Defect inspection device, defect inspection method, and defect inspection system
The defect inspection apparatus simplifies the detection of defects in structures by analyzing deformation distributions and estimating defect shapes using image analysis and machine learning, eliminating the need for physical loading during inspection.
Patent Information
- Application Number
- PCT/JP2025/031842
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2025-05-15
- Filing Date
- 2025-09-09
- Publication Date
- 2026-11-19
AI Technical Summary
Existing defect inspection methods for structures like bridges require applying an inspection load, which is cumbersome and time-consuming, and there is a need for a simpler method that can detect defects without this requirement.
A defect inspection apparatus that includes an imaging unit, deformation measurement unit, defect estimation database, and defect estimation unit to analyze deformation distributions and estimate defect shapes based on virtual defect shapes and load conditions, allowing defect detection without applying a load during inspection.
Enables defect inspection with a simple configuration, reducing inspection time and complexity by measuring deformation distributions and estimating defect shapes using image analysis and machine learning, without the need for physical loading.
Smart Images

Figure JP2025031842_19112026_PF_FP_ABST
Abstract
Description
Defect inspection apparatus, defect inspection method, and defect inspection system
[0001] The present disclosure relates to a defect inspection apparatus, a defect inspection method, and a defect inspection system that perform inspection of an object using an image obtained by capturing an image of the object.
[0002] Structures such as bridges are regularly inspected to check for the presence or absence of defects caused by loads and the like. Further, as one of the methods for determining the presence or absence of defects, a method is known in which the displacement distribution on the surface of the object is obtained based on a plurality of images captured of the object (inspection target), and defects are detected from the displacement distribution (for example, Patent Document 1).
[0003] International Publication No. 2021 / 152811
[0004] In inspections targeting bridges and the like, it is common practice to capture images of the inspection target in an unloaded state and the inspection target in a loaded state, and detect defects using the obtained images. That is, a means for applying an inspection load to the inspection target is required. For this reason, realization of a technique that enables inspection by a simple method without requiring a means for applying a load is desired.
[0005] The present disclosure has been made in view of the above, and an object thereof is to obtain a defect inspection apparatus capable of performing inspection with a simple configuration.
[0006] In order to solve the above-described problems and achieve the object, a defect inspection apparatus according to the present disclosure includes: an imaging unit that images a range including a defect of an inspection target in which a defect has occurred; a deformation measurement unit that obtains a deformation distribution on the surface of the inspection target based on a past image and a current image of the inspection target captured by the imaging unit; a defect estimation database unit in which a plurality of virtual defect shapes, which are shapes of defects predicted to occur in the inspection target, are registered, and deformation distributions on the surface of the inspection target in a case where each of the defects of the plurality of virtual defect shapes has occurred are registered; and a defect estimation unit that estimates the shape of a defect occurring in the inspection target based on the deformation distribution obtained by the deformation measurement unit, and the virtual defect shapes and the deformation distributions registered in the defect estimation database unit.
[0007] This disclosure offers the advantage of providing a defect inspection device that can perform inspections with a simple configuration.
[0008] A diagram showing an example of the configuration of a defect inspection device according to Embodiment 1. A diagram showing an example of the hardware configuration of a defect inspection device according to Embodiment 1. A flowchart showing an example of the operation of a defect inspection device according to Embodiment 1. A first schematic diagram showing an example of how the imaging unit of the defect inspection device according to Embodiment 1 photographs the object to be inspected. A second schematic diagram showing an example of how the imaging unit of the defect inspection device according to Embodiment 1 photographs the object to be inspected. A diagram showing an example of an image of the object to be inspected photographed by the imaging unit of the defect inspection device according to Embodiment 1. A schematic diagram showing a first example of a load assumed when performing structural analysis of the defect shape. A schematic diagram showing a second example of a load assumed when performing structural analysis of the defect shape. A schematic diagram showing an example of measurement results by the deformation measurement unit of the defect inspection device according to Embodiment 1. A diagram showing an example of the configuration of a defect inspection device according to Embodiment 2. A diagram showing an example of measurement results by the deformation measurement unit of the defect inspection device according to Embodiment 2. A diagram showing an example of the configuration of a defect inspection device according to Embodiment 3. A diagram showing an example of the hardware configuration of a defect inspection device according to Embodiment 3. A flowchart showing an example of the operation of a defect inspection device according to Embodiment 3. A diagram showing an example of the configuration of a defect inspection device according to Embodiment 4. A diagram showing an example of the configuration of a defect inspection device according to Embodiment 5. A diagram showing an example of the hardware configuration of a defect inspection device according to Embodiment 5. A diagram showing an example of the configuration of a defect inspection device according to Embodiment 6.
[0009] The defect inspection apparatus, defect inspection method, and defect inspection system according to embodiments of this disclosure will be described in detail below with reference to the drawings.
[0010] Embodiment 1. Figure 1 shows an example of the configuration of the defect inspection device 1 according to Embodiment 1. The defect inspection device 1 shown in Figure 1 has a function to acquire images of the surface of the object to be inspected, calculates the amount of deformation of the surface of the object to be inspected from multiple images acquired at intervals, and estimates the size of defects such as cracks and their shape in the depth direction based on the calculated amount of deformation. The object to be inspected by the defect inspection device 1 is, for example, a bridge.
[0011] The defect inspection device 1 includes a surface deformation acquisition unit 10 that determines the deformation distribution of the surface by comparing a current image of the surface of the object to be inspected with a past image, a defect analysis unit 20 that estimates the size and depth shape of the defect in the object to be inspected based on the deformation distribution of the surface of the object to be inspected, and an estimation result output unit 30 that outputs the estimation results from the defect analysis unit 20.
