Image processing device, image processing method, and image processing program

Abstract

According to the present invention: an acquisition unit 46 acquires, as an input image, a partial image including a target object 38 and a portion of a test object having a plurality of gradations whose positional relationship is known; a test object detection unit 48 detects the test object included in the input image; a gradation detection unit 50 detects the gradations included in the test object; an image size derivation unit 52 detects the size in the image of a pair of the gradations detected by the gradation detection unit 50; an actual size derivation unit 54 derives the actual size of the pair of gradations detected by the gradation detection unit 50; and an actual-dimension conversion information derivation unit 56 derives the derivation results by each of the image size derivation unit 52 and the actual size derivation unit 54 as actual-dimension conversion information.

Description

Image processing device, image processing method, and image processing program

[0001] The present disclosure relates to an image processing device, an image processing method, and an image processing program.

[0002] Japanese Patent Laid-Open Publication No. 5-203427 discloses an optical shape recognition device that includes a light source that illuminates a measurement line on the surface of an object from an oblique direction at a predetermined elevation angle; a scale that has graduations at predetermined intervals and is installed on the surface of the object so that the graduations are parallel to the measurement line; a photodetector that detects light reflected from the measurement line and outputs gray level image information made up of a plurality of pixel data; measurement line dividing means that divides the measurement line on the gray level image information into a plurality of blocks corresponding to each graduation interval of the scale; and interpolation means that calculates pixel data for each predetermined actual size interval on the block by interpolation from the pixel data that make up each of the divided blocks.

[0003] Japanese Patent Application Laid-Open No. 2014-57790 discloses an imaging diagnostic device that, on a calibration image obtained by imaging a calibration object whose actual dimensions are known, specifies a range within the region in which the calibration object is imaged that corresponds to the actual dimensions, generates objective calibration data that indicates the size per pixel that constitutes the calibration image using the number of pixels within the specified range and the actual dimensions of the calibration object, and performs measurement processing of physical quantities using an image of a measurement target.

[0004] Japanese Patent Laid-Open Publication No. 2004-275362 discloses an image actual size calculation method for acquiring image data carrying a radiographic image in which a scale capable of radiography is captured together with a subject, and calculating the actual size of the radiographic image based on the length indicated by the scale in the radiographic image represented by the acquired image data.

[0005] Japanese Patent Laid-Open Publication No. 2004-239731 discloses a radiological non-destructive inspection device that irradiates an object to be inspected with radiation, receives the radiation that has passed through the object to generate visible light in the red region, transmits the generated visible light while absorbing the radiation, and receives the transmitted visible light and converts it into an electrical signal, thereby obtaining image information of the object.

[0006] Japanese Patent Laid-Open Publication No. 2005-27904 discloses a radiological image display device having a magnification setting means for setting the magnification of the image, a measuring means for measuring the distance between predetermined positions on the image, a calculating means for calculating the distance between predetermined positions on the subject corresponding to the predetermined positions on the image based on the magnification set by the magnification setting means and the distance measured by the measuring means, and a display means for simultaneously displaying the magnification set by the magnification setting means, the distance measured by the measuring means, and the distance calculated by the calculating means.

[0007] Various techniques have been proposed to calculate the real-world size of a target object other than a test object from the size in an image and the real-world size of a test object whose real-world size is known. However, if the entire test object is not visible in the image, it is not possible to calculate the real-world size of the target object.

[0008] Therefore, the present disclosure aims to provide an image processing device, an image processing method, and an image processing program that are capable of calculating the real-world size of a target object even when the entire test object is not captured in the image.

[0009] In order to achieve the above object, an image processing device according to a first aspect of the present disclosure includes a processor, which detects a pair of scale marks from a partial image including a part of a test object in an input image in which the input image has as its subject a target object and a test object having a plurality of scale marks whose positional relationship is known, and derives actual size conversion information based on the size of the pair of scale marks on the image and the real-world size of the pair of scale marks.

[0010] An image processing device according to a second aspect of the present disclosure is the image processing device according to the first aspect, wherein the input image is a radiographic image for non-destructive testing, and the test object is at least one of a multi-linear image quality meter and a perforated transmittance meter.

[0011] An image processing device according to a third aspect of the present disclosure is the image processing device according to the first aspect, in which the processor detects, as a pair of scale marks, a pair of scale marks included in the partial image, excluding the pair of scale marks whose distance on the image is the shortest.

[0012] An image processing device according to a fourth aspect of the present disclosure is an image processing device according to the first aspect, in which a processor derives actual size conversion information for a specified position by interpolating the positions of a pair of scale marks and actual size conversion information derived from each of a plurality of partial images.

[0013] An image processing device according to a fifth aspect of the present disclosure is the image processing device according to the first aspect, in which a processor derives the size on the image and the real-world size of an object whose real-world size is known and which is larger than the size on the image of a pair of scales as actual size conversion information.

[0014] An image processing device according to a sixth aspect of the present disclosure is the image processing device according to the first aspect, wherein the processor detects distortion of the test object and corrects the positions of the pair of scales on the image based on the distortion.

[0015] An image processing device according to a seventh aspect of the present disclosure is the image processing device according to the sixth aspect, wherein when a distortion of the test object is detected, the processor selects a pair of scales excluding the pair with the furthest distance on the image.

[0016] An image processing device according to an eighth aspect of the present disclosure is the image processing device according to the sixth aspect, wherein the processor calculates distortion correction information based on a change in the distance between the scale marks.

[0017] An image processing device according to a ninth aspect of the present disclosure is the image processing device according to the sixth aspect, wherein the processor treats the position of the scale on the image as a known variable, the position of the light source relative to the light receiving surface in the real world, the position of the test object relative to the light receiving surface in the real world, and the rotation angle of the test object relative to the light receiving surface in the real world as unknown variables, and treats the unknown variables as distortion correction information.

[0018] An image processing device according to a tenth aspect of the present disclosure is the image processing device according to the first aspect, wherein the processor detects a portion of the test object from the input image.

[0019] An image processing device according to an eleventh aspect of the present disclosure is the image processing device according to the first aspect, wherein the processor calculates the real-world size of the target object from the actual size conversion information and the size of the target object on the image.

[0020] An image processing device according to a twelfth aspect of the present disclosure is an image processing device according to the first aspect, in which a processor calculates the real-world size of the target object from actual size conversion information and the size of the target object on the image, and superimposes the real-world size of the target object and the actual size conversion information on the input image.

