrasonic waves. For example, Patent Document 1 describes an inspection system using ultrasonic flaw detection, which includes a probe that emits ultrasonic waves and receives reflected waves, an amplifier circuit that amplifies the received signal, a memory, an inner surface reflected wave recognition means that recognizes reflected waves from the inner surface, an amplification degree conversion means that reduces the amplification degree of the amplifier circuit when the value amplified by the amplifier circuit becomes equal to or greater than a first threshold value so that the value of the amplified received signal becomes less than a second threshold value, and increases the amplification degree of the amplifier circuit from the reduced state for a predetermined time, an outer surface reflected wave recognition means that recognizes reflected waves from the outer surface, and a thickness detection means that detects the thickness of a wall. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent No. 5042153 Summary of the Invention [Problem to be solved by the invention]

[0004] In ultrasonic testing using an inspection system such as that in Patent Document 1, a specialized inspector must visually check the inspection results to determine the condition of the inspected pipe, such as whether the pipe has defects such as pitting corrosion or thinning, and to find out whether there are any defects and what type of defects there are. Therefore, to determine the condition of a pipe using an inspection system such as that in Patent Document 1, a specialized inspection engineer is required, and there is also the problem that it is time-consuming.

[0005] Therefore, an object of the present disclosure is to provide a condition determination support device, a condition determination model manufacturing device, an ultrasonic inspection device, a condition determination support method, a condition determination model manufacturing method, an ultrasonic inspection method, a condition determination support program, a condition determination model manufacturing program, an ultrasonic inspection program, and a recording medium that can easily determine the condition of a pipe to be inspected. [Means for solving the problem]

[0006] In order to achieve the above object, the state determination support device of the present disclosure comprises: The test result acquisition unit and the state determination unit are included. The inspection result acquisition unit acquires ultrasonic flaw detection inspection data of the inspection target pipe, The condition determining unit determines a condition of the inspection target pipe based on the ultrasonic flaw detection data.

[0007] The state determination model manufacturing device of the present disclosure comprises: The apparatus includes an inspection result acquisition unit, a preprocessing unit, and a trained model generation unit, The inspection result acquisition unit acquires ultrasonic flaw detection inspection data of the inspection target pipe, the ultrasonic inspection data includes image data; The preprocessing unit processes the image data to generate preprocessed data, The trained model generation unit generates, through machine learning using the preprocessed data as training data, a condition determination model that outputs a condition determination result for the inspection target pipe when the preprocessed data is input, as a trained model.

[0008] The ultrasonic inspection device of the present disclosure includes: A state determination unit is included, The state determination unit is the state determination support device of the present disclosure.

[0009] The state determination support method of the present disclosure includes: The method includes an inspection result acquisition step and a state determination step, The inspection result acquisition step acquires ultrasonic inspection data of the inspection target pipe, The state determining step determines a state of the inspection target pipe based on the ultrasonic flaw detection data.

[0010] The present disclosure relates to a method for producing a state determination model, The method includes an inspection result acquisition process, a preprocessing process, and a trained model generation process. The inspection result acquisition step acquires ultrasonic inspection data of the inspection target pipe, the ultrasonic inspection data includes image data; The preprocessing step generates preprocessed data by processing the image data, The trained model generation process generates, through machine learning using the preprocessed data as training data, a condition determination model that outputs a condition determination result for the inspection target pipe when the preprocessed data is input, as a trained model.

[0011] The ultrasonic inspection method of the present disclosure includes: A state determination step is included, The state determination step is the state determination support method of the present disclosure.

[0012] The state determination support program of the present disclosure is Includes a procedure for obtaining test results and a procedure for determining status, The inspection result acquisition step includes acquiring ultrasonic inspection data of the inspection target pipe, The state determination step determines a state of the inspection target pipe based on the ultrasonic flaw detection data. The program causes a computer to execute each of the above procedures.

[0013] The state determination model manufacturing program disclosed herein is This includes a procedure for obtaining inspection results, a preprocessing procedure, and a procedure for generating a trained model. The inspection result acquisition step includes acquiring ultrasonic inspection data of the inspection target pipe, the ultrasonic inspection data includes image data; The pre-processing step generates pre-processed data by processing the image data; The trained model generation step generates, as a trained model, a condition determination model that outputs a condition determination result of the inspection target pipe when the preprocessed data is input, by machine learning using the preprocessed data as training data. The program causes a computer to execute each of the above procedures.

[0014] The ultrasonic inspection program of the present disclosure includes: Including a state determination procedure, The state determination procedure is a program executed by the state determination support program of the present disclosure.

[0015] The recording medium of the present disclosure is a computer-readable recording medium having recorded thereon the condition judgment support program, the condition judgment model manufacturing program, or the ultrasonic inspection program of the present disclosure. Effect of the Invention

[0016] According to the present disclosure, the condition of the tube being inspected can be easily determined. [Brief description of the drawings]

[0017] [Figure 1] FIG. 1 is a block diagram showing a configuration of an example of a state determination support device according to the present disclosure. [Diagram 2] FIG. 2 is a block diagram showing an example of a hardware configuration of the state determination support device of the present disclosure. [Diagram 3] FIG. 3 is a flowchart showing an example of processing in the state determination support device of the present disclosure. [Figure 4] FIG. 4 is a schematic diagram for explaining the process of the test result acquisition unit. [Diagram 5] FIG. 5 is a schematic diagram for explaining the process of the state determination unit. [Figure 6] FIG. 6 is a schematic diagram for explaining the process of the state determination unit. [Figure 7] FIG. 7 is a block diagram showing a configuration of an example of a state determination support device of the present disclosure. [Figure 8]FIG. 8 is a flowchart showing an example of processing in the state determination support device of the present disclosure. [Figure 9] FIG. 9 is a diagram showing an example of an image generated by the image generating unit. [Figure 10] FIG. 10 is a block diagram showing a configuration of an example of a state determination model manufacturing device according to the present disclosure. [Figure 11] FIG. 11 is a block diagram showing an example of a hardware configuration of a state determination model producing device according to the present disclosure. [Figure 12] FIG. 12 is a flowchart showing an example of processing in the state determination model manufacturing device of the present disclosure. [Figure 13] FIG. 13 is a diagram showing the performance index of the state determination model. [Figure 14] FIG. 14 is a diagram showing the performance index of the state determination model. [Figure 15] FIG. 15 is a diagram showing the performance index of the state determination model. [Figure 16] FIG. 16 is a schematic diagram showing an example of the configuration of an ultrasonic inspection device according to the present disclosure. [Figure 17] FIG. 17 is a schematic diagram showing an example of the configuration of a probe device in an ultrasonic inspection device according to the present disclosure. [Figure 18] FIG. 18(a) is a schematic diagram showing an example of a liquid flow in the probe device of the present disclosure, and FIG. 18(b) is a schematic diagram showing an example of the configuration of a reflecting mirror and a rotation mechanism of the probe device. [Figure 19] FIG. 19 is a schematic diagram showing an example of ultrasonic inspection of a circumferential flaw in a pipe to be inspected. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0018] Next, an embodiment of the present disclosure will be described with reference to the drawings. The present disclosure is not limited to the following embodiment. In each of the following drawings, the same parts are given the same reference numerals. In addition, the description of each embodiment can be mutually incorporated unless otherwise specified, and the configurations of each embodiment can be combined unless otherwise specified.

[0019] [Embodiment 1] Fig. 1 is a block diagram showing an example of the configuration of a state determination support device 10 according to the present embodiment. As shown in Fig. 1, the device 10 includes an inspection result acquisition unit 11 and a state determination unit 12. The device 10 may also include a storage unit, although not shown. Although not shown, the device 10 may also include, for example, an input unit, an output unit, a display unit, and / or a storage unit.

[0020] The device 10 may be, for example, a single device including each of the above-mentioned units, or a device to which each of the above-mentioned units can be connected via a communication line network. The device 10 may be connected to an external device described later via the communication line network. The communication line network is not particularly limited, and a known network may be used, for example, wired or wireless. Examples of the communication line network include an Internet line, a World Wide Web (WWW), a telephone line, a Local Area Network (LAN), a Storage Area Network (SAN), a Delay Tolerant Networking (DTN), a Low Power Wide Area (LPWA), and a Local 5G (L5G). Examples of wireless communication include Wi-Fi (registered trademark), Bluetooth (registered trademark), a Local 5G, and a LPWA. Examples of the wireless communication may include a form in which each device directly communicates with each other (Ad Hoc communication), an infrastructure (infrastructure communication), and an indirect communication via an access point. The device 10 may be incorporated into a server as a system, for example. Furthermore, the present device 10 may be, for example, a personal computer (PC, for example, desktop type or notebook type) in which the program of the present disclosure is installed, a smartphone, a tablet terminal, etc. The present device 10 may be in the form of cloud computing or edge computing, for example, in which at least one of the above-mentioned units is on a server and the other units are on a terminal.

[0021] 2 illustrates a block diagram of a hardware configuration of the present device 10. The present device 10 includes, for example, a central processing unit (CPU, GPU, etc.) 101, a memory 102, a bus 103, a storage device 104, an input device 105, an output device 106, a communication device 107, etc. Each part of the present device 10 is connected to each other via the bus 103 by each interface (I / F).

[0022] The central processing unit 101 cooperates with other components through a controller (such as a system controller or an I / O controller) and controls the entire device 10. In the device 10, the central processing unit 101 executes, for example, the program of the present disclosure (the state determination support program) and other programs, and also reads and writes various information. Specifically, for example, the central processing unit 101 functions as an inspection result acquisition unit 11 and a state determination unit 12. The device 10 may include, as a calculation device, other calculation devices such as a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), an APU (Accelerated Processing Unit), or a combination of these.

[0023] The bus 103 can also be connected to, for example, an external device. Examples of the external device include a state determination model manufacturing device disclosed herein, which will be described later, an ultrasonic inspection device disclosed herein, an external storage device (external database, etc.), a printer, an external input device, an external display device, an external imaging device, etc. The present device 10 can be connected to an external network (the above-mentioned communication line network) by, for example, a communication device 107 connected to the bus 103, and can also be connected to other devices via the external network.