[0012] The surface deformation acquisition unit 10 includes an imaging unit 11, an image storage unit 12, and a deformation measurement unit 13. The defect analysis unit 20 includes a defect structure analysis unit 21, a data input unit 22, a defect estimation database unit 23, and a defect estimation unit 24.
[0013] The imaging unit 11 of the surface deformation acquisition unit 10 photographs the surface of the object to be inspected. The imaging unit 11 photographs the area in which defects have already occurred, that is, the area in which defects have already occurred, so that the area in which defects have already occurred is included in the image.
[0014] The image storage unit 12 of the surface deformation acquisition unit 10 stores images of the surface of the object to be inspected, which are captured by the imaging unit 11.
[0015] The deformation measurement unit 13 of the surface deformation acquisition unit 10 uses multiple images held in the image storage unit 12, specifically past images of the object to be inspected and the most recent image of the object to be inspected, and measures the deformation distribution of the surface of the object to be inspected by comparing these images. The deformation measurement unit 13 measures the deformation distribution of the surface of the object to be inspected, for example, by using the digital image correlation method.
[0016] The defect structural analysis unit 21 of the defect analysis unit 20 performs a three-dimensional structural analysis using the shape of the object to be inspected, load conditions, and virtual defect shapes to determine the deformation distribution on the surface of the object to be inspected. The load conditions are the load applied to the object to be inspected when a defect is propagating and the load applied to the object to be inspected during inspection. The virtual defect shape represents the shape of a defect that is predicted to occur when the load indicated by the load conditions is applied to the object to be inspected. Multiple virtual defect shapes are prepared for each location with a high probability of defect occurrence, such as weld toes, unwanted welds, and locations where stress is concentrated. Multiple virtual defect shapes are prepared for each location. That is, multiple types of virtual defect shapes with different shapes are prepared for each location. The defect structural analysis unit 21 repeatedly performs a three-dimensional structural analysis while changing the virtual defect shapes used, without changing the shape of the object to be inspected or the load conditions used, and determines multiple deformation distributions (deformation distributions on the surface of the object to be inspected) corresponding to each of the multiple virtual defect shapes. At this time, the defect structural analysis unit 21 analyzes not only the size of defects visible on the surface of the object to be inspected, but also the size and shape of defects in the invisible depth direction. The deformation distribution corresponding to the virtual defect shape is the deformation distribution of the surface of the object being inspected when a defect of the virtual defect shape occurs.
[0017] The data input unit 22 of the defect analysis unit 20 acquires the shape of the object to be inspected, the load conditions, and multiple virtual defect shapes, and inputs them into the defect structural analysis unit 21.
[0018] The defect estimation database unit 23 of the defect analysis unit 20 stores the deformation distribution of the surface of the object to be inspected, which is the analysis result from the defect structural analysis unit 21, in association with the virtual defect shape and load conditions used by the defect structural analysis unit 21 when obtaining the analysis result.
[0019] The defect estimation unit 24 of the defect analysis unit 20 estimates the shape of a defect based on data stored in the defect estimation database unit 23 and measurement results from the deformation measurement unit 13, using machine learning or statistical methods. Specifically, the defect estimation unit 24 estimates the shape of a defect based on the deformation distribution of the surface of the object under inspection, the virtual defect shape, and the load conditions stored in the defect estimation database unit 23, and the measurement results of the deformation distribution of the surface of the object under inspection by the deformation measurement unit 13. For example, the defect estimation unit 24 uses Euclidean distance to determine the similarity between the deformation distribution of the surface of the object under inspection measured by the deformation measurement unit 13 and the deformation distribution of the surface of the object under inspection associated with each of the multiple virtual defect shapes stored in the defect estimation database unit 23, and estimates the shape of the defect by taking a weighted average of the virtual defect shapes from the determined similarity. A specific example of how the defect estimation unit 24 estimates the shape of a defect is shown below.
[0020] For the sake of simplicity, we will assume there are two virtual defect shapes, which will be referred to as virtual defect shape A and virtual defect shape B. Furthermore, the deformation distribution associated with virtual defect shape A will be referred to as deformation distribution A, and the deformation distribution associated with virtual defect shape B will be referred to as deformation distribution B. For example, if the similarity between the deformation distribution of the surface of the inspection target measured by the deformation measurement unit 13 and deformation distribution A is 0.1, and the similarity between the deformation distribution of the surface of the inspection target measured by the deformation measurement unit 13 and deformation distribution B is 10, the defect estimation unit 24 will obtain the defect estimation result according to the following equation (1). Note that the closer the similarity is to 0, the more similar the deformation distributions are considered to be. In equation (1), shape A is virtual defect shape A, shape B is virtual defect shape B, and σ is a constant representing the magnitude of the noise.
[0021] (Estimated result of defects) = (Shape A * exp(-0.1 / σ) + Shape B * exp(-10 / σ)) / (exp(-0.1 / σ) + exp(-10 / σ)) ... (1)
[0022] When the defect estimation unit 24 estimates the shape of a defect using machine learning, for example, a trained model is created in advance that has learned the relationship between a virtual defect shape and the deformation distribution of the surface of the object being inspected. The defect estimation unit 24 then estimates the shape of the defect using the trained model and the deformation distribution of the surface of the object being inspected estimated by the deformation measurement unit 13.