[0021] An image processing device according to a thirteenth aspect of the present disclosure includes a processor, which detects scale marks from a partial image including a part of a test object in an input image including the test object having a plurality of scale marks whose positional relationship is known, and calculates distortion correction information for correcting distortion of the test object from the positions of the detected scale marks on the image.

[0022] An image processing method according to a fourteenth aspect of the present disclosure involves a computer detecting a pair of scale marks from a partial image including a part of a test object in an input image in which the test object has a target object and a plurality of scale marks whose positional relationship is known as its subject, and deriving actual size conversion information based on the size of the pair of scale marks on the image and the real-world size of the pair of scale marks.

[0023] An image processing program according to a fifteenth aspect of the present disclosure causes a computer to execute a process of detecting a pair of scale marks from a partial image including a part of a test object in an input image having a target object and a test object having a plurality of scale marks whose positional relationship is known as subjects, and deriving actual size conversion information based on the size of the pair of scale marks on the image and the real-world size of the pair of scale marks.

[0024] According to the present disclosure, it is possible to provide an image processing device, an image processing method, and an image processing program that are capable of calculating the real-world size of a target object even when the entire test object is not captured in the image.

[0025] 1 is a block diagram showing an example of the overall configuration of a radiographic imaging system according to the present embodiment. FIG. 2 is a block diagram showing an example of the configuration of a main part of a control device. FIG. 3 is a schematic diagram of a method of simultaneously imaging test objects and converting them into actual size. FIG. 4 is a schematic diagram of a test object that is not entirely captured. FIG. 5 is a diagram showing an example of a test object having a scale. FIG. 6 is a diagram showing variations of test objects having scales. FIG. 7 is a functional block diagram showing the functional configuration of a control unit in a control device of a radiographic imaging system according to the first embodiment. FIG. 8 is a flowchart showing an example of the flow of processing performed by the control device of the radiographic imaging system. FIG. 9 is a diagram showing an example of a display of the real-world size of a target object. FIG. 10 is a diagram for explaining a situation where interpolation is required when deriving actual size conversion information of an arbitrary position. FIG. 11 is a flowchart showing an example of the flow of processing performed by the control device of the radiographic imaging system according to the present embodiment to derive actual size conversion information of an arbitrary position. FIG. 12 is a diagram showing an example of a display when deriving actual size conversion information using a non-test object when there is a known non-test object that is larger than the test object. FIG. 13 is a flowchart showing an example of the flow of processing performed by the control device of the radiographic imaging system according to the present embodiment to derive actual size conversion information. FIG. 14 is a diagram for explaining a situation where consideration of distortion of the test object is required. FIG. 15 is a diagram showing an example of the undistorted shape and distorted shape of a test object. FIG. 16 is a flowchart showing an example of the flow of processing performed by the control device of the radiographic imaging system according to the present embodiment to generate distortion correction information. FIG. 10 is a schematic diagram for explaining an example of a distortion model.

[0026] Hereinafter, embodiments of the present disclosure will be described in detail with reference to the drawings. However, the present invention is not limited to the embodiments.

[0027] An embodiment of the present disclosure relates to a technique for converting an image size to a real-world size, particularly using a test object whose real-world size is known. That is, the real-world size of an object other than the test object is calculated from the image size and real-world size of the test object shown in the image.

[0028] In this disclosure, "size" refers to length and / or area. Size in this disclosure is broadly divided into size on an image and size in the real world. An example of a unit of size on an image is pixels, etc. An example of a unit of size in the real world is meters, etc.

[0029] Furthermore, in a broad sense, "actual size conversion" includes not only the above processing but also the function of storing and displaying actual size conversion information.

[0030] In addition, in this disclosure, "actual size conversion information" refers to numerical values ​​required for actual size conversion. It may be at least one of a pair of an image size and a real-world size, or a ratio of an image size and a real-world size. For ease of explanation to users, the position of the test object and the position of its scale, which are the basis for the calculation, may also be included.

[0031] In addition, in this disclosure, a "test object" refers to an object with a known shape and size. In particular, in this disclosure, it refers to an object with a scale. A new test object may be created for the purpose of this disclosure, or an "Image Quality Indicator (IQI)" test object used for image quality evaluation in NDT (Non-destructive Testing) may be utilized. For example, a radiographic image from a non-destructive test may be used as the input image, and at least one of a multi-linear image quality meter and a perforated penetrometer may be used as the test object.

[0032] It should be noted that the "scale" is not necessarily limited to a line pattern, but may be a dot, etc. In other words, it is sufficient that the distance between any two scales is known or can be calculated.

[0033] First Embodiment In this embodiment, a radiographic image capturing system including a control device as an example of an image processing device will be described as an example. Fig. 1 is a block diagram showing an example of the overall configuration of the radiographic image capturing system according to this embodiment.

[0034] As shown in FIG. 1, the radiation image capturing system 10 includes a radiation irradiation device 12, a radiation image capturing device 16, and a control device 18.

[0035] The radiation irradiator 12 according to this embodiment includes a radiation source 14 that irradiates a subject W, which is an example of an imaging target, with radiation R, such as X-rays (X-rays). The method of instructing the radiation irradiator 12 to irradiate radiation R is not particularly limited. For example, if the radiation irradiator 12 includes an irradiation button or the like, a user such as a radiologist may issue an instruction to irradiate radiation R using the irradiation button, thereby causing the radiation irradiator 12 to irradiate radiation R. Alternatively, for example, a user such as a radiologist may issue an instruction to irradiate radiation R by operating the control device 18, thereby causing the radiation irradiator 12 to irradiate radiation R.

[0036] When the radiation irradiator 12 receives an instruction to irradiate radiation R, it irradiates radiation R from the radiation source 14 in accordance with the irradiation conditions, such as the set tube voltage, tube current, and irradiation period.

[0037] The radiographic imaging device 16 according to this embodiment includes a radiation detector 20 that detects radiation R that is irradiated from the radiation irradiator 12 and that has passed through the subject W. The radiographic imaging device 16 uses the radiation detector 20 to capture a radiographic image of the subject W (the radiographic imaging device 16 generates a radiographic image showing the subject W by capturing an image of the subject W using the radiation detector 20).

[0038] 2, the control device 18 includes a control unit 22, a display unit 32, an operation unit 34, and a communication I / F (interface) unit 36. FIG. 2 is a block diagram showing an example of the configuration of the main parts of the control device 18.