[0024] An example of the memory 102 is a main memory (primary storage device). When the central processing unit 101 performs processing, the memory 102 reads various operation programs, such as the program of the present disclosure, stored in the storage device 104 described below, and the central processing unit 101 receives data from the memory 102 and executes the program. The main memory is, for example, a RAM (random access memory). The memory 102 may also be, for example, a ROM (read only memory).

[0025] The storage device 104 is also called an auxiliary storage device, for example, in contrast to the main memory (primary storage device). As described above, the storage device 104 stores an operating program including the program of the present disclosure. The storage device 104 may be, for example, a combination of a recording medium and a drive that reads and writes from and to the recording medium. The recording medium is not particularly limited, and may be, for example, an internal or external type, such as a HD (hard disk), a CD-ROM, a CD-R, a CD-RW, an MO, a DVD, a flash memory, a memory card, and the like. The storage device 104 may be, for example, a hard disk drive (HDD) in which a recording medium and a drive are integrated, and a solid state drive (SSD). When the present device 10 includes the storage unit, for example, the storage device 104 functions as the storage unit. The storage unit can store, for example, information of the user of the present device (name, identification number, attribute information (gender, age, organization, address, contact information, etc.), biological information, etc.), ultrasonic flaw inspection data, image data, preprocessing data, state determination model, determination result, etc.

[0026] In the present device 10, the memory 102 and the storage device 104 can also store various information such as log information, information acquired from an external database (not shown) or an external device, information generated by the present device 10, and information used when the present device 10 executes processing. In this case, the memory 102 and the storage device 104 may store, for example, the above-mentioned information on the user of the present device. Note that at least a portion of the information may be stored, for example, in an external server other than the memory 102 and the storage device 104, or may be distributed and stored in multiple terminals using block chain technology or the like.

[0027] The device 10 further includes, for example, an input device 105 and an output device 106. The input device 105 may be, for example, a pointing device such as a touch panel, a track pad, or a mouse; a keyboard; an imaging means such as a camera or a scanner; a card reader such as an IC card reader or a magnetic card reader; and an audio input means such as a microphone. The output device 106 may be, for example, a display device such as an LED display or a liquid crystal display; an audio output device such as a speaker; a printer; or the like. In the first embodiment, the input device 105 and the output device 106 are configured separately, but the input device 105 and the output device 106 may be configured as an integrated device, such as a touch panel display.

[0028] Next, an example of the state determination support method of this embodiment will be described with reference to the flowchart of Fig. 3. The state determination support method of this embodiment is implemented as follows, for example, using the state determination support device 10 of Fig. 1 or Fig. 2. Note that the state determination support method of this embodiment is not limited to use of the state determination support device 10 of Fig. 1 or Fig. 2.

[0029] First, the inspection result acquisition unit 11 acquires ultrasonic inspection data of the inspection target pipe (S1, inspection result acquisition step). The ultrasonic inspection data is, for example, data showing the result of an ultrasonic inspection test performed on the inspection target pipe. The ultrasonic inspection test may be, for example, a pipe insertion ultrasonic inspection test using an internal rotating ultrasonic inspection system (IRIS), which is called a water immersion UT test or the like. The ultrasonic inspection data includes, for example, image data. The image data is, for example, a contour image including at least one color. The inspection result acquisition unit 11 can acquire ultrasonic inspection data including the image data by, for example, cutting out an area of ​​the contour image from the ultrasonic inspection data output by an ultrasonic inspection device. The contour image is, for example, also called a color contour map. The color contour map is, for example, an image in which the state of the inspection target pipe is expressed by a plurality of colors. As a specific example, the color contour map includes, for example, blue, green, yellow, and red areas. In the color contour map, the blue region indicates no thinning (also called normal region), the green region indicates a light thinning region, the yellow region indicates a medium thinning region, and the red region indicates a heavy thinning region. The "light", "moderate", and "heavy" may be, for example, a relative index (for example, the region with the largest amount of thinning is heavy, the region with the second largest amount of thinning is medium, and the region with thinning but with a relatively small amount of thinning is light) or an absolute index (for example, a heavy thinning threshold, a medium thinning threshold, and a light thinning threshold are set, and when each threshold is exceeded, it is determined as heavy, medium, or light). Note that the combination of the color in the color contour map and the state of the pipe to be inspected is an example, and the present disclosure is not limited thereto. Specifically, the inspection result acquisition unit 11 may, for example, directly acquire ultrasonic flaw inspection data from an ultrasonic inspection device, or may acquire the ultrasonic flaw inspection data from an external database or server in which ultrasonic flaw inspection data measured by the ultrasonic inspection device is stored. In the former case, the ultrasonic inspection device may be, for example, an ultrasonic inspection device disclosed in the present disclosure, which will be described later, or may be another ultrasonic inspection device. The processing by the inspection result acquisition unit 11 may be performed, for example, using a publicly known image processing library.The image processing library may be, for example, Open CV.

[0030] The ultrasonic inspection data may include, for example, other data. The other data may include various information output by the ultrasonic inspection device. As a specific example, the ultrasonic inspection data may include, for example, a pipe cross-sectional image. The pipe cross-sectional image is, for example, an image showing a cross-section of the inspection target pipe output by the ultrasonic inspection device. The inspection result acquisition unit 11 may further acquire, for example, the pipe cross-sectional image. In this case, the inspection result acquisition unit 11 can acquire the pipe cross-sectional image by, for example, cutting out the area of ​​the pipe cross-sectional image from the ultrasonic inspection data.

[0031] The processing of the inspection result acquisition unit 11 will be specifically described with reference to Fig. 4. Fig. 4(A) is a schematic diagram showing an example of ultrasonic inspection data 110. As shown in Fig. 4(A), the ultrasonic inspection data 110 includes, for example, a contour image (color contour map) 111 including at least one color, and a pipe cross-sectional image 112. The inspection result acquisition unit 11 can acquire the color contour map 111 and the pipe cross-sectional image 112 shown in Fig. 4(B) by, for example, cutting out an area surrounded by a black frame in the ultrasonic inspection data 110 shown in Fig. 4(A).

[0032] The state determination unit 12 determines the state of the inspection target pipe based on the ultrasonic inspection data (S2, state determination step). The "state" of the inspection target pipe may be, for example, a parameter of the inspection target pipe (e.g., thickness), the presence or absence of a defect in the inspection target pipe, the type of defect in the inspection target pipe, or the state of the defect. The state determination unit 12 can determine, for example, the thickness of the inspection target pipe, the presence or absence of corrosion, pitting, or thinning of the inspection target pipe, the thickness at the defective portion of the inspection target pipe, etc., based on the ultrasonic inspection data. The state determination unit 12 may determine the state of the inspection target pipe by, for example, image analysis processing, or may determine the state of the inspection target pipe by using a machine learning model.

[0033] When making a judgment using the machine learning model, the condition judgment unit 12 can, for example, input the ultrasonic flaw detection inspection data into a condition judgment model to judge the condition of the inspection target pipe. The state determination model is, for example, a model (AI) that has been trained to output a state determination result of the inspection target pipe when the ultrasonic inspection data is input. The ultrasonic inspection data input to the state determination model may be, for example, image data included in the ultrasonic inspection data, and may further be pre-processed data described later. The state determination result may be, for example, the presence or absence of a defective portion (e.g., a corroded portion, a pitting portion, a thinning portion, a damaged portion such as a crack, etc.) in the inspection target pipe; the type of defect such as corrosion, pitting, thinning, or crack; the location of defect occurrence; the probability of defect occurrence; and the like. The state determination model may be, for example, a model constructed by a state determination model manufacturing device of the present disclosure described later. The state determination model may be, for example, stored in advance in the storage unit of the present device 10, or may be stored outside the present device 10.

[0034] The state determination model includes, for example, an input layer for inputting ultrasonic inspection data, an output layer for outputting defects in the inspection target pipe, and at least one intermediate layer provided between the input layer and the output layer. The state determination model may be a program module that is a part of artificial intelligence software. An example of the multi-layered network is a neural network. An example of the neural network is a convolution neural network (CNN), but is not limited to CNN, and may be a trained model constructed with other learning algorithms such as a neural network other than CNN, a support vector machine (SVM), a Bayesian network, and a regression tree. For example, Faster R-CNN can be used as a specific example of an object detection network using CNN. Faster R-CNN has high estimation accuracy in object detection, and therefore, it is expected to improve estimation accuracy.

[0035] The condition determination model can be generated by machine learning using, for example, the ultrasonic flaw detection data and labels for the pre-processed data (for example, a label indicating a defect in the pipe to be inspected and a label indicating noise) as training data. The generation of the condition determination model will be described later. The condition determination model may be, for example, a trained model generated in advance. The trained model may also be a trained model (derived model) that has been retrained using the training data and a trained model that has already been generated. Furthermore, the trained model may be a trained model obtained by transfer learning using a trained model generated using the training data, or a trained model generated by model compression of a trained model generated using the training data.

[0036] When performing the state determination by the image analysis process, the state determination unit 12 extracts, for example, characteristic information included in the ultrasonic inspection data by image analysis process. When the ultrasonic inspection data includes image data and the image data is the color contour map, the characteristic information may include, for example, color information included in the color contour map. The state determination unit 12 can determine, for example, based on the extracted color for each pixel of the image data, that the corresponding portion of the inspection target pipe is in a state corresponding to the color. The state determination can be performed, for example, by referring to a state determination database in which types of colors are associated with types of states (for example, defects). The state determination database is, for example, a database in which the state of the inspection target pipe (for example, the presence or absence of defects such as a corroded portion, a pitting portion, a thinning portion, or a damaged portion such as a crack; types of defects such as corrosion, pitting, thinning, or crack; positions of defects; probability of defects; thickness of the pipe) is associated with the color of the image data (color contour map) included in the ultrasonic inspection data. In the state determination database, for example, it is preferable that the labels of the red area (pcred) of the pitting area, the yellow area (pcyellow) of the pitting area, the red area (cwred) of the thinning area, the yellow area (cwyellow) of the thinning area, and the noise (noise) are associated with the color features (color, area, etc.). In the state determination database, for example, the color and a threshold value of the area of ​​the area including the color may be further linked. In this case, the state determination unit 12 can determine the state of the inspection target pipe by, for example, comparing the color and the area of ​​the area including the color in the preprocessing data with the state determination database. The state determination database may be stored in, for example, a storage unit of the present device 10, or may be stored outside the present device 10.