[0023] Figure 2 shows an example of the hardware configuration of the defect inspection device 1 according to Embodiment 1. As shown in Figure 2, the defect inspection device 1 consists of a camera 101, a storage device 102, a personal computer (PC) 103 for deformation measurement, a PC 104 for defect estimation, a database 105 for defect estimation, and a PC 106 for defect shape structure analysis.
[0024] The camera 101, storage device 102, and deformation measurement PC 103 constitute the surface deformation acquisition unit 10 described above. Specifically, the camera 101 constitutes the imaging unit 11, the storage device 102 constitutes the image storage unit 12, and the deformation measurement PC 103 constitutes the deformation measurement unit 13. In addition, the defect estimation PC 104, defect estimation database 105, and defect shape structure analysis PC 106 constitute the defect analysis unit 20 described above. Specifically, the defect shape structure analysis PC 106 constitutes the defect structure analysis unit 21, the defect estimation database 105 constitutes the defect estimation database unit 23, and the defect estimation PC 104 constitutes the defect estimation unit 24. Although Figure 2 shows an example in which one processing unit is configured with one piece of hardware (PC), multiple processing units may be configured with one piece of hardware. For example, the defect structure analysis unit 21 and the defect estimation unit 24 of the defect analysis unit 20 may be configured with a single PC.
[0025] Next, we will explain an example of the operation of the defect inspection device 1. Figure 3 is a flowchart showing an example of the operation of the defect inspection device 1 according to Embodiment 1.
[0026] As shown in Figure 3, the defect inspection device 1 performs a surface deformation distribution acquisition step S10 to acquire the deformation distribution of the surface of the object to be inspected, and a defect estimation step S20 to estimate the shape of the defect occurring on the object to be inspected, and outputs the estimated defect shape as the defect estimation result in step S30.
[0027] The surface deformation distribution acquisition step S10 is a process performed by the surface deformation acquisition unit 10. In this surface deformation distribution acquisition step S10, first, the imaging unit 11 photographs the object to be inspected (step S11), and the image storage unit 12 stores an image of the surface of the object to be inspected (step S12). Next, the deformation measurement unit 13 calculates the deformation distribution of the surface of the object to be inspected (step S13). The deformation measurement unit 13 uses past images of the surface of the object to be inspected stored in the image storage unit 12 and the current image of the surface of the object to be inspected to calculate the deformation distribution, for example, by the digital image correlation method.
[0028] Here, the method by which the imaging unit 11 photographs the object to be inspected will be explained using Figures 4 to 6. As an example, we will explain the case where the object to be inspected is a bridge.
[0029] Figure 4 is a first schematic diagram showing an example of how the imaging unit 11 of the defect inspection device 1 according to Embodiment 1 photographs an object to be inspected, and Figure 5 is a second schematic diagram showing an example of how the imaging unit 11 of the defect inspection device 1 according to Embodiment 1 photographs an object to be inspected. Figures 4 and 5 show an example in which the camera constituting the imaging unit 11 photographs a defect that has occurred in a bridge beam. As shown in Figures 4 and 5, the imaging unit 11 (camera) photographs the range that includes the location where the defect has occurred. The setting of the shooting range may be done by the user of the defect inspection device 1, or it may be done automatically by the defect inspection device 1. An example of an image taken by the imaging unit 11 is shown in Figure 6.
[0030] Furthermore, if the surface of the object being inspected that has a defect does not have a pattern, a random pattern may be applied to the surface of the object being inspected before the imaging unit 11 takes an image. By taking an image with a random pattern applied, the measurement accuracy of the digital image correlation method can be improved.
[0031] Next, the loads assumed by the defect inspection device 1 when performing structural analysis of the defect shape will be explained using Figures 7 and 8. Figure 7 is a schematic diagram showing a first example of the loads assumed when performing structural analysis of the defect shape, and Figure 8 is a schematic diagram showing a second example of the loads assumed when performing structural analysis of the defect shape.
[0032] There are two types of loads assumed in structural analysis. The first type of load is the load that causes defects in the object under inspection to progress under normal conditions. In the case of a bridge under inspection, the loads that cause defects to progress under normal conditions are the weight of the bridge itself and the load from vehicles traveling on the bridge, as shown in Figure 7.
[0033] The second load is the load applied to the object being inspected when it is being inspected using image diagnostics. When the object being inspected is a bridge, as shown in Figure 8, the load applied to the object being inspected using image diagnostics includes the weight of the bridge itself and the load from vehicles parked on the bridge for the purpose of deforming it. The defect inspection device 1 according to this embodiment can measure plastic deformation without applying a load to the object being inspected during inspection. However, if images are acquired with a load applied to the object being inspected, as in the conventional method, the load from vehicles parked on the bridge for the purpose of deforming it will also be incorporated into the structural analysis as a load applied during inspection.
[0034] Next, the measurement results by the deformation measurement unit 13, that is, the deformation distribution of the surface of the object to be inspected, will be explained using Figure 9. Figure 9 is a schematic diagram showing an example of the measurement results by the deformation measurement unit 13 of the defect inspection device 1 according to Embodiment 1.
[0035] As a defect propagates, a plastic deformation region of several millimeters or more may occur at the tip of the defect due to tension (see Figure 9). In Figure 9, the area labeled "plastic region" corresponds to the plastic deformation region. The deformation measurement unit 13 measures the plastic deformation region using methods such as digital image correlation. The defect inspection device 1 performs inspection using the plastic deformation region measured by the deformation measurement unit 13. Furthermore, if a load is applied to the object being inspected during inspection (as in the example shown in Figure 8), elastic deformation is also added to the deformation distribution measured by the deformation measurement unit 13.