[0039] The main part of the control unit 22 is configured as a general computer, and controls the overall operation of the control device 18 .

[0040] The control unit includes a CPU 24, a ROM 26, and a RAM 28. The ROM 26 pre-stores various programs 30 executed by the CPU 24, such as an image capture processing program and a display processing program executed when controlling image capture. The RAM 28 temporarily stores various data.

[0041] The display unit 32 displays a screen for operating the radiation irradiating device 12 and the radiographic image capturing device 16, as well as captured radiographic images (radiographic images obtained by capturing images), etc. The display unit 32 is, for example, a direct-view electronic display.

[0042] The operation unit 34 is used by the user to input instructions and various information related to photography, etc. The operation unit 34 is not particularly limited, and examples thereof include various switches, a touch panel, a touch pen, and a mouse.

[0043] The communication I / F unit 36 ​​communicates captured images such as radiographic images and various information between the radiation irradiating device 12 and the radiographic image capturing device 16 via wireless or wired communication.

[0044] In business processes such as image inspection and image analysis, there are cases where it is necessary to measure the real-world size of a target object contained in an input image. For example, in the case of inspection, a pass / fail decision may be made based on whether or not the shape of a recognized abnormality exceeds a predetermined standard value. In the case of analysis, the cause of a phenomenon may be inferred from the shape information of the target pattern. In both cases, the premise is the quantification of shape, which is often the real-world size.

[0045] When converting the size on an image to a size in the real world, if information about the imaging system, such as the relative positions of the light source, target object, and light receiving surface, as well as the three-dimensional structure of the target object, is known, the size on the captured image (the image obtained by capturing) can be converted to a size in the real world.

[0046] However, since it is difficult to always keep the imaging system under precise control, it is difficult to realize precise actual size conversion from information on the imaging system.

[0047] Therefore, there is a method to achieve real-world size conversion by simultaneously photographing a test object whose real-world size is known.

[0048] For example, Figure 3 shows a schematic diagram of a method for simultaneously photographing a test object and converting it to actual size. Instead of photographing only the target object 38 as shown on the left side of Figure 3, a test object 40 is placed near the target object 38 and photographed so that it appears in the same image as the target object 38, as shown on the right side of Figure 3. To determine the real-world size of the deformed portion 42, for example, the calculation can be made as follows: real-world size of the deformed portion 42 = size of the deformed portion 42 on the image × (real-world size of the test object 40 / size of the test object 40 on the image).

[0049] However, in the case of radiographic images of this embodiment, the entire test object 40 may not always be captured, depending on the imaging. Fig. 4 shows a schematic diagram of a test object 40 that is not captured in its entirety. The test object 40 originally has an overall shape as shown in the upper part of Fig. 4, but there are situations where the overall shape cannot be determined, as shown in the lower part of Fig. 4. In such situations, the size of the test object 40 on the image cannot be measured, and conversion to actual size cannot be achieved.

[0050] Specifically, this refers to a situation in which the overall shape of the test object 40 cannot be determined due to low contrast, the presence of an obstacle (e.g., occlusion), etc. For example, in radiography as in this embodiment, contrast adjustment is difficult, and the entire test object 40 may not be captured due to the presence of an internal structure of the object that cannot be seen during imaging. In this way, if the entire test object 40 is not included in the image, it is not possible to convert it to actual size.

[0051] Therefore, in the radiographic imaging system 10 according to this embodiment, the control device 18 detects a pair of scale marks from a partial image including a part of the test object 40 in an input image including the target object 38 and the test object 40 having multiple scale marks whose positional relationship is known, and performs a derivation process to derive actual size conversion information based on the size of the detected pair of scale marks on the image and the real-world size of the pair of scale marks.

[0052] That is, in this embodiment, the individual scales of the test object 40 with scales included in the partial image are used instead of the overall shape of the test object 40. Specifically, the test object 40 with scales is utilized, and individual scales are detected from the partial image, and pairs of scales are detected. Then, actual size conversion information is derived based on the detected pairs of scales. In this way, actual size conversion information can be derived based on the individual scales even if the test object 40 is only partially visible.

[0053] For example, a test object 40 having a scale as shown in Fig. 5 is used. Even in a situation where the overall shape of the test object 40 cannot be determined, as shown in the middle of Fig. 5, it is possible to detect individual scales from a partial image 44 containing part of the test object 40, as shown in the lower part of Fig. 5, and associate the size of the test object 40 on the image with its size in the real world based on the distance between the scales.

[0054] The size of the test object 40 on the image may be determined as long as the positional relationship of the scale is known. Furthermore, various variations of the test object 40 with the scale are conceivable. For example, an IQI (Image Quality Indicator) used in X-ray nondestructive testing may be used.

[0055] 6 shows variations of a test object 40 with a scale. In the example of Fig. 6, a test object 40 with a linear scale, a test object 40 with a pair of linear scales, a test object 40 with a circular scale, a test object 40 with a star-shaped scale, a test object 40 with linear scales both vertically and horizontally, a test object 40 with a grid-shaped scale, a test object 40 with a honeycomb-shaped scale, and a test object 40 with a concentric circular scale are shown as examples.

[0056] Furthermore, a partial image 44 including a part of a test object 40 in an input image including a target object 38 and a test object 40 having multiple scales whose positional relationship is known is not an image in which parts other than the scales are missing, but an image in which part of the scales are missing.

[0057] Furthermore, individual scale marks can be detected using an automatic method such as pattern recognition technology. Examples include template matching, feature point extraction, and machine learning. For a test object 40 with scale marks arranged continuously in a specific direction, such as an IQI used in X-ray nondestructive testing, a profile of pixel values ​​can be obtained in that direction and the scale marks can be detected from the characteristic shape of the profile.

[0058] Furthermore, it is desirable to detect a pair of scales from among the scales included in a partial image 44 including a portion of the test object 40 in an input image including the target object 38 and a test object 40 having multiple scales with known positional relationships, excluding the pair of scales with the closest distance on the image. It is even more desirable to detect a pair of scales from among the scales included in a partial image 44 including a portion of the test object 40 in an input image including the target object 38 and the test object 40, the pair of scales being farther than the distance from the edge to the center. In particular, employing the pair of scales that are farthest apart minimizes measurement errors, thereby stabilizing conversion accuracy. In other words, in digital images, information is discretized in pixel units, and information on sizes smaller than a pixel is lost. By using a larger image size, this effect can be reduced.