[0037] When the ultrasonic inspection data includes the pipe cross-sectional image, the state determination unit 12 may estimate a pipe cross-sectional profile based on the pipe cross-sectional image. The pipe cross-sectional profile is, for example, information indicating an estimated defect location in the pipe cross-sectional image. The state determination unit 12 may estimate the pipe cross-sectional profile from the pipe cross-sectional image by, for example, image analysis processing, or may estimate the pipe cross-sectional profile using machine learning. In the former case, the state determination unit 12 can detect the shape of the inspection target pipe shown in the pipe cross-sectional image by, for example, object detection processing, and estimate the pipe cross-sectional profile based on the detected shape. In this case, the state determination unit 12 can estimate the pipe cross-sectional profile by, for example, referring to a database in which the shape of the inspection target pipe shown in the pipe cross-sectional image and the pipe cross-sectional profile are associated with each other. In the latter case, the state determination unit 12 can estimate the pipe cross-sectional profile by, for example, applying semantic segmentation (SS) to the pipe cross-sectional image. The semantic segmentation can be performed by using a deep neural network (DNN) for semantic segmentation, such as a fully convolutional network (FCN) or a pyramid scene parsing network (PSPNet). The NN for semantic segmentation may be trained by teacher data that is labeled to remove noise (e.g., ultrasonic leakage) contained in the pipe cross-sectional image. Specifically, the state determination unit 12 applies semantic segmentation to the pipe cross-sectional image, for example, to distinguish the pipe cross-section from the background in the pipe cross-sectional image. Then, the locations of pitting corrosion and thinning can be estimated based on the shape of the identified pipe cross-sectional portion.

[0038] Furthermore, the state determination unit 12 may generate pre-processed data by processing the image data (pre-processing step), and determine the state of the inspection target pipe based on the pre-processed data. In this case, the state determination unit 12 generates split image data by, for example, dividing the image data into four at the positions of two straight lines intersecting at the center of the image data (splitting step). The split image data is, for example, image data divided into four at the center positions in the vertical and horizontal directions of the image data. The state determination unit 12 can then generate pre-processed data by, for example, doubling the vertical and horizontal dimensions of each of the split image data (enlarging step). A specific example of the pre-processed data generation process will be described with reference to FIG. 5. As shown in FIG. 5(A), for example, it is assumed that the image data 111 is image data with a size of 307 pixels vertically and 813 pixels horizontally. In this case, the state determination unit 12 divides the image data 111 into four at the position of the broken line shown in FIG. 5(A) as the dividing step, and generates divided image data 111a, 111b, 111c, and 111d each having a size of 406.5 pixels in width and 153.5 pixels in height as shown in FIG. 5(B). Then, the state determination unit 12 doubles the vertical size and horizontal size of each of the divided image data 111a, 111b, 111c, and 111d as the enlarging step, and generates pre-processed data having a size of 307 pixels in height and 813 pixels in width, and determines the state of the inspection target pipe based on the pre-processed data. FIG. 5(C) shows an example of pre-processed data 111A obtained by processing the divided image data 111a. The generation process of the pre-processed data by the state determination unit 12 may be performed using, for example, a known image processing library. The image processing library may be, for example, Open CV.

[0039] When the image data is a contour image including at least one color, the state determination unit 12 may further detect the at least one color in the contour image and perform an enhancement process on the area including the detected color (enhancement process step). The color detected by the state determination unit 12 is not particularly limited and may be any color. For example, the state determination unit 12 may detect a color (e.g., red, yellow, etc.) indicating a defect (e.g., a thinning portion, a pitting portion) of the inspection target pipe in the color contour map. In addition, the state determination unit 12 may perform the enhancement process on the area by drawing the periphery of the area including the color with a line of the same color as the detected color. By performing such a process by the state determination unit 12, for example, it is expected to improve the accuracy of judging a relatively small defect (e.g., pitting, thinning) portion in the inspection target pipe. A specific example of the enhancement process on the contour image will be described with reference to FIG. 6. First, as shown in FIG. 6(A), the state determination unit 12 performs color recognition processing on the contour image to detect a red region 1111 and a yellow region 1112. Then, as shown in FIG. 6(B), the state determination unit 12 draws the periphery of the detected red region 1111, i.e., the boundary between the red region 1111 and other colors, with a red frame 1111a, and draws the periphery of the detected yellow region 1112, i.e., the boundary between the yellow region 1112 and other colors, with a yellow frame 1112a. This makes it possible to highlight, for example, the yellow region and the red region indicating a defective portion, and improve the detection accuracy of a relatively small defective portion in the inspection target pipe. Note that the red region 1111, the yellow region 1112, the red frame 1111a, and the yellow frame 1112a shown in FIG. 6 are examples, and the present disclosure is not limited to the example of FIG. 6.

[0040] For example, when the condition determination unit 12 determines the condition of the inspection target pipe based on the contour image after the enhancement process, the condition determination unit 12 may further execute a verification process. The verification process is, for example, a process of reflecting the condition determination result (for example, the position of a defective portion, etc.) in the contour image before the enhancement process and verifying the difference in color before and after the enhancement process. This makes it possible to suppress, for example, erroneous determination based on the enhancement process.

[0041] The device 10 may, for example, output the state determination result. The output of the state determination result will be described later in the second embodiment.

[0042] In the state judgment support method of the present embodiment, the case where S1 to S2 are executed sequentially has been described as an example, but the present disclosure is not limited thereto. In the present disclosure, S1 to S2 may be executed simultaneously or separately, for example. In the latter case, the order of execution is not particularly limited and is arbitrary.

[0043] According to the condition determination support device of this embodiment, by inputting ultrasonic inspection data of the inspection target pipe, the condition of the inspection target pipe can be determined based on image data included in the ultrasonic inspection data. Therefore, according to the condition determination support device of this embodiment, for example, the condition of the inspection target pipe can be easily determined. In addition, for example, by connecting the condition determination support device of this embodiment to an ultrasonic inspection device, it becomes possible to automatically determine the condition of the inspection target pipe.

[0044] [Embodiment 2] This embodiment is similar to the state judgment support device 10 of embodiment 1, except that an image generation unit and an output unit are included in addition to the configuration of the state judgment support device 10 of embodiment 1, and the description thereof can be used. The state judgment support device of this embodiment includes, for example, an image generation unit and an output unit, and the image generation unit generates a judgment image by superimposing the judgment result of the state judgment unit on the ultrasonic flaw detection inspection data, and the output unit outputs the judgment image.

[0045] Fig. 7 is a block diagram showing an example of the configuration of a state determination support device 10A of this embodiment. As shown in Fig. 7, state determination support device 10A includes an image generation unit 13 and an output unit 14 in addition to the configuration of state determination support device 10 of embodiment 1. The hardware configuration of state determination support device 10A is similar to that of state determination support device 10 in Fig. 2, except that CPU 101 includes the configuration of state determination support device 10A in Fig. 7 instead of the configuration of state determination support device 10 in Fig. 1.

[0046] Next, the state determination support method of this embodiment will be described with reference to the flowchart of Fig. 8. The state determination support method of this embodiment can be implemented, for example, by using the state determination support device 10A of this embodiment shown in Fig. 7. Note that the state determination support method disclosed herein is not limited to use of the state determination support device 10A.

[0047] First, steps S1 and S2 are carried out in the same manner as steps S1 and S2 in the state determination support method of the first embodiment.

[0048] Next, the image generating unit 13 generates a judgment image by, for example, superimposing the judgment result of the state judgment unit on the ultrasonic inspection data (S3, image generating step). The image generating unit 13 can generate the judgment image by, for example, superimposing the judgment result on an image showing the ultrasonic inspection data generated by an ultrasonic inspection device. In addition, in S2, if a pipe cross-sectional profile is estimated, the image generating unit 13 can further generate a pipe cross-sectional profile estimated image based on the ultrasonic inspection data. The pipe cross-sectional profile estimated image is, for example, an image in which a defective portion estimated based on the pipe cross-sectional profile is superimposed on a pipe cross-sectional image of the inspection target pipe. In this case, the image generating unit 13 can generate the pipe cross-sectional profile estimated image by, for example, superimposing the pipe cross-sectional profile on the pipe cross-sectional image generated by an ultrasonic inspection device.

[0049] The output unit 14 outputs, for example, the judgment image (S4, output step). The output unit 14 may output (display) the judgment image to, for example, the output device (display) 106 of the state judgment support device 10A, or may output (display) the judgment image on a display device external to the device. Examples of the display device external to the device include a monitor provided in an ultrasonic inspection device, an external display, etc. Also, in the case where the pipe cross-sectional profile estimated image is generated in S3, the output unit 14 may output, for example, the pipe cross-sectional profile estimated image. The judgment image and the pipe cross-sectional profile estimated image may be output simultaneously or separately.

[0050] Fig. 9 shows an example of a determination image and a pipe cross-sectional profile estimation image generated by the image generation unit 13 and output by the output unit 14, but the present disclosure is not limited to the example of Fig. 9. The monitor (1) shown in the upper part of Fig. 9 is the screen of an ultrasonic inspection device (IRIS), and the monitor (2) shown in the lower part of Fig. 9 is the monitor (output device) 106 of the present device 10A. As shown in Fig. 9, the present device 10A can mirror (duplicate) ultrasonic flaw detection data including a contour image 111 and a pipe cross-sectional image 112 on the screen of the monitor (1) of the ultrasonic inspection device on the monitor (2), and can superimpose and display a determination image 113 and a pipe cross-sectional profile estimation image 112A on the ultrasonic flaw detection data.

[0051] In the state determination support method of the present embodiment, the case where S1 to S4 are executed sequentially has been described as an example, but the present disclosure is not limited thereto. In the present disclosure, S1 to S4 may be executed simultaneously or separately, for example. In the latter case, the order of execution is not particularly limited and is arbitrary.

[0052] The condition determination support device of this embodiment can output a determination image, which allows a user to easily determine the condition of the inspection target pipe, for example, by checking the output determination image.

[0053] [Embodiment 3] The third embodiment is an example of a state determination model manufacturing device according to the present disclosure.