[0036] As shown in Figure 9, the plastic deformation region (plastic zone) persists after defect propagation, making it possible to measure the plastic deformation region without applying a load during inspection. Furthermore, since the size of the plastic deformation region is related to the load during defect propagation, the load during defect propagation can be determined from the size of the plastic deformation region. Conventional digital image correlation methods determine the deformation distribution of the surface of the object being inspected from the deformation before and after load application, but the defect inspection device 1 does not require applying a load to the object being inspected during inspection, thus shortening inspection time and simplifying the inspection process.
[0037] Returning to the explanation of Figure 3, the defect estimation step S20 is a process performed by the defect analysis unit 20. In this defect estimation step S20, first, the defect structural analysis unit 21 performs a structural analysis using multiple virtual defect shapes (step S21) and registers the analysis results in a database (step S22). Specifically, the defect structural analysis unit 21 performs a three-dimensional structural analysis using the shape of the object to be inspected, load conditions, and multiple virtual defect shapes, and determines the deformation distribution of the surface of the object to be inspected for each of the multiple virtual defect shapes. Then, the defect structural analysis unit 21 associates the determined deformation distribution with the load conditions and virtual defect shapes used in the three-dimensional structural analysis and registers the analysis results in the defect estimation database unit 23. Next, the defect estimation unit 24 estimates the defect (step S23). Based on the deformation distribution calculated in step S13 and the analysis results registered in the defect estimation database unit 23, the defect estimation unit 24 estimates the shape of the defect in the object to be inspected.
[0038] In Figure 3, an example is shown in which the surface deformation acquisition unit 10 executes steps S11 to S13 of the surface deformation distribution acquisition step S10, and then the defect analysis unit 20 executes steps S21 to S23 of the defect estimation step S20. However, the execution order of steps S11 to S13 and steps S21 to S23 may be partially reversed. Specifically, steps S21 and S22 may be executed first, followed by steps S11 to S13 and S23. Alternatively, steps S11 to S13 and steps S21 to S22 may be executed simultaneously and in parallel.
[0039] Furthermore, if the object to be inspected has multiple defects, the defect inspection device 1 takes a photograph of each of the multiple defects and estimates the shape based on the obtained images. In this case, steps S21 and S22 described above only need to be performed once at the beginning. Also, if a single image obtained by photographing the object to be inspected contains multiple defects, the defect inspection device 1 estimates the shape of each of the multiple defects.
[0040] As described above, the defect inspection device 1 according to Embodiment 1 includes an imaging unit 11 that photographs the area containing the defect occurring on the object to be inspected, and a deformation measurement unit 13 that determines the deformation distribution of the surface of the object to be inspected based on the current image and past images of the object to be inspected. The defect inspection device 1 also includes a defect structure analysis unit 21 that determines the deformation distribution of the surface of the object to be inspected from a plurality of virtual defect shapes representing the shape of defects that are predicted to occur on the object to be inspected, and the load conditions when defects of the virtual defect shapes occur, and a defect estimation database unit 23 that stores the deformation distribution determined by the defect structure analysis unit 21 in association with the virtual defect shapes and load conditions used when determining this deformation distribution. The defect inspection device 1 also includes a defect estimation unit 24 that estimates the shape of the defect occurring on the object to be inspected based on the deformation distribution of the surface of the object to be inspected determined by the deformation measurement unit 13 and the deformation distribution and virtual defect shapes stored in the defect estimation database unit 23. According to the defect inspection device 1 according to Embodiment 1, it is possible to estimate the shape of a defect based on an image taken without applying a load to the object to be inspected. In other words, the defect inspection device 1 can perform inspections, including defect shape estimation, with a simple configuration.
[0041] Note that in the present embodiment, the defect inspection apparatus 1 configured to include the surface deformation acquisition unit 10 that captures an image of the surface of an inspection object to obtain the surface deformation distribution and the defect analysis unit 20 that estimates the shape of a defect has been described, but the configuration is not limited thereto. For example, a configuration that inspects an inspection object by combining a first apparatus provided with the surface deformation acquisition unit 10 and a second apparatus provided with the defect analysis unit 20 may be adopted. In this case, a defect inspection system may be configured that includes the first apparatus provided with the surface deformation acquisition unit 10 and the second apparatus provided with the defect analysis unit 20, and the first apparatus and the second apparatus transmit and receive data and the like via a communication network or the like. As another example, a configuration in which the defect estimation database unit 23 included in the defect analysis unit 20 is provided outside, such as in a cloud, may be adopted.
[0042] Second Embodiment. FIG. 10 is a diagram illustrating a configuration example of a defect inspection apparatus 1a according to the second embodiment. In FIG. 10, the same reference numerals are assigned to constituent elements common to those of the defect inspection apparatus 1 according to the first embodiment illustrated in FIG. 1. Descriptions of the constituent elements common to those of the defect inspection apparatus 1 will be omitted. As illustrated in FIG. 10, the defect inspection apparatus 1a has a configuration in which a propagation-time stress estimation unit 40 is added to the defect inspection apparatus 1 according to the first embodiment.
[0043] The propagation-time stress estimation unit 40 estimates the stress that was generated when a defect in an inspection object propagates. As described in the first embodiment, a deformation distribution including a plastic deformation region (plastic zone) due to defect propagation is obtained as a measurement result by the deformation measurement unit 13 (see FIG. 11). Note that FIG. 11 is a diagram illustrating an example of a measurement result obtained by the deformation measurement unit 13 of the defect inspection apparatus 1a according to the second embodiment. The propagation-time stress estimation unit 40 determines the size of the plastic deformation region from the deformation distribution obtained by the measurement by the deformation measurement unit 13, and estimates the stress that was generated when the defect propagates, based on one or both of the radius and the area of the plastic deformation region.