[0059] In this embodiment, in order to perform the above-described derivation process, the CPU 24 executes a program 30 stored in the ROM 26 of the control unit 22, thereby achieving the functions shown in Fig. 7. Fig. 7 is a functional block diagram showing the functional configuration of the control unit 22 in the control device 18 of the radiographic imaging system 10 according to this embodiment.

[0060] That is, as shown in FIG. 7 , the control unit 22 has the functions of an acquisition unit 46, a test object detection unit 48, a scale detection unit 50, an image size derivation unit 52, an actual size derivation unit 54, an actual size conversion information derivation unit 56, an image size measurement unit 58, an actual size conversion unit 60, and a presentation unit 62.

[0061] The acquisition unit 46 acquires, as an input image, a radiographic image captured by the radiographic imaging device 16 (a radiographic image obtained by performing imaging using the radiographic imaging device 16). In particular, the acquisition unit 46 acquires, as an input image, a partial image 44 including the target object 38 and a part of the test object 40 having a plurality of scales whose positional relationship is known. As described above, a radiographic image from non-destructive testing can be used as an example of the input image.

[0062] The test object detection unit 48 detects the test object 40 included in the input image acquired by the acquisition unit 46. For example, the detection is performed using a method represented by pattern recognition techniques such as template matching, feature point extraction, and machine learning. Note that the user may specify an area of ​​the test object 40, and the specified area may be detected as the test object 40. Alternatively, the test object 40 may be placed in a predetermined location, and a predetermined area in the input image may be detected as the test object 40.

[0063] The scale detection unit 50 detects scales included in the test object 40. For example, the scale detection unit 50 uses a method such as template matching, feature point extraction, or machine learning to detect each of the scales included in the test object 40. Furthermore, the scale detection unit 50 identifies the pair of scales that are the furthest apart from each other among the detected scales.

[0064] The image size deriving unit 52 detects the size of the pair of scale marks on the image detected by the scale detection unit 50. For example, the image size deriving unit 52 derives the distance between the pair of scale marks by counting the number of pixels between the pair of scale marks.

[0065] The real size derivation unit 54 derives the real size of the pair of scale marks detected by the scale detection unit 50. That is, it derives the real-world size of the pair of scale marks by referring to information about the known scale marks of the test object 40. The known size may be input by the user, or may be stored in advance and read out.

[0066] The actual size conversion information deriving section 56 derives the results derived by the image size deriving section 52 and the actual size deriving section 54 as actual size conversion information.

[0067] The image size measurement unit 58 measures the size of the target object 38 on the image. For example, the user specifies the target object 38 to be measured, and the image size of the specified target object 38 is measured. Alternatively, the target object 38 may be specified by an automatic method such as a pattern recognition technique.

[0068] The actual size conversion unit 60 derives the size of the target object 38 in the real world using the actual size conversion information derived by the actual size conversion information derivation unit 56 and the measurement results of the size of the target object 38 on the image measured by the image size measurement unit 58.

[0069] The presentation unit 62 presents the real-world size of the target object 38 to the user by displaying the real-world size of the target object 38 derived by the actual size conversion unit 60 on the display unit 32 .

[0070] Next, specific processing performed in the radiographic imaging system 10 according to this embodiment configured as described above will be described. Fig. 8 is a flowchart showing an example of the flow of processing performed by the control device 18 of the radiographic imaging system 10. Note that the processing in Fig. 8 starts, for example, when an instruction is given to measure an object included in a radiographic image captured by the radiographic imaging device 16 (a radiographic image obtained by performing imaging using the radiographic imaging device 16).

[0071] In step 100, the CPU 24 acquires an input image and proceeds to step 102. That is, the acquisition unit 46 acquires, as an input image, a radiographic image captured by the radiographic imaging device 16 (a radiographic image obtained by imaging using the radiographic imaging device 16). The input image includes the target object 38 and a test object 40 having a plurality of scales whose positional relationship is known.

[0072] In step 102, the CPU 24 detects the test object 40 and proceeds to step 104. That is, the test object detection unit 48 detects the test object 40 included in the input image acquired by the acquisition unit 46. For example, the detection is performed using a method typified by pattern recognition techniques such as template matching, feature point extraction, and machine learning. Note that the user may specify an area of ​​the test object 40, and the specified area may be detected as the test object 40.

[0073] In step 104, the CPU 24 detects the scale marks on the test object 40 and proceeds to step 106. That is, the scale mark detection unit 50 detects the scale marks included in the test object 40. For example, the CPU 24 detects the individual scale marks included in the test object 40 using a method such as template matching, feature point extraction, or machine learning. Note that if the number of scale marks detected is one or less, the CPU 24 suspends the subsequent processing and displays an alert to that effect to the user.

[0074] In step 106, the CPU 24 identifies a pair of scale marks and proceeds to step 108. That is, the scale detection unit 50 identifies the pair of scale marks that are the furthest apart from the detected scale marks. Here, an example will be described in which the pair of scale marks that are the furthest apart from each other is identified, but it is sufficient to detect a pair of scale marks that are the closest apart from the pair of scale marks included in the input image.

[0075] In step 108, the CPU 24 derives the size of the pair of scale marks on the image, and then proceeds to step 110. That is, the image size deriving unit 52 derives the size of the pair of scale marks on the image detected by the scale detection unit 50. For example, the distance between the pair of scale marks is derived by counting the number of pixels between the pair of scale marks.

[0076] In step 110, the CPU 24 derives the real size of the pair of scale marks and proceeds to step 112. That is, the real size deriving unit 54 derives the real size of the pair of scale marks detected by the scale detection unit 50. That is, by referring to information about the scale marks of the test object 40, which is known, the real-world size of the pair of scale marks is derived. The known size may be input by the user, or may be stored in advance and then read out.

[0077] In step 112, the CPU 24 derives actual size conversion information, and proceeds to step 114. That is, the actual size conversion information deriving unit 56 derives the results of the image size deriving unit 52 and the actual size deriving unit 54 as actual size conversion information.

[0078] In step 114, the CPU 24 determines whether or not the target object 38 in the input image has been designated. This determination determines whether or not the target object 38 to be measured has been designated by the user. The process waits until the determination is affirmative, and then proceeds to step 116.

[0079] In step 116, the CPU 24 measures the image size of the specified target object 38, and then the process proceeds to step 118. That is, the image size measurement unit 58 measures the number of pixels as an example of the size of the target object 38 on the image.