[0054] The state judgment model manufacturing apparatus of this embodiment will be described with reference to Fig. 10. Fig. 10 is a block diagram showing an example of the configuration of the state judgment model manufacturing apparatus 20 of this embodiment. As shown in Fig. 10, the state judgment model manufacturing apparatus 20 includes an inspection result acquisition unit 21, a preprocessing unit 22, and a trained model generation unit 23. Although not shown, the state judgment model manufacturing apparatus 20 may also include, for example, a storage unit.

[0055] The device 20 may be, for example, a single device including each of the above-mentioned units, or a device to which each of the above-mentioned units can be connected via a communication line network. The device 20 may be connected to an external device described later via the communication line network. The communication line network is not particularly limited, and a known network may be used, for example, wired or wireless. Examples of the communication line network include an Internet line, a World Wide Web (WWW), a telephone line, a Local Area Network (LAN), a Storage Area Network (SAN), a Delay Tolerant Networking (DTN), a Low Power Wide Area (LPWA), and a Local 5G (L5G). Examples of wireless communication include Wi-Fi (registered trademark), Bluetooth (registered trademark), a Local 5G, and a LPWA. Examples of the wireless communication may include a form in which each device directly communicates with each other (Ad Hoc communication), an infrastructure (infrastructure communication), and an indirect communication via an access point. The device 20 may be incorporated into a server as a system, for example. Furthermore, the present device 20 may be, for example, a personal computer (PC, for example, desktop type or notebook type) in which the program of the present disclosure is installed, a smartphone, a tablet terminal, etc. The present device 20 may be in the form of cloud computing or edge computing, for example, in which at least one of the above-mentioned units is on a server and the other units are on a terminal.

[0056] 11 illustrates a block diagram of a hardware configuration of the present device 20. The present device 20 includes, for example, a central processing unit (CPU, GPU, etc.) 201, a memory 202, a bus 203, a storage device 204, an input device 205, an output device 206, a communication device 207, etc. Each part of the present device 20 is connected to each other via the bus 203 by each interface (I / F).

[0057] The central processing unit 201 cooperates with other components through a controller (such as a system controller or an I / O controller) and controls the entire device 20. In the device 20, the central processing unit 201 executes, for example, the program of the present disclosure (the state determination support program) and other programs, and also reads and writes various information. Specifically, for example, the central processing unit 201 functions as an inspection result acquisition unit 21, a preprocessing unit 22, and a trained model generation unit 23. The device 20 may include other calculation devices such as a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), an APU (Accelerated Processing Unit), or a combination of these as a calculation device.

[0058] The bus 203 can also be connected to, for example, an external device. Examples of the external device include the state determination support device disclosed herein, an ultrasonic inspection device disclosed herein later, an external storage device (external database, etc.), a printer, an external input device, an external display device, an external imaging device, etc. The present device 20 can be connected to an external network (the communication line network) by, for example, a communication device 207 connected to the bus 203, and can also be connected to other devices via the external network.

[0059] The memory 202 may be, for example, a main memory (primary storage device). When the central processing unit 201 performs processing, the memory 202 reads various operation programs, such as the program of the present disclosure, stored in the storage device 204 described below, and the central processing unit 201 receives data from the memory 202 and executes the program. The main memory may be, for example, a RAM (random access memory). The memory 202 may also be, for example, a ROM (read only memory).

[0060] The storage device 204 is also called an auxiliary storage device, for example, in contrast to the main memory (primary storage device). As described above, the storage device 204 stores an operating program including the program of the present disclosure. The storage device 204 may be, for example, a combination of a recording medium and a drive for reading and writing from and to the recording medium. The recording medium is not particularly limited, and may be, for example, an internal or external type, such as a HD (hard disk), a CD-ROM, a CD-R, a CD-RW, an MO, a DVD, a flash memory, a memory card, and the like. The storage device 204 may be, for example, a hard disk drive (HDD) in which a recording medium and a drive are integrated, and a solid state drive (SSD). When the present device 20 includes the storage unit, for example, the storage device 204 functions as the storage unit. The storage unit can store, for example, information of the user of the present device (name, identification number, attribute information (gender, age, organization, address, contact information, etc.), biological information, etc.), ultrasonic flaw inspection data, image data, preprocessing data, a state determination model, and the like.

[0061] In the present device 20, the memory 202 and the storage device 204 can also store various information such as log information, information acquired from an external database (not shown) or an external device, information generated by the present device 20, and information used when the present device 20 executes processing. In this case, the memory 202 and the storage device 204 may store, for example, the above-mentioned information on the user of the present device. Note that at least a part of the information may be stored, for example, in an external server other than the memory 202 and the storage device 204, or may be distributed and stored in multiple terminals using block chain technology or the like.

[0062] The device 20 further includes, for example, an input device 205 and an output device 206. The input device 205 may be, for example, a pointing device such as a touch panel, a track pad, or a mouse; a keyboard; an imaging means such as a camera or a scanner; a card reader such as an IC card reader or a magnetic card reader; and an audio input means such as a microphone. The output device 206 may be, for example, a display device such as an LED display or a liquid crystal display; an audio output device such as a speaker; a printer; or the like. In the first embodiment, the input device 205 and the output device 206 are configured separately, but the input device 205 and the output device 206 may be configured as an integrated device, such as a touch panel display.

[0063] Next, an example of the state determination model manufacturing method of this embodiment will be described with reference to the flowchart of Fig. 12. The trained model manufacturing method of this embodiment is carried out as follows, for example, using the state determination model manufacturing device 20 of Figs. 10 and 11. Note that the state determination model manufacturing method of this embodiment is not limited to the use of the state determination model manufacturing device 20 of Figs. 10 and 11.

[0064] First, ultrasonic flaw detection inspection data of the inspection target pipe is acquired by the inspection result acquisition unit 21 (S21, inspection result acquisition step). The processing of the inspection result acquisition unit 21 can be performed, for example, in the same manner as S1 in the first embodiment, and the description thereof can be cited.

[0065] Next, the preprocessing unit 22 processes the image data to generate preprocessed data (S22, preprocessing step). The processing by the preprocessing unit 22 may be the same as the processing for generating the preprocessed data in S2 of the first embodiment, or may be a different processing. When the preprocessing unit 22 performs the same processing as the preprocessing steps (the division step, the enlargement step, and the emphasis processing step) in S2 of the first embodiment, for example, the preprocessing unit 22 can be implemented in the same manner as the preprocessing steps in the first embodiment, and the description thereof can be used.

[0066] Furthermore, the pre-processing unit 22 of the present device 20 may further assign at least one of a label indicating a defect and a label indicating noise to the region including the color. The label indicating the defect includes at least one of a label indicating pitting corrosion of the pipe and a label indicating thinning of the pipe. The pre-processing unit 22 may assign a label to each region of the pre-processed data based on an input to the input device 205, for example. Specifically, the pre-processing unit 22 may assign five types of labels, such as "red region of pitting corrosion (pcred)", "yellow region of pitting corrosion (pcyellow)", "red region of thinning region (cwred)", "yellow region of thinning region (cwyellow)", and "noise", to the region including the color. Note that the present device 20 is not limited thereto, and may automatically assign a label estimated by a machine learning model learned by machine learning (for example, learning using a neural network).

[0067] Furthermore, the pre-processing unit 22 may generate additional learning data in which the detected colors are changed to other colors. The combination of colors changed by the pre-processing unit 22 is not particularly limited, and any color can be changed to any other color. As a specific example, if yellow and red are colors associated with defects in the inspected pipe in the color contour map, the pre-processing unit 22 can change yellow to red and green to yellow. This makes it possible to generate learning data in which the number of defective portions is increased, thereby improving the accuracy of defect detection.

[0068] In addition, the pre-processing unit 22 may further generate additional learning data by, for example, a Generative Adversarial Network (GAN) method. In this case, the image data may be expanded. The pipe cross-sectional image data may be expanded. This makes it possible to perform highly accurate state determination even when the number of teacher data is relatively small.

[0069] Next, the trained model generation unit 23 generates, as a trained model, a condition determination model that outputs a condition determination result of the inspection target pipe when the preprocessed data is inputted, by machine learning using the preprocessed data as teacher data (S23, learning step). Specifically, the trained model generation unit 23 constructs the condition determination model by machine learning using, for example, the preprocessed data as teacher data. The machine learning is not particularly limited, and is, for example, machine learning using CNN, SVM, Bayesian network, regression tree, etc. Furthermore, the trained model generation unit 23 may generate a trained model (derived model) that has been retrained, for example, by using the teacher data and a trained model that has already been generated. Furthermore, the trained model generation unit 23 may generate a trained model obtained by transfer learning using a trained model generated using the teacher data, or may generate the trained model by model compression of a trained model generated using the teacher data.

[0070] Hereinafter, the processing by the pre-processing unit 22 and the trained model generating unit 23 will be described with specific examples, but the present disclosure is not limited to the following examples. First, as the ultrasonic flaw detection inspection data for learning, a large number of color contour maps of the inspected pipe obtained from an ultrasonic inspection device (IRIS) were prepared, and five types of labels were assigned to each data: red area (pcred) of pitting corrosion area, yellow area (pcyellow) of pitting corrosion area, red area (cwred) of thinning area, yellow area (cwyellow) of thinning area, and noise. In addition, in each data, the periphery of the yellow area was highlighted with a yellow frame, and the periphery of the red area was highlighted with a red frame.

[0071] Next, the preprocessed data after labeling was divided into four at the position of two straight lines intersecting at the center of the image data to generate divided image data, and furthermore, each divided image data was enlarged twice in both length and width to generate preprocessed data. Then, the enlarged preprocessed data or the preprocessed data before division was trained with Detectron2 to create a state determination model. The number of batches during training was 6, the learning rate was 0.01, and the number of training times was 50,000. Figure 13 shows the performance of each state determination model. As shown in Figure 13, models 1 and 2 are models that use the preprocessed data after division and enlargement, and models 3 and 4 are models that use the preprocessed data before division. In addition, models 1 and 3 use 8,449 pieces of image data as training data, and models 2 and 4 use 16,369 pieces of image data as training data. The detection degree (red) and detection degree (yellow) were used as indicators to evaluate the models. As shown in Figure 13, the detection rate (red) can be calculated by the ratio of the number of detections of red regions (red) in the color classification results of all training images to the total value of the red regions of pitting corrosion areas (pcred) and the red regions of thinning areas (cwred) in all training images. The detection rate (yellow) can be calculated by the ratio of the number of detections of yellow regions (yellow) in the color classification results of all training images to the total value of the yellow regions of pitting corrosion areas (pcyellow) and the yellow regions of thinning areas (cwyellow) in all training images. As shown in Figure 13, all models 1 to 4 were able to determine the condition of the inspected pipe based on the input color contour map. In addition, it was found that models 1 and 2, which performed a vertical and horizontal doubling process after dividing the input image into four as preprocessing, improved both the detection rate (red) and the detection rate (yellow).