[0044] As described above, according to the defect inspection apparatus 1a according to the second embodiment, it is possible to obtain an estimated value of the stress that was generated when a defect occurring in an inspection object propagates.
[0045] Embodiment 3. FIG. 12 is a diagram showing a configuration example of a defect inspection apparatus 1b according to the third embodiment. In FIG. 12, the same reference numerals are assigned to components common to the defect inspection apparatus 1 according to the first embodiment shown in FIG. 1. Descriptions of the components common to the defect inspection apparatus 1 are omitted. As shown in FIG. 12, the defect inspection apparatus 1b has a configuration in which a load analysis unit 50 and a convergence determination unit 60 are added to the defect inspection apparatus 1 according to the first embodiment.
[0046] The load analysis unit 50 estimates the load applied to the inspection object based on the defect shape obtained by the estimation by the defect estimation unit 24 and the deformation distribution of the surface of the inspection object obtained by the measurement by the deformation measurement unit 13. Specifically, the load analysis unit 50 estimates the load applied to the inspection object at the time of defect propagation and the load applied to the inspection object at the time of inspection.
[0047] The load analysis unit 50 includes a load structural analysis unit 51, a data input unit 52, a load estimation database unit 53, and a load estimation unit 54.
[0048] The load structural analysis unit 51 of the load analysis unit 50 performs three-dimensional structural analysis using the shape of the inspection object, load conditions, and a virtual defect shape, to obtain the deformation distribution of the surface of the inspection object. The load conditions are the load at the time of defect propagation and the load at the time of inspection. A plurality of load conditions are prepared. The load structural analysis unit 51 obtains the deformation distribution of the surface of the inspection object by the same method as the defect structural analysis unit 21. The load structural analysis unit 51 repeatedly executes the three-dimensional structural analysis while changing the used load conditions without changing the used shape of the inspection object and the virtual defect shape, to obtain a plurality of deformation distributions (deformation distributions of the surface of the inspection object) respectively corresponding to the plurality of load conditions. Here, the load structural analysis unit 51 uses the defect shape estimated by the defect estimation unit 24 as the virtual defect shape.
[0049] The data input unit 52 of the load analysis unit 50 acquires the shape of the inspection object, the plurality of load conditions, and the virtual defect shape, and inputs them to the load structural analysis unit 51. At this time, the data input unit 52 acquires the defect shape estimated by the defect estimation unit 24 from the defect estimation unit 24, and inputs the acquired shape to the load structural analysis unit 51 as the virtual defect shape.
[0050] The load estimation database unit 53 of the load analysis unit 50 stores the deformation distribution of the surface of the object to be inspected, which is the analysis result from the load structural analysis unit 51, in association with the virtual defect shape and load conditions used by the load structural analysis unit 51 when obtaining the analysis result.
[0051] The load estimation unit 54 of the load analysis unit 50 estimates the load applied to the object under inspection based on data stored in the load estimation database unit 53 and measurement results from the deformation measurement unit 13, using machine learning or statistical methods. Specifically, the load estimation unit 54 estimates the load that propagates defects in the object under inspection under normal conditions and the load applied to the object under inspection during inspection. Specifically, the load estimation unit 54 estimates the load applied to the object under inspection based on the deformation distribution of the surface of the object under inspection, the virtual defect shape and load conditions stored in the load estimation database unit 53, and the measurement results of the deformation distribution of the surface of the object under inspection by the deformation measurement unit 13. For example, the load estimation unit 54 uses Euclidean distance to determine the similarity between the deformation distribution of the surface of the object under inspection measured by the deformation measurement unit 13 and the deformation distribution of the surface of the object under inspection associated with each of the multiple load conditions stored in the load estimation database unit 53, and estimates the load applied to the object under inspection by taking a weighted average of the load conditions from the determined similarity.
[0052] The convergence determination unit 60 compares the load conditions used in the defect analysis unit 20 with the estimation results from the load estimation unit 54 and determines whether the similarity between them is below a threshold. The similarity between the load conditions used in the defect analysis unit 20 and the estimation results from the load estimation unit 54 is expressed, for example, by the Euclidean distance between the vectors when the load indicated by the load conditions and the load indicated by the estimation results are each represented as vectors. In this case, the convergence determination unit 60 determines that the similarity is below a threshold if, for example, the Euclidean distance between the length of the vector when the load indicated by the load conditions is represented as a vector and the length of the vector when the load indicated by the estimation results is represented as a vector is, for example, 5% or less.
[0053] The convergence determination unit 60 determines that the defect shape estimation process by the defect analysis unit 20 and the load estimation process by the load analysis unit 50 have converged if the similarity is below a threshold, and outputs the load estimation result by the load estimation unit 54 and the defect shape estimation result by the defect estimation unit 24 to the estimation result output unit 30. On the other hand, if the similarity is greater than the threshold, the convergence determination unit 60 outputs the load estimation result by the load estimation unit 54 to the data input unit 22. The data input unit 22 inputs the load estimation result by the load estimation unit 54 as a new load condition to the defect structure analysis unit 21. The defect structure analysis unit 21 performs a three-dimensional structural analysis using the new load condition to determine the deformation distribution of the surface of the object to be inspected. The defect structure analysis unit 21 also updates the data stored in the defect estimation database unit 23 based on the deformation distribution obtained using the new load condition. That is, the defect structure analysis unit 21 registers the obtained deformation distribution in the defect estimation database unit 23 in association with the new load condition and virtual defect shape. The defect estimation unit 24 uses the updated data stored in the defect estimation database unit 23 to re-estimate the shape of the defect.