[0080] In step 118, the CPU 24 performs an actual size conversion of the target object 38, and then proceeds to step 120. That is, the actual size conversion unit 60 derives the size of the target object 38 in the real world using the actual size conversion information derived by the actual size conversion information derivation unit 56 and the measurement results of the size of the target object 38 on the image measured by the image size measurement unit 58.

[0081] In step 120, the CPU 24 displays the conversion result and proceeds to step 122. That is, the presentation unit 62 displays the real-world size of the target object 38 derived by the actual size conversion unit 60 on the display unit 32, thereby presenting the real-world size of the target object 38 to the user. For example, as shown in FIG. 9 , the real-world size of the target object 38 is displayed. In addition to the real-world size of the target object 38, as shown in FIG. 9 , the size of the target object 38 on the image, and the real-world and image sizes of the test object 40 may also be displayed. Note that, in addition to the real-world size of the target object 38, at least one of the size on the image of the target object 38, the real-world size of the test object 40, and the size on the image of the test object 40 may also be displayed.

[0082] In step 122, the CPU 24 determines whether or not another target object 38 has been designated. If the determination is affirmative, the process returns to step 116 to repeat the above-described processing. If the determination is negative because another target object has not been designated, the process proceeds to step 124.

[0083] In step 124, the CPU 24 determines whether or not the process is to be terminated. This determination is made by determining whether or not an instruction to terminate the process has been issued. If the determination is negative, the process returns to step 122 and the above-described process is repeated. If the determination is positive, the process ends.

[0084] Second Embodiment A process performed by the control device 18 according to the second embodiment will be described. In this embodiment, an example will be described in which actual size conversion information at an arbitrary position is calculated by interpolation from a plurality of test objects 40 in the radiographic image capturing system 10 described in the first embodiment.

[0085] A situation requiring interpolation will be described with reference to Fig. 10. Fig. 10 is a diagram for explaining a situation requiring interpolation when deriving actual size conversion information at an arbitrary position.

[0086] In the case of transmission photography as in this embodiment, light rays are emitted radially from the light source 64, so even if the size in the real world is the same, an object closer to the light source 64 appears larger in the image on the light receiving surface 66.

[0087] 10, it may not be possible to place the test object 40 at the same distance from the light source 64 due to constraints imposed by the three-dimensional structure of the target object 38. In such a situation, the actual size conversion information calculated from the test object 40 shown in the image varies depending on the position, so it is desirable to interpolate the actual size conversion information depending on the position, especially when the surface of the target object 38 is of interest.

[0088] The interpolation method may be a simple calculation method such as the following formula, for example.

[0089] r ｐ = p.r. ０ +(1-p)・r １

[0090] r ０ , r １ is the ratio of actual size of test object 0 and test object 1.

[0091] p is the position on the line connecting test object 0 and test object 1. Normalized

[0092] r ｐ is the ratio of actual size at position p.

[0093] Next, specific processing performed in the radiographic imaging system 10 according to this embodiment will be described. Fig. 11 is a flowchart showing an example of the flow of processing performed in the control device 18 of the radiographic imaging system 10 according to this embodiment to derive actual size conversion information at an arbitrary position. Note that the processing in Fig. 11 is performed instead of, for example, steps 100 to 112 in Fig. 8.

[0094] In step 200, the CPU 24 acquires an input image and proceeds to step 202. That is, the acquisition unit 46 acquires, as an input image, a radiographic image captured by the radiographic imaging device 16 (a radiographic image obtained by imaging using the radiographic imaging device 16). The input image includes the target object 38 and a test object 40 having a plurality of scales whose positional relationship is known.

[0095] In step 202, the CPU 24 detects the test object 40 and proceeds to step 204. That is, the test object detection unit 48 detects the test object 40 included in the input image acquired by the acquisition unit 46. For example, the detection is performed using a method typified by pattern recognition techniques such as template matching, feature point extraction, and machine learning. Note that the user may specify an area of ​​the test object 40, and the specified area may be detected as the test object 40.

[0096] In step 204, the CPU 24 determines whether or not a plurality of test objects 40 have been detected. If the determination is negative, the process proceeds to step 206, and if the determination is affirmative, the process proceeds to step 208.

[0097] In step 206, the CPU 24 performs actual size conversion information derivation processing, and then proceeds to step 210. The actual size conversion information derivation processing performs the processing of steps 104 to 112 described above.

[0098] Meanwhile, in step 208, the CPU 24 performs processing to derive actual size conversion information including position information, and then proceeds to step 210. The processing to derive actual size conversion information including position information performs the processing of step 206 on a plurality of test objects 40 to derive actual size conversion information, detects the position of each test object 40, and includes the detected position information in the actual size conversion information.

[0099] In step 210, the CPU 24 determines whether or not the actual size converted position has been designated. The CPU 24 waits until the determination is affirmative, and then proceeds to step 212.

[0100] In step 212, the CPU 24 determines whether or not there is actual size conversion information for the specified position. If the result is negative, the process proceeds to step 214, and if the result is positive, the process proceeds to step 216. For example, if the detection of multiple test objects 40 is negative in step 204 and actual size conversion information for one test object 40 is derived in step 206, the determination in step 212 that there is no actual size conversion information for the specified position is denied, and the process proceeds to step 214.

[0101] In step 214, the CPU 24 selects the actual size conversion information for the specified position and ends the process of deriving actual size conversion information for a series of arbitrary positions. For example, if actual size conversion information for one test object 40 is derived in step 206, that one actual size conversion information is selected.

[0102] On the other hand, in step 216, the CPU 24 ends the process of interpolating the actual size conversion information of the designated position to derive the actual size conversion information of a series of arbitrary positions. For example, the actual size conversion information of the designated position is interpolated using the interpolation method described above. In this way, the actual size conversion information of the arbitrary position can be obtained.

[0103] Third Embodiment A process performed by the control device 18 according to a third embodiment will be described. In this embodiment, when deriving actual size conversion information, if there is a known non-test object that is larger than the test object 40, the actual size conversion information is derived using the non-test object.

[0104] To improve the accuracy of the actual size conversion, it is desirable to use a larger size on the image. In this embodiment, if information on a non-test object having a size on the image larger than that of the test object 40 is obtained, this information is used for the actual size conversion.

[0105] Specifically, the size of a pair of scales on the test object 40 is compared with the size of a non-test object, and the larger one is adopted. For example, in parts manufacturing and / or piping inspection, the size of the object to be image-inspected is often known, so a portion of a known size is used.