[0072] Next, additional learning data was generated by converting yellow in the image to red and green to yellow by image processing for the pre-processed data, and a model was manufactured by learning the pre-processed data and the additional learning data with Detectron2 at a batch number of 6 and a learning rate of 0.01. The number of learning times was also changed to 50,000, 80,000, and 120,000. Figure 14 shows the performance of the state determination model. In Figure 14, model 1 shows the results of the same model as model 1 in Figure 13, model 5 is a model using the additional learning data and learning times of 80,000, and model 6 is a model using the additional learning data and learning times of 120,000. As shown in Figure 14, all of models 1, 5, and 6 were able to determine the state of the inspection target pipe based on the input color contour map. It was also found that the detection degree (red) was improved by using the additional learning data and learning times of 80,000.

[0073] Next, the difference in the pre-processed data due to the presence or absence of highlighting was confirmed. Figure 15 shows the performance of the condition determination model. In Figure 15, model 1 shows the results of the same model as model 1 in Figure 13, model 7 is a model in which the pre-processed data is not highlighted, and model 8 is a model in which the pre-processed data is not highlighted and uses the additional learning data. As shown in Figure 15, all of models 1, 7, and 8 were able to determine the condition of the inspected pipe based on the input color contour map. In addition, model 1 had the highest detection rate (red) and detection rate (yellow), which shows that highlighting significantly improves the defect detection accuracy.

[0074] [Embodiment 4] In the fourth embodiment, an ultrasonic inspection device according to the present disclosure will be described. The ultrasonic inspection device according to the present disclosure includes, for example, a state determination unit, and the state determination unit is the state determination support device according to the present disclosure, and other configurations are not particularly limited.

[0075] The ultrasonic inspection device of the present disclosure may include, for example, various components that are included in known ultrasonic inspection devices. Examples of the components include a probe device, a liquid pump, an ultrasonic control device, a lead cable, and an ultrasonic cable. FIG. 16 shows a schematic diagram of an example of the configuration of the ultrasonic inspection device of this embodiment. As shown in FIG. 16, the ultrasonic inspection device 30 of this embodiment includes a state determination unit (state determination support device) 10, a probe device 31, a liquid pump 32, an ultrasonic control device 33, a lead cable 34, and an ultrasonic cable (not shown).

[0076] The liquid pump 32 can send liquid (e.g., water) to the rear end of the cylindrical container 311 of the probe device 31 shown in Fig. 17 through, for example, the lead cable 34. Specifically, the liquid pump 32 can send the liquid by, for example, increasing the water pressure.

[0077] The lead cable 34 is a cable for connecting, for example, the liquid pump 32 and the probe device 31, and the ultrasonic control device 33 and the probe device 31, and the ultrasonic cable is inserted into the lead cable 34. The ultrasonic cable connects the ultrasonic control device 33 and the probe 312. The lengths of the lead cable 34 and the ultrasonic cable are not particularly limited and can be set appropriately depending on the length of the pipe to be inspected.

[0078] The ultrasonic control device 33 is, for example, a device that controls the transmission and reception of ultrasonic waves by the probe device 31, and a control signal of the ultrasonic control device 33 is transmitted to the probe 312 via the ultrasonic cable, thereby controlling the transmission and reception of ultrasonic waves by the probe 312. In addition, the ultrasonic control device 33 analyzes ultrasonic flaw inspection data received by the probe 312 and transmitted to the ultrasonic control device 33, and generates the ultrasonic flaw inspection data.

[0079] The ultrasonic control device 33 and the state judgment support device 10 are, for example, devices that can be connected to each other. The connection between the ultrasonic control device 33 and the state judgment support device 10 may be, for example, wired or wireless. The wireless connection may be, for example, communication via a communication line network. The communication line network is, for example, as described above.

[0080] The condition determination support device 10 is the condition determination support device of the present disclosure, and determines the condition of the pipe based on ultrasonic flaw detection data acquired from the ultrasonic control device 33.

[0081] The probe device 31 is not particularly limited, and for example, a known probe device can be used. An example of the configuration of the probe device will be described with reference to Figs. 16 and 17. Fig. 17 is a schematic diagram showing an example of the configuration of the probe device 31. The probe device 31 includes a probe device main body and a probe device tip. As shown in Fig. 17, the probe body includes a cylindrical container 311 and a probe 312. The probe 312 is housed inside the cylindrical container 311. The cylindrical container 311 has a tip on the side in the insertion direction into the tube and a rear end on the opposite side to the tip. The probe device tip includes a reflecting mirror 313 and a rotation mechanism 314.

[0082] The size of the tip of the probe and the cylindrical container 311 is not particularly limited as long as they can be inserted into the pipe 2 to be inspected.

[0083] FIG. 18(a) is a schematic diagram showing an example of a liquid flow. As shown in FIG. 18(a), a flow path through which the liquid 3 can flow is formed between the probe 312 and the inner wall of the cylindrical container 311. In the flow path, the liquid 3 can flow from the rear end to the front end. The liquid 3 is not particularly limited, but is, for example, water. For example, as described above, the liquid 3 is sent from the liquid pump 32 to the rear end of the cylindrical container 311 through the lead cable 34. The inside of the tube 2 to be inspected may be filled with the liquid 3 in advance, or may be filled with the liquid 3 by the liquid pump 32 sending the liquid.

[0084] The probe 312 is capable of transmitting and receiving ultrasonic waves. The frequency of the ultrasonic waves can be appropriately set according to the types of the pipe 2 and liquid 3 to be inspected.

[0085] The reflecting mirror 313 can reflect ultrasonic waves transmitted from the probe 312 and emit the ultrasonic waves in a direction perpendicular to the inner wall of the pipe 2 to be inspected, and can also reflect ultrasonic waves reflected by the inner wall of the pipe 2 to be inspected so that the ultrasonic waves are incident on the probe 312. The reflecting mirror 313 is rotatably attached to the tip of the cylindrical container 311 by a rotation mechanism 314. The angle of the reflecting mirror 313 is not particularly limited and can be set appropriately.

[0086] The rotation mechanism 314 can rotate the reflection mirror 313 using the flow of the liquid 3 in the flow path as a driving force. FIG. 18(b) is a schematic diagram showing an example of the configuration of the reflection mirror 313 and the rotation mechanism 314. As shown in FIG. 18(b), the rotation mechanism 314 includes, for example, a cylindrical rotor 314a and a plate-shaped rotor 314b. The cylindrical rotor 314a has a tip portion on the insertion direction side in the tube 2 to be inspected and a rear end portion on the opposite side to the tip portion. The reflection mirror 313 is disposed at the tip portion of the cylindrical rotor 314a. The plate-shaped rotor 314b is disposed at the rear end portion of the cylindrical rotor 314a. The cylindrical rotor 314a is rotatably attached to the tip portion of the cylindrical container via the plate-shaped rotor 314b. The plate-shaped rotor 314b has a groove (not shown). In the groove, the liquid 3 can flow from the tip of the cylindrical container 311 toward the tip of the cylindrical rotor 314a. A discharge hole for the liquid 3 is formed on the side of the cylindrical rotor 314a. That is, the liquid 3 sent to the rear end of the cylindrical container 311 flows through the flow path and flows through the groove of the plate-shaped rotor 314b. The plate-shaped rotor 314b rotates using the force of the liquid 3 flowing through the groove as a driving force, and the cylindrical rotor 314a rotates in conjunction with the plate-shaped rotor 314b. The direction of rotation may be, for example, clockwise or counterclockwise. The cylindrical rotor 314a and the plate-shaped rotor 314b may be integrally molded products.

[0087] 18(b), the rotation mechanism 314 may further include, for example, a bearing 314c and a fixing jig 314d. The bearing 314c is disposed in contact with, for example, the cylindrical container 311, the cylindrical rotor 314a, and the plate-shaped rotor 314b. The fixing jig 314d is disposed in contact with, for example, the cylindrical container 311, the bearing 314c, and the cylindrical rotor 314a. The bearing 314c and the fixing jig 314d are not particularly limited as long as they can fix and support the cylindrical rotor 314a to the tip portion of the cylindrical container 311.

[0088] [Embodiment 5] Next, the ultrasonic inspection method of the present disclosure will be described. The ultrasonic inspection method of the present disclosure includes, for example, a state determination step, and the state determination step is characterized by being the state determination support method of the present disclosure, and other steps and conditions are not limited in any way.

[0089] The ultrasonic inspection method of the present disclosure can be implemented, for example, by using the state determination support device of the present disclosure or the ultrasonic inspection device of the present disclosure, but the ultrasonic inspection method of the present disclosure is not limited to a method using the state determination support device of the present disclosure or the ultrasonic inspection device of the present disclosure. An example of the ultrasonic inspection method of the present embodiment will be described below with reference to the drawings, but the present disclosure is not limited to the following example.

[0090] First, as shown in FIG. 19, an ultrasonic transmitter / receiver is inserted into the pipe 2 to be inspected. The ultrasonic transmitter / receiver may be any device capable of transmitting and receiving ultrasonic waves, and is not particularly limited. For example, the ultrasonic transmitter / receiver may be a probe device in an ultrasonic inspection device, that is, the probe device 31 of the fourth embodiment. Hereinafter, the ultrasonic transmitter / receiver is referred to as the probe device 31. For example, the probe device 31 is preferably inserted into the center of the pipe 2. The pipe 2 to be inspected is filled with a fluid 3. Next, the probe device 31 transmits ultrasonic waves 4 in a vertical direction to the inner wall of the pipe 2 while rotating in the circumferential direction of the pipe 2. Then, the probe device 31 receives the reflection of the transmitted ultrasonic waves 4. Then, the probe device 31 transmits the obtained ultrasonic flaw inspection data to the ultrasonic control device 33 via the lead cable 34. The ultrasonic control device 33 analyzes the obtained ultrasonic flaw inspection data. Then, the ultrasonic control device 33 transmits the analyzed ultrasonic flaw inspection data to the state judgment support device 10, and the state judgment support device 10 judges the state of the pipe 2 based on the ultrasonic flaw inspection data. The state determination by the state determination support device 10 is, for example, as described above in the first and second embodiments.