[0054] Next, using the defect shape estimated again by the defect estimation unit 24, each part of the load analysis unit 50 executes the above-described process again to estimate the load applied to the object being inspected.
[0055] The defect analysis unit 20 and the load analysis unit 50 repeat the above operations until the convergence determination unit 60 determines that the defect shape estimation process and the load estimation process have converged.
[0056] Figure 13 shows an example of the hardware configuration of the defect inspection device 1b according to Embodiment 3. As shown in Figure 13, the defect inspection device 1b consists of a camera 101, a storage device 102, a deformation measurement PC 103, a defect estimation PC 104, a defect estimation database 105, a defect shape structure analysis PC 106, a load estimation PC 107, a load estimation database 108, a load structure analysis PC 109, and a convergence determination PC 110. In Figure 13, the same reference numerals are used for hardware that is common to the hardware that constitutes the defect inspection device 1 according to Embodiment 1 shown in Figure 2.
[0057] The camera 101, storage device 102, and deformation measurement PC 103 constitute the surface deformation acquisition unit 10, while the defect estimation PC 104, defect estimation database 105, and defect shape structure analysis PC 106 constitute the defect analysis unit 20.
[0058] The load estimation PC 107, the load estimation database 108, and the load structure analysis PC 109 constitute the load analysis unit 50. Specifically, the load structure analysis PC 109 constitutes the load structure analysis unit 51, the load estimation database 108 constitutes the load estimation database unit 53, and the load estimation PC 107 constitutes the load estimation unit 54. In addition, the convergence determination PC 110 constitutes the convergence determination unit 60. Although Figure 13 shows an example in which one processing unit is configured with one piece of hardware (PC), multiple processing units may be configured with one piece of hardware. For example, the defect structure analysis unit 21 and the defect estimation unit 24 of the defect analysis unit 20 may be configured with one PC, and the load structure analysis unit 51 and the load estimation unit 54 of the load analysis unit 50 may be configured with one PC.
[0059] Next, an example of the operation of the defect inspection device 1b will be described. Figure 14 is a flowchart showing an example of the operation of the defect inspection device 1b according to Embodiment 3. In Figure 14, the same steps are assigned to the same processes as in the flowchart of Figure 3, which shows an example of the operation of the defect inspection device 1 according to Embodiment 1. The processes of steps S11 to S13 and S21 to S23, which are the same processes as in the flowchart of Figure 3, will not be explained.
[0060] After performing the surface deformation distribution acquisition step S10 and the defect estimation step S20, the defect inspection device 1b performs the load estimation step S40 to estimate the load applied to the object being inspected.
[0061] The load estimation step S40 is a process performed by the load analysis unit 50. In this load estimation step S40, first, the load structural analysis unit 51 performs a structural analysis using multiple load conditions (step S41) and registers the analysis results in a database (step S42). Specifically, the load structural analysis unit 51 performs a three-dimensional structural analysis using the shape of the object to be inspected, multiple load conditions, and a virtual defect shape to determine the deformation distribution of the surface of the object to be inspected for each of the multiple load conditions. Then, the load structural analysis unit 51 associates the determined deformation distribution with the load conditions and virtual defect shape used in the three-dimensional structural analysis and registers the analysis results in the load estimation database unit 53. Next, the load estimation unit 54 estimates the load applied to the object to be inspected (step S43). The load estimation unit 54 estimates the load applied to the object to be inspected based on the deformation distribution calculated in step S13 and the analysis results registered in the load estimation database unit 53.
[0062] After executing the load estimation step S40, the defect inspection device 1b determines whether the defect shape estimation process and the load estimation process have converged (step S50). This step S50 is a process executed by the convergence determination unit 60. If the defect shape estimation process and the load estimation process have not converged (step S50: No), the defect inspection device 1b returns to step S21 and executes the defect estimation step S20 and the load estimation step S40 again. On the other hand, if the defect shape estimation process and the load estimation process have converged (step S50: Yes), the defect inspection device 1b outputs the defect estimation result and the load estimation result (step S31). In this step S31, the estimation result output unit 30 outputs the shape of the defect on the inspection target estimated in step S23 and the load on the inspection target estimated in step S43.
[0063] As described above, the defect inspection device 1b according to Embodiment 3 performs the process of estimating the shape of the defect to be inspected and the process of estimating the load applied to the inspection target separately, thereby enabling efficient estimation of the shape of the defect and the load applied to the inspection target, and reducing the cost required for analysis. Furthermore, since the defect inspection device 1b repeatedly performs the process of estimating the shape of the defect and the process of estimating the load applied to the inspection target, the accuracy of estimating the shape of the defect and the load applied to the inspection target can be improved.
[0064] Embodiment 4. Figure 15 shows an example of the configuration of the defect inspection device 1c according to Embodiment 4. In Figure 15, the same reference numerals are used for components that are common to the defect inspection device 1b according to Embodiment 3 shown in Figure 12. The components that are common to the defect inspection device 1b will not be described. As shown in Figure 15, the defect inspection device 1c has a configuration in which a stress progression estimation unit 40 is added to the defect inspection device 1b according to Embodiment 3. This stress progression estimation unit 40 is the same as the stress progression estimation unit 40 provided in the defect inspection device 1a according to Embodiment 2 (see Figure 10).