[0106] An example of execution of this embodiment is shown in Fig. 12. Fig. 12 shows an example in which the actual size conversion information of the non-test object 68 is used, and "Ratio: x.xx = xxx [mm] / xxx [px]" is displayed as the actual size conversion information of the non-test object 68.

[0107] It is also possible to accept a user instruction to adopt the test object 40 .

[0108] In addition, the real-world size of the non-test object 68 may be input by the user, or an automatic method may be used in which the approximate size of the object is calculated from the standard resolution information of the image and the closest size corresponding to the object's specifications is adopted.

[0109] The size of the non-test object 68 on the image may be input by the user or may be measured by an automatic means such as a pattern recognition technique.

[0110] Next, specific processing performed in the radiographic imaging system 10 according to this embodiment will be described. Fig. 13 is a flowchart showing an example of the flow of processing for deriving actual size conversion information, which is performed in the control device 18 of the radiographic imaging system 10 according to this embodiment. Note that the processing in Fig. 13 is performed instead of, for example, steps 100 to 112 in Fig. 8.

[0111] In step 300, the CPU 24 acquires an input image and proceeds to step 302. That is, the acquisition unit 46 acquires, as an input image, a radiographic image captured by the radiographic imaging device 16 (a radiographic image obtained by imaging using the radiographic imaging device 16). The input image includes the target object 38 and a test object 40 having a plurality of scales whose positional relationship is known.

[0112] In step 302, the CPU 24 determines whether or not there is a non-test object 68 of known size in the input image. If the determination is negative, the process proceeds to step 304, and if the determination is affirmative, the process proceeds to step 306.

[0113] In step 304, the CPU 24 completes the process of deriving a series of actual size conversion information by performing the process of deriving actual size conversion information for the test object 40. The process of deriving actual size conversion information for the test object 40 includes the processes of steps 104 to 112 described above.

[0114] On the other hand, in step 306, the CPU 24 determines whether the non-test object 68 is larger than the test object 40. If the determination is negative, the process proceeds to step 304 described above, and if the determination is positive, the process proceeds to step 308.

[0115] In step 308, the CPU 24 completes the process of deriving a series of actual size conversion information by performing the actual size conversion information derivation process for the non-test object 68. The actual size conversion information derivation process for the non-test object 68 derives the actual size conversion information using the non-test object 68 whose size is known.

[0116] Fourth Embodiment A control device 18 according to a fourth embodiment will be described. In this embodiment, the distortion of the test object 40 is taken into consideration when making corrections.

[0117] A situation in which consideration of distortion of the test object 40 is required will be described with reference to Fig. 14. Fig. 14 is a diagram for explaining a situation in which consideration of distortion of the test object 40 is required.

[0118] Assuming the imaging system shown in Fig. 14, when the test object 40 is not tilted with respect to the light-receiving surface 66, as shown on the left side of Fig. 14, no distortion occurs in the shape of the image of the test object 40. On the other hand, when the test object 40 is tilted with respect to the light-receiving surface 66, as shown in the center of Fig. 14, distortion occurs in the shape of the image of the test object 40. If the actual size conversion information is calculated without taking this distortion into consideration, an error will occur.

[0119] Therefore, in this embodiment, distortion correction information for the test object 40 is calculated from the detected scales, and the positions of the scales are corrected, and then a pair of scales is identified.

[0120] Note that distortion correction may be completed in advance for a partial image 44 of the test object 40. Furthermore, distortion correction does not necessarily have to be based on the scale; it is also possible to calculate distortion correction information from the shape of the test object 40. For example, as shown in FIG. 15 , this can be achieved using a general method by using the undistorted shape shown in the upper part of FIG. 15 , which is prior information on the test object 40, and the distorted shape shown in the image in the lower part of FIG. 15 . Even if the overall shape of the test object 40 cannot be determined, distortion correction information can be sufficiently calculated from the degree of distortion of straight lines, etc.

[0121] Here, an example of a specific method for distortion correction will be described. Here, the situation shown in the center of Fig. 14 is assumed. The scales on the test object 40 are arranged in one dimension. For simplicity, it is assumed that the scales are arranged at equal intervals.

[0122] As an example of approximation, as shown on the left side of FIG. 14, the position of the kth scale on the image is expressed as p ｋ First, the distance between the scale marks on the image, p ｋ＋１ -p ｋ Next, by calculating the amount of change in distance on the image using regression or the like, the distortion model k and the scale position P' are calculated. ｋ The following equation is obtained as the relationship between

[0123] P' ｋ＋１ -P' ｋ = a k + b

[0124] The intercept b may be determined by regression, and P １ -p ０ etc. may also be used.

[0125] Also, the value p' calculated from the equation obtained by regression ｋ and the measured value p on the image ｋ It distinguishes between the following.

[0126] This is used as distortion correction information, ｋ and / or P' ｋ From the above, the position of the scale on the image after distortion correction, P^ ｋ This is calculated as p^ ｋ This is possible because the relationship between k and k is obvious. For example, the following formula can be used.

[0127] P^ ｋ = p ０ +kb

[0128] Measurement value p ０ The values ​​are arranged at equal intervals with a width b determined by regression, based on p'. ｋ →k becomes a quadratic equation. In this case, k is not necessarily a natural number.

[0129] The distortion modeling is not limited to the above, and may be more sophisticated. For example, p' may be calculated by referring to the position information of the light source 64 and / or the light receiving surface 66. ｋ＋１ -p' ｋ One possible method is to make it a nonlinear function.

[0130] Next, specific processing performed in the radiographic imaging system 10 according to this embodiment will be described. Fig. 16 is a flowchart showing an example of the flow of processing for generating distortion correction information, which is performed in the control device 18 of the radiographic imaging system 10 according to this embodiment. Note that the processing in Fig. 16 is performed, for example, between step 104 and step 106 in Fig. 8.

[0131] In step 400, the CPU 24 estimates the distortion and proceeds to step 402. That is, the distortion is estimated from the scale of the test object 40 detected in step 104.

[0132] In step 402, the CPU 24 determines whether or not there is a scale distortion. If the determination is affirmative, the process proceeds to step 404, and if negative, the process of generating a series of distortion correction information is terminated.

[0133] In step 404, the CPU 24 outputs a distortion alert and proceeds to step 406. For example, the alert is output by displaying a message indicating that the scale is distorted.