[0091] In addition, the ultrasonic inspection method of this embodiment may, for example, be configured to send liquid 3 into pipe 2, and transmit ultrasonic waves 4 while rotating in the circumferential direction of pipe 2, as shown in Figure 19, with pipe 2 filled with liquid 3.

[0092] As described above, the ultrasonic inspection method of the present embodiment judges the state of a pipe by the state judgment support method of the present disclosure based on the obtained ultrasonic flaw detection inspection data. Therefore, according to the ultrasonic inspection method of the present embodiment, for example, the state of the inspection target pipe can be easily judged.

[0093] [Embodiment 6] The status determination support program of this embodiment is a program for causing a computer to execute each step of the status determination support method described above. Specifically, the status determination support program of this embodiment is a program for causing a computer to execute an inspection result acquisition procedure and a status determination procedure.

[0094] The inspection result acquisition step includes acquiring ultrasonic inspection data of the inspection target pipe, The state determination step includes generating preprocessed data by processing the image data; The state determination step determines a state of the inspection target pipe based on the ultrasonic flaw detection data.

[0095] The state determination assistance program of this embodiment can also be said to be a program that causes a computer to function as an inspection result acquisition procedure and a state determination procedure.

[0096] The state determination support program of this embodiment can be implemented by using the description of the state determination support device and the state determination support method of the present disclosure. In each of the steps, for example, the "step" can be replaced with "processing". In addition, the program of this embodiment may be recorded in a computer-readable recording medium. The recording medium is, for example, a non-transitory computer-readable storage medium. The recording medium is not particularly limited, and examples thereof include random access memory (RAM), read-only memory (ROM), hard disk (HD), flash memory (for example, SSD (Solid State Drive), USB flash memory, SD / SDHC card, etc.), optical disk (for example, CD-R / CD-RW, DVD-R / DVD-RW, BD-R / BD-RE, etc.), magneto-optical disk (MO), floppy (registered trademark) disk (FD), etc. In addition, the program of this embodiment (for example, also referred to as a programming product or program product) may be in a form distributed from an external computer, for example. The "distribution" may be, for example, distribution via a communication network or distribution via a device connected by a wire. The program of the present embodiment may be installed and executed in the device to which it is distributed, or may be executed without being installed.

[0097] [Embodiment 7] The state determination model manufacturing program of this embodiment is a program for causing a computer to execute each step of the above-mentioned state determination model manufacturing method. Specifically, the state determination model manufacturing program of this embodiment is a program for causing a computer to execute an inspection result acquisition procedure, a preprocessing procedure, and a trained model generation procedure.

[0098] The inspection result acquisition step includes acquiring ultrasonic inspection data of the inspection target pipe, the ultrasonic inspection data includes image data; The pre-processing step generates pre-processed data by processing the image data; The trained model generation procedure generates, as a trained model, a condition determination model that outputs a condition determination result for the inspection target pipe when the preprocessed data is input, through machine learning using the preprocessed data as training data.

[0099] In addition, the state determination model manufacturing program of this embodiment can also be said to be a program that causes a computer to function as an inspection result acquisition procedure, a preprocessing procedure, and a trained model generation procedure.

[0100] The state judgment model manufacturing program of this embodiment can be used in the description of the state judgment model manufacturing device and the state judgment model manufacturing method of the present disclosure. For example, the "procedure" in each of the procedures can be read as "processing". The program of this embodiment may be recorded in a computer-readable recording medium. The recording medium is, for example, a non-transitory computer-readable storage medium. The recording medium is not particularly limited, and examples thereof include random access memory (RAM), read-only memory (ROM), hard disk (HD), flash memory (for example, SSD (Solid State Drive), USB flash memory, SD / SDHC card, etc.), optical disk (for example, CD-R / CD-RW, DVD-R / DVD-RW, BD-R / BD-RE, etc.), magneto-optical disk (MO), floppy (registered trademark) disk (FD), etc. The program of this embodiment (for example, also referred to as a programming product or program product) may be in a form distributed from an external computer, for example. The "distribution" may be, for example, distribution via a communication network or distribution via a device connected by a wire. The program of the present embodiment may be installed and executed in the device to which it is distributed, or may be executed without being installed.

[0101] [Embodiment 8] The ultrasonic inspection program of the present embodiment includes a state determination procedure, and the state determination procedure is executed by the state determination support program of the present disclosure.

[0102] The ultrasonic inspection program of this embodiment can be implemented by using the description of the ultrasonic inspection device and ultrasonic inspection method of the present disclosure. In each of the procedures, for example, "procedure" can be replaced with "processing". In addition, the program of this embodiment may be recorded, for example, in a computer-readable recording medium. The recording medium is, for example, a non-transitory computer-readable storage medium. The recording medium is not particularly limited, and examples thereof include random access memory (RAM), read-only memory (ROM), hard disk (HD), flash memory (for example, SSD (Solid State Drive), USB flash memory, SD / SDHC card, etc.), optical disk (for example, CD-R / CD-RW, DVD-R / DVD-RW, BD-R / BD-RE, etc.), magneto-optical disk (MO), floppy (registered trademark) disk (FD), etc. In addition, the program of this embodiment (for example, also referred to as a programming product or program product) may be, for example, distributed from an external computer. The "distribution" may be, for example, distribution via a communication network or distribution via a device connected by a wire. The program of the present embodiment may be installed and executed in the device to which it is distributed, or may be executed without being installed.

[0103] Although the present disclosure has been described above with reference to the embodiments, the present disclosure is not limited to the above-mentioned embodiments. Various modifications that can be understood by a person skilled in the art can be made to the configuration and details of the present disclosure within the scope of the present disclosure.