[0065] In the defect inspection device 1c, the "load applied to the inspection target during defect progression" included in each of the multiple load conditions input by the data input unit 52 of the load analysis unit 50 to the load structural analysis unit 51 is the "stress that occurred when the defect in the inspection target progressed," which is estimated by the stress estimation unit 40 during defect progression. The operation of the load analysis unit 50 of the defect inspection device 1c is the same as the operation of the load analysis unit 50 of the defect inspection device 1b according to Embodiment 3. For this reason, the explanation of the operation of the load analysis unit 50 of the defect inspection device 1c is omitted.
[0066] As described above, the defect inspection apparatus 1c according to Embodiment 4 includes a stress estimation unit 40 that estimates the stress that occurred when a defect in the object under inspection propagates, and uses the stress estimated by the stress estimation unit 40 as the "load applied to the object under inspection when the defect propagates" included in the load conditions input to the load structural analysis unit 51. This makes it possible to reduce the effort and improve the accuracy of the three-dimensional structural analysis performed by the load structural analysis unit 51.
[0067] Embodiment 5. Figure 16 shows an example of the configuration of a defect inspection device 1d according to Embodiment 5. In Figure 16, the same reference numerals are used for components that are common to the defect inspection device 1b according to Embodiment 3 shown in Figure 12. The components that are common to the defect inspection device 1b will not be described. As shown in Figure 16, the defect inspection device 1d has a configuration in which a defect progression rate estimation unit 70 is added between the convergence determination unit 60 and the estimation result output unit 30 of the defect inspection device 1b according to Embodiment 3.
[0068] In the defect inspection device 1d, the convergence determination unit 60 outputs the defect estimation result and the load estimation result to the defect propagation rate estimation unit 70. The defect propagation rate estimation unit 70 estimates the defect propagation rate based on fracture mechanics, using the defect shape input as the defect estimation result and the load conditions input as the load estimation result. The defect propagation rate estimation unit 70 outputs the estimated defect propagation rate, along with the defect shape (defect estimation result) and load conditions (load estimation result) used to estimate the defect propagation rate, to the estimation result output unit 30.
[0069] Figure 17 shows an example of the hardware configuration of the defect inspection device 1d according to Embodiment 5. As shown in Figure 17, the defect inspection device 1d consists of a camera 101, a storage device 102, a PC for deformation measurement 103, a PC for defect estimation 104, a database for defect estimation 105, a PC for defect shape structure analysis 106, a PC for load estimation 107, a database for load estimation 108, a PC for load structure analysis 109, a PC for convergence determination 110, and a PC for defect propagation rate estimation 111. In Figure 17, the same reference numerals are used for hardware that is common to the hardware that constitutes the defect inspection device 1b according to Embodiment 3 shown in Figure 13.
[0070] The camera 101, storage device 102, and deformation measurement PC 103 constitute the surface deformation acquisition unit 10, while the defect estimation PC 104, defect estimation database 105, and defect shape structure analysis PC 106 constitute the defect analysis unit 20. Furthermore, the load estimation PC 107, load estimation database 108, and load structure analysis PC 109 constitute the load analysis unit 50. The defect propagation rate estimation PC 111 constitutes the defect propagation rate estimation unit 70. Although Figure 17 shows an example where one processing unit is configured with one piece of hardware (PC), multiple processing units may be configured with one piece of hardware.
[0071] Thus, according to the defect inspection device 1d of Embodiment 5, it is possible to obtain an estimated value of the rate of progression of defects occurring in the object being inspected.
[0072] Embodiment 6. Figure 18 shows an example of the configuration of a defect inspection device 1e according to Embodiment 6. In Figure 18, the same reference numerals are used for components that are common to the defect inspection device 1b according to Embodiment 3 shown in Figure 12. The components that are common to the defect inspection device 1b will not be described. As shown in Figure 18, the defect inspection device 1e has a configuration in which a repair determination unit 80 is added between the convergence determination unit 60 and the estimation result output unit 30 of the defect inspection device 1b according to Embodiment 3.
[0073] In the defect inspection device 1e, the convergence determination unit 60 outputs the defect estimation result and the load estimation result to the repair determination unit 80. The repair determination unit 80 estimates the defect progression rate using the defect shape input as the defect estimation result and the load conditions input as the load estimation result. The repair determination unit 80 also determines, based on the estimated progression rate, whether the inspection target will lose its function by the next inspection. If it loses its function, it outputs an inspection result indicating that the inspection target needs to be repaired; otherwise, it outputs an inspection result indicating that the inspection target does not need to be repaired. The repair determination unit 80 estimates the defect progression rate in the same way as the defect progression rate estimation unit 70 of the defect inspection device 1d according to Embodiment 5. If the repair determination unit 80 determines that the inspection target does not need to be repaired, it may also estimate the remaining time until the inspection target loses its function and output the estimation result.
[0074] Thus, the defect inspection device 1e according to Embodiment 6 can determine whether or not repair is necessary for defects occurring in the object being inspected and output it as an inspection result.
[0075] The configurations shown in the above embodiments are merely examples, and it is possible to combine them with other known technologies, combine different embodiments, and omit or modify parts of the configuration without departing from the gist of the invention.