[0134] In step 406, the CPU 24 generates distortion correction information and proceeds to step 408. For example, the distortion correction information is generated by the method described above.

[0135] In step 408, the CPU 24 corrects the distortion of the scale and ends the process of generating a series of distortion correction information. That is, the CPU 24 corrects the distortion of the scale based on the generated distortion correction information.

[0136] Fifth Embodiment A control device 18 according to a fifth embodiment will be described. In this embodiment, another method of correcting distortion in the fourth embodiment will be described.

[0137] When the test object 40 is parallel to the light-receiving surface 66, the projected image of the test object 40 on the image is similar to the shape of the test object 40 in the real world, and is simply enlarged. However, when the test object 40 is tilted with respect to the light-receiving surface 66, the projected image of the test object 40 on the image is distorted from the shape of the test object 40 in the real world.

[0138] Therefore, the tilt of the test object 40 is determined from the distortion of the projected image of the test object 40, and if it is tilted, correction is performed to obtain actual size conversion information.

[0139] When the test object 40 is tilted, its projected image is distorted depending on the position of the light source 64 and the position and tilt of the test object 40. Therefore, the position of the light source 64 and the position and tilt of the test object 40 are identified based on the distortion of the projected image. Specifically, a total of nine unknown parameters are identified: the XYZ position of the light source 64, and the XYZ position and rotation angle around the three XYZ axes of the test object 40. Then, actual size conversion information is calculated. A specific method for deriving actual size conversion information is described below. FIG. 17 is a schematic diagram for explaining an example of a distortion model.

[0140] 17, the coordinate system of the test object 40 is represented by (Xt, Yt, Zt), and the coordinate system of the light-receiving surface is represented by (X, Y, Z). Note that the surface of the test object 40 coincides with the surface where zt = 0, and the light-receiving surface 66 coincides with the surface where Z = 0.

[0141] The (Xt, Yt, Zt) coordinate system is defined as a coordinate system obtained by moving the (X, Y, Z) coordinate system by xt0, yt0, zt0 in the X direction, then rotating it by θx around the X axis, rotating it by θy around the Y axis after the rotation, and rotating it by θz around the Z axis after the rotation.

[0142] In the (Xt, Yt, Zt) coordinate system, the ends of each scale are taken as feature points, and the coordinates of each feature point are taken as (xt[1], yt[1], zt[1]), (xt[2], yt[2], zt[2]), ... (xt[n], yt[n], zt[n]). Note that the coordinates of the first feature point (xt[1], yt[1], zt[1]) are taken as the origin (0,0,0) of the (Xt, Yt, Zt) coordinate system. If the coordinates of feature point (xt[i], yt[i], zt[i]) in the (X, Y, Z) coordinate system are taken as (x[i], y[i], z[i]), then these coordinates are given by the following equation (1):

[0143]

[0144] however,

[0145]

[0146] The coordinates of each feature point of the test object 40 are converted into the (X, Y, Z) coordinate system using the above equation (1).

[0147] (B) The coordinates of each feature point are projected onto the light receiving surface 66. The line passing through the coordinates (xs, ys, zs) of the light source 64 and the feature points (x[i], y[i], z[i]) is given by the following equation, where t represents a parameter.

[0148]

[0149] The point t where the line in the above equation intersects with the light receiving surface 66, that is, the surface where Z=0, is zs / (zs-z[i]).

[0150] Therefore, when x and y of the light receiving surface 66 are respectively designated as xp and yp, they are given by the following equation (2).

[0151]

[0152] (C) The actual coordinates of each feature point on the light receiving surface 66 (the coordinates of each feature point detected on the image) are assumed to be (xpr[i], ypr[i], 0).

[0153] The coordinates (xpr[1], ypr[1], 0) of the first feature point are set as the origin (0, 0, 0) of the coordinate system (XYZ coordinate system) of the light receiving surface 66 .

[0154] The units of (xpr[i], ypr[i], 0) are the same as the units of (xp[i], yp[i], 0). For example, if the units of (xp[i], yp[i], 0) are mm, (xpr[i], ypr[i], 0) is multiplied by the pixel pitch (mm / pixel) to change the units to mm.

[0155] The sum of square errors between the coordinates of each feature point calculated using the distortion models (A) and (B) and the actual coordinates is defined as Er.

[0156] Er is given by the following equation (3).

[0157]

[0158] From the above equations (1), (2), and (3), the error Er is a function of a total of nine variables: the position (xt0, yt0, zt0) and rotation angle (θx, θy, θz) of the coordinate system of the test object 40, and the coordinates (xs, ys, zs) of the light source 64.

[0159] If the nine variables that minimize the error Er are found, the ratio between the size on the image and the size in the real world, that is, the actual size conversion information, can be obtained using the following formula.

[0160]

[0161] In the above formula, for example, the unit of the size on the image is pixels, the unit of the size in the real world is mm, the unit of the pixel pitch is mm / pixel, and the ratio of the size on the image to the size in the real world represents mm per pixel.

[0162] By expanding equations (1), (2), and (3), the error Er can be differentiated with respect to each of the nine variables, and therefore the values ​​of the variables that minimize the error Er can be found by gradient descent. In this case, annealing and / or stochastic gradient descent may be used to obtain a global solution without falling into a local solution. A genetic algorithm may also be used. Many methods have been proposed for finding a solution that minimizes a function, and any known method may be used.

[0163] Before carrying out the above correction, it is necessary to detect the feature points of each scale mark on the image, which can be detected using known corner detection methods such as Moravec and / or Harris.

[0164] In performing the above correction, it is desirable that the scales on the test object 40 are asymmetrical in order to identify the feature points of each scale on the image (to correspond to the feature points of each scale in the real world). For example, as shown in Fig. 17, if the widths of the scales are different, it is possible to identify the feature points.

[0165] Furthermore, before performing the above correction, it is preferable to first determine the tilt of the test object 40 in the real world from the distortion of the shape of the test object 40 in the image, and perform correction only if the test object 40 is tilted. The distortion of the test object 40 in the image can be evaluated based on the deviation of the shape of the test object 40 from a similar shape in the real world. For example, if the distance and / or angle between the scale marks on the test object 40 in the image deviates by more than a predetermined value compared to the distance and / or angle between the scale marks when the shape of the test object 40 in the real world is enlarged while maintaining its similar shape, the test object 40 in the image is distorted. In other words, it can be determined that the test object 40 is tilted in the real world.