[0104] <Additional Notes> Some or all of the above embodiments may be described as follows, but are not limited to the following supplementary notes. (Appendix 1) The test result acquisition unit and the state determination unit are included. The inspection result acquisition unit acquires ultrasonic flaw detection inspection data of the inspection target pipe, The condition determination unit determines the condition of the inspection target pipe based on the ultrasonic flaw detection data. (Appendix 2) The ultrasonic inspection data includes image data, The state determination unit is Generate divided image data by dividing the image data into four parts along two lines that intersect at the center of the image data; 2. A condition determination support device as described in claim 1, which generates pre-processed data by enlarging each of the divided image data by twice its size vertically and horizontally, and determines the condition of the inspection target pipe based on the pre-processed data. (Appendix 3) the image data is a contour image including at least one color; 3. A condition determination support device as described in claim 2, wherein the condition determination unit detects the at least one color in the contour image, performs an enhancement process on an area including the detected color, and determines the condition of the inspection target pipe based on the data after the enhancement process. (Appendix 4) The state determination support device according to claim 3, wherein the state determination unit, as the emphasis processing, draws an outer periphery of an area including the color with a line of the same color as the detected color. (Appendix 5) An image generating unit and an output unit are included, The image generating unit generates a determination image by superimposing a determination result of the state determining unit on the ultrasonic flaw detection inspection data, 5. The condition judgment support device according to claim 1, wherein the output unit outputs the judgment image. (Appendix 6) The image generating unit generates a pipe cross-sectional profile estimated image based on the ultrasonic flaw detection data, The condition determination support device according to claim 5, wherein the output unit outputs the pipe cross-sectional profile estimated image. (Appendix 7) 7. A condition determination support device according to any one of claims 1 to 6, wherein the condition determination unit inputs the ultrasonic flaw detection inspection data into a condition determination model to determine the condition of the inspection target pipe. (Appendix 8) The apparatus includes an inspection result acquisition unit, a preprocessing unit, and a trained model generation unit, The inspection result acquisition unit acquires ultrasonic flaw detection inspection data of the inspection target pipe, the ultrasonic inspection data includes image data; The preprocessing unit processes the image data to generate preprocessed data, The trained model generation unit generates, as a trained model, a condition determination model that outputs a condition determination result of the inspection target pipe when the preprocessed data is input, through machine learning using the preprocessed data as training data. The pipe condition determination model manufacturing device. (Appendix 9) the image data is a contour image including at least one color; 9. The state determination model manufacturing device according to claim 8, wherein the preprocessing unit detects the at least one color in the contour image and performs emphasis processing on an area including the detected color. (Appendix 10) 10. The state judgment model manufacturing device according to claim 9, wherein the pre-processing unit draws an outer periphery of an area including the color with a line of the same color as the detected color. (Appendix 11) 11. The condition determination model manufacturing device according to claim 9, wherein the pre-processing unit assigns at least one of a label indicating a defect and a label indicating noise to the region including the color. (Appendix 12) 12. The condition judgment model manufacturing apparatus according to claim 11, wherein the labels indicating defects include at least one of a label indicating pitting corrosion of a tube and a label indicating thinning of a tube. (Appendix 13) The preprocessing unit generates additional learning data by changing the detected color to another color, 13. A state determination model manufacturing apparatus according to any one of appendices 9 to 12, wherein the trained model generation unit uses the preprocessed data and the additional learning data as teacher data. (Appendix 14) A state determination unit is included, 8. An ultrasonic inspection device, wherein the state determination unit is a state determination support device according to any one of appendices 1 to 7. (Appendix 15) a probe device, a liquid pump, an ultrasonic control device, a lead cable, and an ultrasonic cable; The probe device includes a probe device body and a probe device tip, The probe body includes a cylindrical container and a probe, the cylindrical container is sized so as to be insertable into a pipe to be inspected; The cylindrical container has a front end portion on a side in a direction in which the cylindrical container is inserted into the pipe, and a rear end portion opposite the front end portion, The probe is housed inside the cylindrical container, a flow path through which a liquid can flow is formed between the probe and an inner wall of the cylindrical container; In the flow path, the liquid can flow from the rear end portion to the front end portion, The probe is capable of transmitting and receiving ultrasonic waves, the probe tip is sized to be insertable into the canal; The tip of the probe device includes a reflecting mirror and a rotating mechanism. the reflection mirror is capable of reflecting ultrasonic waves transmitted from the probe and emitting the ultrasonic waves in a direction perpendicular to the inner wall of the pipe, and is also capable of reflecting ultrasonic waves reflected by the inner wall of the pipe so that the ultrasonic waves are incident on the probe; the reflecting mirror is rotatably attached to the tip of the cylindrical container by the rotation mechanism, the rotation mechanism is capable of rotating the reflection mirror using a flow of the liquid in the flow channel as a driving force; the liquid pump is capable of pumping the liquid to the rear end of the cylindrical container of the probe device through the lead cable; The ultrasonic cable is inserted into the lead cable, The ultrasonic control device and the probe are connected by the ultrasonic cable, The ultrasonic control device controls transmission and reception of ultrasonic waves by the probe, and ultrasonic flaw detection data is generated based on the ultrasonic signal data. The ultrasonic control device and the state determination unit are connected to each other, 15. The ultrasonic inspection device of claim 14, wherein the state determination unit determines the state of the pipe based on ultrasonic flaw detection data acquired from the ultrasonic control device. (Appendix 16) The method includes an inspection result acquisition step and a state determination step, The inspection result acquisition step acquires ultrasonic inspection data of the inspection target pipe, The condition determination step determines the condition of the inspection target pipe based on the ultrasonic flaw detection data. (Appendix 17) The ultrasonic inspection data includes image data, The state determination step includes: Generate divided image data by dividing the image data into four parts along two lines that intersect at the center of the image data; A condition determination support method as described in Appendix 16, further comprising generating pre-processed data by doubling the vertical and horizontal dimensions of each of the divided image data, and determining the condition of the inspection target pipe based on the pre-processed data. (Appendix 18) the image data is a contour image including at least one color; 18. A condition determination support method as described in claim 17, wherein the condition determination step detects the at least one color in the contour image, performs an enhancement process on an area including the detected color, and determines the condition of the inspection target pipe based on the data after the enhancement process. (Appendix 19) The state determination support method according to claim 18, wherein the state determination step, as the emphasis processing, draws an outer periphery of an area including the color with a line of the same color as the detected color. (Appendix 20) An image generating step and an output step are included, The image generating step generates a determination image in which a determination result of the state determining step is superimposed on the ultrasonic flaw detection inspection data, 20. The condition judgment support method according to any one of appendices 16 to 19, wherein the output step outputs the judgment image. (Appendix 21) The image generating step generates a pipe cross-sectional profile estimated image based on the ultrasonic flaw detection data, The condition determination support method according to claim 20, wherein the output step outputs the pipe cross-sectional profile estimated image. (Appendix 22) 22. A condition judgment support method according to any one of appendices 16 to 21, wherein the condition judgment step inputs the ultrasonic flaw detection inspection data into a condition judgment model to judge the condition of the inspection target pipe. (Appendix 23) The method includes an inspection result acquisition step, a state determination step, and a trained model generation step, The inspection result acquisition step acquires ultrasonic inspection data of the inspection target pipe, the ultrasonic inspection data includes image data; The state determination step includes generating preprocessed data by processing the image data, The trained model generation process of this method for manufacturing a pipe condition determination model generates, as a trained model, a condition determination model that outputs a condition determination result of the inspection target pipe when the preprocessed data is input, by machine learning using the preprocessed data as training data. (Appendix 24) the image data is a contour image including at least one color; 24. The state determination model manufacturing method according to claim 23, wherein the pre-processing step detects the at least one color in the contour image and performs an emphasizing process on an area including the detected color. (Appendix 25) 25. The state judgment model manufacturing method according to claim 24, wherein the preprocessing step draws an outer periphery of an area including the color with a line of the same color as the detected color. (Appendix 26) 26. The method for producing a condition judgment model according to claim 24 or 25, wherein the preprocessing step assigns at least one of a label indicating a defect and a label indicating noise to the region including the color. (Appendix 27) 27. The condition judgment model manufacturing method of claim 26, wherein the labels indicating defects include at least one of a label indicating pitting corrosion of a tube and a label indicating thinning of a tube. (Appendix 28) The pre-processing step generates additional learning data in which the detected color is changed to another color, A state determination model manufacturing method described in any one of Appendices 24 to 27, wherein the trained model generation process uses the preprocessed data and the additional learning data as teacher data. (Appendix 29) A state determination step is included, 23. An ultrasonic inspection method, wherein the condition determination step is a condition determination support method according to any one of Appendices 16 to 22. (Appendix 30) Includes a procedure for obtaining test results and a procedure for determining status, The inspection result acquisition step includes acquiring ultrasonic inspection data of the inspection target pipe, The state determination step determines a state of the inspection target pipe based on the ultrasonic flaw detection data. A status determination support program for causing a computer to execute each of the above procedures. (Appendix 31) the ultrasonic inspection data includes image data; The state determination procedure includes: Generate divided image data by dividing the image data into four parts along two lines that intersect at the center of the image data; A condition determination support program according to claim 30, which generates pre-processed data by doubling the vertical and horizontal dimensions of each of the divided image data, and determines the condition of the inspection target pipe based on the pre-processed data. (Appendix 32) the image data is a contour image including at least one color; 32. A condition determination support program as described in claim 31, wherein the condition determination step detects the at least one color in the contour image, enhances an area including the detected color, and determines the condition of the inspection target pipe based on the data after the enhancement process. (Appendix 33) 33. The state determination support program according to claim 32, wherein the state determination step, as the emphasis processing, draws an outer periphery of an area including the color with a line of the same color as the detected color. (Appendix 34) An image generating step and an output step are included, the image generating step generates a determination image in which a determination result of the state determining step is superimposed on the ultrasonic flaw detection inspection data; 34. The state judgment support program according to any one of appendices 31 to 33, wherein the output step outputs the judgment image. (Appendix 35) The image generating step generates a pipe cross-sectional profile estimated image based on the ultrasonic flaw detection data, The condition determination support program according to claim 34, wherein the output step outputs the pipe cross-sectional profile estimated image. (Appendix 36) 36. A condition determination support program according to any one of appendices 31 to 35, wherein the condition determination step inputs the ultrasonic flaw detection inspection data into a condition determination model to determine the condition of the inspection target pipe. (Appendix 37) The method includes a procedure for obtaining inspection results, a procedure for determining a state, and a procedure for generating a trained model. The inspection result acquisition step includes acquiring ultrasonic inspection data of the inspection target pipe, the ultrasonic inspection data includes image data; The state determination step includes generating preprocessed data by processing the image data; The trained model generation step generates, as a trained model, a condition determination model that outputs a condition determination result of the inspection target pipe when the preprocessed data is input, by machine learning using the preprocessed data as training data. A state determination model manufacturing program for causing a computer to execute each of the above procedures. (Appendix 38) the image data is a contour image including at least one color; 38. The state determination model generating program according to claim 37, wherein the preprocessing step detects the at least one color in the contour image and highlights an area including the detected color. (Appendix 39) 39. The state determination model manufacturing program according to claim 38, wherein the preprocessing step draws an outer periphery of an area including the color with a line of the same color as the detected color. (Appendix 40) 40. The condition determination model generating program according to claim 38 or 39, wherein the preprocessing step assigns at least one of a label indicating a defect and a label indicating noise to the region including the color. (Appendix 41) 41. The condition judgment model manufacturing program of claim 40, wherein the labels indicating defects include at least one of a label indicating pitting corrosion of a pipe and a label indicating thinning of a pipe. (Appendix 42) The pre-processing step includes generating additional learning data by changing the detected color to another color; A state judgment model manufacturing program described in any one of Appendices 36 to 41, wherein the trained model generation procedure uses the preprocessed data and the additional learning data as teacher data. (Appendix 43) Including a state determination procedure, An ultrasonic inspection program, wherein the state determination procedure is executed by a state determination support program according to any one of appendices 31 to 36. (Appendix 44) Includes a procedure for obtaining test results and a procedure for determining status, The inspection result acquisition step includes acquiring ultrasonic inspection data of the inspection target pipe, The state determination step determines a state of the inspection target pipe based on the ultrasonic flaw detection data. A computer-readable recording medium having recorded thereon a state determination support program for causing a computer to execute each of the above procedures. (Appendix 45) the ultrasonic inspection data includes image data; The state determination procedure includes: Generate divided image data by dividing the image data into four parts along two lines that intersect at the center of the image data; The recording medium of claim 44, further comprising: generating pre-processed data by doubling the vertical and horizontal dimensions of each of the divided image data; and determining the state of the inspection target pipe based on the pre-processed data. (Appendix 46) the image data is a contour image including at least one color; The recording medium of claim 45, wherein the condition determination step detects the at least one color in the contour image, enhances an area including the detected color, and determines the condition of the inspection target pipe based on the data after the enhancement process. (Appendix 47) 47. The recording medium according to claim 46, wherein the state determination step includes drawing an outer periphery of an area including the color with a line of the same color as the detected color as the highlighting process. (Appendix 48) An image generating step and an output step are included, the image generating step generates a determination image in which a determination result of the state determining step is superimposed on the ultrasonic flaw detection inspection data; 48. The recording medium according to any one of Appendices 44 to 47, wherein the output step outputs the judgment image. (Appendix 49) The image generating step generates a pipe cross-sectional profile estimated image based on the ultrasonic flaw detection data, The recording medium according to claim 48, wherein the output step outputs the pipe cross-sectional profile estimated image. (Appendix 50) 50. The recording medium according to any one of appendices 44 to 49, wherein the condition determination step inputs the ultrasonic inspection data into a condition determination model to determine the condition of the inspection target pipe. (Appendix 51) This includes a procedure for obtaining inspection results, a preprocessing procedure, and a procedure for generating a trained model. The inspection result acquisition step includes acquiring ultrasonic inspection data of the inspection target pipe, the ultrasonic inspection data includes image data; The pre-processing step generates pre-processed data by processing the image data; The trained model generation step generates, as a trained model, a condition determination model that outputs a condition determination result of the inspection target pipe when the preprocessed data is input, by machine learning using the preprocessed data as training data. A computer-readable recording medium having recorded thereon a state determination model manufacturing program for causing a computer to execute each of the above procedures. (Appendix 52) the image data is a contour image including at least one color; 52. The recording medium of claim 51, wherein the preprocessing step detects the at least one color in the contour image and highlights an area including the detected color. (Appendix 53) 53. The recording medium of claim 52, wherein the preprocessing step includes drawing an outer periphery of an area including the color with a line of the same color as the detected color. (Appendix 54) 54. The storage medium of claim 52 or 53, wherein the preprocessing step includes assigning at least one of a label indicating a defect and a label indicating noise to the region including the color. (Appendix 55) 55. The recording medium of claim 54, wherein the label indicating the defects includes at least one of a label indicating pitting corrosion of the pipe and a label indicating thinning of the pipe. (Appendix 56) The pre-processing step includes generating additional learning data by changing the detected color to another color; A recording medium described in any one of Appendix 51 to 55, wherein the trained model generation procedure uses the preprocessing data and the additional training data as teacher data. (Appendix 57) Including a state determination procedure, A computer-readable recording medium having an ultrasonic inspection program recorded thereon, the condition determination procedure being executed by a condition determination support program described in any one of Appendices 31 to 36. [Industrial Applicability]