[0076] 1, 1a, 1b, 1c, 1d, 1e Defect inspection device, 10 Surface deformation acquisition unit, 11 Imaging unit, 12 Image storage unit, 13 Deformation measurement unit, 20 Defect analysis unit, 21 Defect structural analysis unit, 22, 52 Data input unit, 23 Defect estimation database unit, 24 Defect estimation unit, 30 Estimation result output unit, 40 Stress estimation unit during propagation, 50 Load analysis unit, 51 Load structural analysis unit, 53 Load estimation database unit, 54 Load estimation unit, 60 Convergence determination unit, 70 Defect propagation rate estimation unit, 80 Repair determination unit.
Claims
1. A defect inspection apparatus comprising: an imaging unit that images the area of an inspection target in which a defect has occurred that includes the defect; a deformation measurement unit that determines the deformation distribution of the surface of the inspection target based on past and present images of the inspection target captured by the imaging unit; a defect estimation database unit that registers a plurality of virtual defect shapes which are the shapes of defects that are predicted to occur in the inspection target, and also registers the deformation distribution of the surface of the inspection target in the case where each of the plurality of virtual defect shapes occurs; and a defect estimation unit that estimates the shape of a defect occurring in the inspection target based on the deformation distribution determined by the deformation measurement unit and the virtual defect shapes and deformation distribution registered in the defect estimation database unit.
2. The defect inspection apparatus according to claim 1, characterized in that the defect estimation unit estimates the shape of a defect occurring in the object to be inspected using a statistical method that uses the deformation distribution obtained by the deformation measurement unit and the virtual defect shape and deformation distribution registered in the defect estimation database unit, or using a trained model generated by machine learning.
3. The defect inspection apparatus according to claim 1 or 2, further comprising: a deformation measurement unit that analyzes the deformation distribution obtained by the deformation measurement unit to estimate the plastic deformation region occurring at the tip of the defect, and a stress estimation unit that estimates the stress that occurred in the inspection target when the defect propagated based on the estimated plastic deformation region.
4. A defect inspection apparatus according to claim 1 or 2, comprising: a load estimation database unit which registers multiple load conditions indicating the load during defect progression and the load during inspection, and which registers the deformation distribution of the surface of the inspection target when each of the loads indicated by the multiple load conditions is applied to the inspection target; and a load estimation unit which estimates the load applied to the inspection target based on the deformation distribution obtained by the deformation measurement unit and the load conditions and deformation distribution registered in the load estimation database unit.
5. A defect inspection apparatus according to claim 4, comprising: a load structural analysis unit that performs a three-dimensional structural analysis using a plurality of load conditions registered in the load estimation database unit and the shape of the defect estimated by the defect estimation unit, and determines a deformation distribution to be registered in the load estimation database unit.
6. The defect inspection apparatus according to claim 5, comprising: an evolutionary stress estimation unit that analyzes the deformation distribution obtained by the deformation measurement unit to estimate the plastic deformation region occurring at the tip of the defect, and estimates the stress that occurred in the inspection target when the defect propagated based on the estimated plastic deformation region, wherein the load structural analysis unit uses the stress estimated by the evolutionary stress estimation unit in the three-dimensional structural analysis.
7. A defect inspection apparatus according to any one of 4 to 6, comprising: a load condition determination unit that determines whether the load conditions used when determining the deformation distribution to be registered in the defect estimation database unit and the load estimated by the load estimation unit satisfy a predetermined condition, wherein the defect estimation unit repeatedly performs estimation of the defect shape and the load estimation unit repeatedly performs estimation of the load by the load estimation unit until the convergence determination unit determines that the conditions are satisfied.
8. A defect inspection apparatus according to any one of 4 to 7, comprising a defect progression rate estimation unit that estimates the rate at which a defect is progressing in the object to be inspected based on the shape of the defect estimated by the defect estimation unit and the load estimated by the load estimation unit.
9. A defect inspection apparatus according to any one of 4 to 7, comprising: a repair determination unit that estimates the rate of progression of a defect occurring in the object to be inspected based on the shape of the defect estimated by the defect estimation unit and the load estimated by the load estimation unit, and determines whether or not the object to be inspected needs to be repaired based on the estimated rate of progression.
10. A defect inspection method performed by a defect inspection device for inspecting defects in an object to be inspected, comprising: a first step of photographing the area of the object to be inspected that includes the defect where the defect occurs; a second step of determining the deformation distribution of the surface of the object to be inspected based on past and present images of the object to be inspected taken in the first step; a third step of registering a plurality of virtual defect shapes, which are the shapes of defects predicted to occur in the object to be inspected, in a defect estimation database, and registering the deformation distribution of the surface of the object to be inspected in the defect estimation database when each of the plurality of virtual defect shapes occurs; and a fourth step of estimating the shape of the defect occurring in the object to be inspected based on the deformation distribution obtained in the second step and the virtual defect shapes and deformation distributions registered in the defect estimation database.
11. A defect inspection system comprising: a first device for measuring the deformation distribution of the surface of an object to be inspected; and a second device for estimating the shape of a defect occurring on the object to be inspected, wherein the first device comprises: an imaging unit for imaging the area of the object to be inspected that includes the defect where the defect occurs; and a deformation measurement unit for determining the deformation distribution of the surface of the object to be inspected based on past and present images of the object to be inspected taken by the imaging unit; and the second device comprises: a defect estimation database unit for which a plurality of virtual defect shapes, which are the shapes of defects predicted to occur on the object to be inspected, are registered, and for which the deformation distribution of the surface of the object to be inspected is registered when each of the plurality of virtual defect shapes occurs; and a defect estimation unit for estimating the shape of a defect occurring on the object to be inspected based on the deformation distribution determined by the deformation measurement unit and the virtual defect shapes and deformation distribution registered in the defect estimation database unit.