[0105] According to the present disclosure, the condition of an inspected pipe can be easily determined. Therefore, the present disclosure is useful in various fields such as inspection, maintenance, and inspection of piping. [Explanation of symbols]

[0106] 10, 10A Status Judgment Support Device 11. Test result acquisition section 12 Status determination section 13 Image Generation Unit 14 Output section 101 CPU 102 Memory 103 Bus 104 Storage device 105 Input Device 106 Output Device 107 Communication Devices 20. Condition Judgment Model Manufacturing Equipment 21 Test Results Acquisition Department 22 Pretreatment section 23 Trained model generation unit 201 CPU 202 Memory 203 Bus 204 Storage device 205 Input Device 206 Output Device 207 Communication Devices 30 Ultrasonic inspection equipment 2. Pipes to be inspected 3 Liquid (water) 4. Ultrasound 31 Probing device 32 Pump 33 Ultrasonic control device 34 Lead Cable 311 Cylindrical containers 312 Probe 313 Reflective Mirror 314 Rotation Mechanism 314a Cylindrical rotating body 314b Plate-shaped rotating body 314c bearing 314d Fixture

Claims

1. The test result acquisition unit and the state determination unit are included. The inspection result acquisition unit acquires ultrasonic flaw detection inspection data of the inspection target pipe, The condition determination unit determines the condition of the inspection target pipe based on the ultrasonic flaw detection data.

2. The ultrasonic inspection data includes image data, The state determination unit is generating divided image data by dividing the image data into four parts at the position of two straight lines that intersect at the center of the image data; 2. A condition determination support device according to claim 1, further comprising: generating pre-processed data by enlarging each of said divided image data by twice its size vertically and horizontally; and determining the condition of said inspection target pipe based on said pre-processed data.

3. the image data is a contour image including at least one color; 3. The condition determination support device according to claim 2, wherein the condition determination unit detects the at least one color in the contour image, performs an enhancement process on an area including the detected color, and determines the condition of the inspection target pipe based on the data after the enhancement process.

4. The state determination support device according to claim 3 , wherein the state determination unit performs the emphasizing process by drawing a periphery of an area including the color with a line of the same color as the detected color.

5. An image generating unit and an output unit are included, The image generating unit generates a determination image by superimposing a determination result of the state determining unit on the ultrasonic flaw detection inspection data, The condition determination support device according to claim 1 , wherein the output unit outputs the determination image.

6. The image generating unit generates a pipe cross-sectional profile estimated image based on the ultrasonic flaw detection data, The condition determination support device according to claim 5 , wherein the output unit outputs the pipe cross-sectional profile estimated image.

7. The condition determination support device according to claim 1 , wherein the condition determination unit inputs the ultrasonic inspection data into a condition determination model to determine the condition of the inspection target pipe.

8. The apparatus includes an inspection result acquisition unit, a preprocessing unit, and a trained model generation unit, The inspection result acquisition unit acquires ultrasonic flaw detection inspection data of the inspection target pipe, the ultrasonic inspection data includes image data; The preprocessing unit processes the image data to generate preprocessed data, The trained model generation unit generates, as a trained model, a condition determination model that outputs a condition determination result of the inspection target pipe when the preprocessed data is input, through machine learning using the preprocessed data as training data. The pipe condition determination model manufacturing device.

9. the image data is a contour image including at least one color; 9. The apparatus for producing a condition judgment model according to claim 8, wherein the preprocessing unit detects the at least one color in the contour image and performs an emphasizing process on an area including the detected color.

10. 10. The device for producing a condition judgment model according to claim 9, wherein the preprocessing unit draws an outer periphery of an area including the color with a line of the same color as the detected color.

11. 11. The condition judgment model manufacturing device according to claim 9, wherein the preprocessing unit assigns at least one of a label indicating a defect and a label indicating noise to the region including the color.

12. 12. The condition judgment model manufacturing apparatus according to claim 11, wherein the labels indicating defects include at least one of a label indicating pitting corrosion of a pipe and a label indicating thinning of a pipe.

13. The preprocessing unit generates additional learning data by changing the detected color to another color, The state determination model manufacturing device according to claim 9 or 10, wherein the trained model generation unit uses the preprocessed data and the additional learning data as teacher data.

14. A state determination unit is included, An ultrasonic inspection device, wherein the state determination unit is the state determination support device according to claim 1 .

15. a probe device, a liquid pump, an ultrasonic control device, a lead cable, and an ultrasonic cable; The probe device includes a probe device body and a probe device tip, The probe body includes a cylindrical container and a probe, the cylindrical container is sized so as to be insertable into a pipe to be inspected; The cylindrical container has a front end portion on a side in a direction in which the cylindrical container is inserted into the pipe, and a rear end portion opposite the front end portion, The probe is housed inside the cylindrical container, a flow path through which a liquid can flow is formed between the probe and an inner wall of the cylindrical container; In the flow path, the liquid can flow from the rear end portion to the front end portion, The probe is capable of transmitting and receiving ultrasonic waves, the probe tip is sized to be insertable into the canal; The tip of the probe device includes a reflecting mirror and a rotating mechanism. the reflection mirror is capable of reflecting ultrasonic waves transmitted from the probe and emitting the ultrasonic waves in a direction perpendicular to the inner wall of the pipe, and is also capable of reflecting ultrasonic waves reflected by the inner wall of the pipe so that the ultrasonic waves are incident on the probe; the reflecting mirror is rotatably attached to the tip of the cylindrical container by the rotation mechanism, the rotation mechanism is capable of rotating the reflection mirror using a flow of the liquid in the flow channel as a driving force; the liquid pump is capable of pumping the liquid to the rear end of the cylindrical container of the probe device through the lead cable; The ultrasonic cable is inserted into the lead cable, The ultrasonic control device and the probe are connected by the ultrasonic cable, The ultrasonic control device controls transmission and reception of ultrasonic waves by the probe, and ultrasonic flaw detection data is generated based on the ultrasonic signal data. The ultrasonic control device and the state determination unit are connected to each other, The ultrasonic inspection device according to claim 14 , wherein the state determination unit determines the state of the pipe based on ultrasonic flaw detection data acquired from the ultrasonic control device.

16. The method includes an inspection result acquisition step and a state determination step, The inspection result acquisition step acquires ultrasonic inspection data of the inspection target pipe, The condition determination step determines the condition of the inspection target pipe based on the ultrasonic flaw detection data.

17. The method includes an inspection result acquisition step, a state determination step, and a trained model generation step, The inspection result acquisition step acquires ultrasonic inspection data of the inspection target pipe, the ultrasonic inspection data includes image data; The state determination step includes generating preprocessed data by processing the image data, The trained model generation process of this method for manufacturing a pipe condition determination model generates, as a trained model, a condition determination model that outputs a condition determination result of the inspection target pipe when the preprocessed data is input, by machine learning using the preprocessed data as training data.

18. A state determination step is included, 17. An ultrasonic inspection method, wherein the condition determination step is a condition determination support method according to claim 16.

19. Includes a procedure for obtaining test results and a procedure for determining status, The inspection result acquisition step includes acquiring ultrasonic inspection data of the inspection target pipe, The state determination step determines a state of the inspection target pipe based on the ultrasonic flaw detection data. A status determination support program for causing a computer to execute each of the above procedures.

20. The method includes a procedure for obtaining inspection results, a procedure for determining a state, and a procedure for generating a trained model. The inspection result acquisition step includes acquiring ultrasonic inspection data of the inspection target pipe, the ultrasonic inspection data includes image data; The state determination step includes generating preprocessed data by processing the image data; The trained model generation step generates, as a trained model, a condition determination model that outputs a condition determination result of the inspection target pipe when the preprocessed data is input, by machine learning using the preprocessed data as training data. A state determination model manufacturing program for causing a computer to execute each of the above procedures.

21. Including a state determination procedure, 20. An ultrasonic inspection program, wherein the state determination step is executed by a state determination support program according to claim 19.

22. Includes a procedure for obtaining test results and a procedure for determining status, The inspection result acquisition step includes acquiring ultrasonic inspection data of the inspection target pipe, The state determination step determines a state of the inspection target pipe based on the ultrasonic flaw detection data. A computer-readable recording medium having recorded thereon a state determination support program for causing a computer to execute each of the above procedures.

23. This includes a procedure for obtaining inspection results, a preprocessing procedure, and a procedure for generating a trained model. The inspection result acquisition step includes acquiring ultrasonic inspection data of the inspection target pipe, the ultrasonic inspection data includes image data; The pre-processing step generates pre-processed data by processing the image data; The trained model generation step generates, as a trained model, a condition determination model that outputs a condition determination result of the inspection target pipe when the preprocessed data is input, by machine learning using the preprocessed data as training data. A computer-readable recording medium having recorded thereon a state determination model manufacturing program for causing a computer to execute each of the above procedures.

24. Including a state determination procedure, 20. A computer-readable recording medium having an ultrasonic inspection program recorded thereon, the state determination procedure being executed by a state determination support program according to claim 19.

Images