[0166] When it is determined that the test object 40 is tilted, if at least the scale closest to the light source 64 is known in advance, the above-described correction may be omitted and actual size conversion information may be obtained simply based on multiple scales excluding the scale farthest from the light source 64. The scales closer to the light source 64 (scales where the angle of incidence of the light from the light source 64 on the scale is closer to perpendicular) are less affected by the tilt of the test object 40 than the scales farther from the light source 64, resulting in less distortion, and therefore accurate actual size conversion information may be obtained. Alternatively, after determining the nine variables above, actual size conversion information may be obtained by selecting scales excluding the scale farthest from the light source 64 based on the relationship between the coordinates of the light source 64 and the coordinates of each scale, without performing correction.

[0167] If it is determined that the test object 40 is not tilted, actual size conversion information can be obtained based on the scales on the test object 40 excluding the pair of scales that are closest in distance within the range shown on the image.

[0168] In the above embodiment, a radiological image is used as an example of an input image, but the input image is not limited to a radiological image, and other images may be used.

[0169] Furthermore, the various processes performed by the CPU in the above embodiments by executing software (programs) may be executed by a computer equipped with various processors other than a CPU. Examples of such processors include programmable logic devices (PLDs) (such as field-programmable gate arrays (FPGAs)) whose circuit configuration can be changed after manufacture, and dedicated electrical circuits, such as application-specific integrated circuits (ASICs), which are processors with circuit configurations specifically designed to execute specific processes. The various processes may be executed by one of these processors, or by a combination of two or more processors of the same or different types (e.g., multiple FPGAs, or a combination of a CPU and an FPGA). The hardware structure of these processors is, more specifically, an electrical circuit that combines circuit elements such as semiconductor devices.

[0170] In the above embodiment, the various programs are pre-stored (installed) in the ROM 20B, but the present invention is not limited to this. The various programs may be provided in a form recorded on a recording medium such as a CD-ROM (Compact Disk Read Only Memory), a DVD-ROM (Digital Versatile Disk Read Only Memory), or a USB (Universal Serial Bus) memory. The various programs may also be downloaded from an external information processing device or the like via a network.

[0171] The program of the present disclosure can be provided as a program product. The program product includes any product for providing the program. For example, the program product includes a program provided over a network such as the Internet, and a non-transitory computer-readable recording medium such as a CD-ROM or DVD on which the program is stored.

[0172] Furthermore, the configuration, operation, etc. of the radiation image capturing system 10 described in the above embodiment are merely examples, and it goes without saying that they can be modified according to the circumstances within the scope of the present disclosure.

[0173] The following supplementary note is further disclosed regarding the above embodiment: (Supplementary Note 1) An image processing device including a processor, the processor detecting a pair of scale marks from a partial image including a part of a test object in an input image having as its subject a target object and a test object having a plurality of scale marks whose positional relationship is known, and deriving actual size conversion information based on the size of the pair of scale marks on the image and the size of the pair of scale marks in the real world.

[0174] (Supplementary Note 2) The image processing device according to Supplementary Note 1, wherein the input image is a radiographic image for non-destructive testing, and the test object is at least one of a multi-linear image quality meter and a perforated penetrameter.

[0175] (Supplementary Note 3) The image processing device according to Supplementary Note 1 or Supplementary Note 2, wherein the processor detects, as the pair of scale marks, a pair of scale marks that is not closest to each other on the image among the scale marks included in the partial image.

[0176] (Supplementary Note 4) The image processing device according to any one of Supplementary Notes 1 to 3, wherein the processor derives the actual size conversion information of a specified position by interpolating the positions of the pair of scale marks and the actual size conversion information derived from each of the plurality of partial images.

[0177] (Supplementary Note 5) The image processing device according to any one of Supplementary Notes 1 to 4, wherein the processor derives the size on the image and the real-world size of an object whose real-world size is known and is larger than the size on the image of the pair of scales as actual size conversion information.

[0178] (Supplementary Note 6) The image processing device according to any one of Supplementary Notes 1 to 5, wherein the processor detects distortion of the test object and corrects positions of the pair of scales on the image based on the distortion.

[0179] (Supplementary Note 7) The image processing device according to Supplementary Note 6, wherein the processor, when detecting distortion of the test object, selects a pair of scales excluding the pair of scales whose distance on the image is the greatest.

[0180] (Supplementary Note 8) The image processing device according to Supplementary Note 6 or Supplementary Note 7, wherein the processor calculates the distortion correction information based on a change in a distance between the scale marks.

[0181] (Supplementary Note 9) The image processing device according to Supplementary Note 6 or Supplementary Note 7, wherein the processor defines the position of the scale on the image as a known variable, and defines the position of a light source relative to a light receiving surface in the real world, the position of the test object relative to the light receiving surface in the real world, and the rotation angle of the test object relative to the light receiving surface in the real world as unknown variables, and defines the unknown variables as the distortion correction information.

[0182] (Supplementary Note 10) The image processing device according to any one of Supplementary Notes 1 to 9, wherein the processor detects a portion of the test object from the input image.

[0183] (Supplementary Note 11) The image processing device according to any one of Supplementary Notes 1 to 10, wherein the processor calculates a real-world size of the target object from the actual size conversion information and a size of the target object on the image.

[0184] (Supplementary Note 12) The image processing device according to any one of Supplementary Notes 1 to 11, wherein the processor calculates a real-world size of the target object from the actual size conversion information and the size of the target object on the image, and superimposes the real-world size of the target object and the actual size conversion information on an input image.

[0185] (Supplementary Note 13) An image processing device comprising: a processor that detects a test object having a plurality of scales whose positional relationships are known from a partial image including a part of the test object in an input image including the test object; and calculates distortion correction information that corrects distortion of the test object from the positions of the detected scales on the image.

[0186] (Supplementary Note 14) An image processing method in which a computer performs a process to detect a pair of scale marks from a partial image including a part of a test object in an input image in which the test object has a target object and a plurality of scale marks whose positional relationship is known as subjects, and derive actual size conversion information based on the size of the pair of scale marks on the image and the real-world size of the pair of scale marks.

[0187] (Supplementary Note 15) An image processing program for causing a computer to execute a process of: detecting a pair of scale marks from a partial image including a part of a test object in an input image having a target object and a test object having a plurality of scale marks whose positional relationship is known as subjects; and deriving actual size conversion information based on the size of the pair of scale marks on the image and the real-world size of the pair of scale marks.

Claims

1. An image processing device comprising a processor, which detects a pair of scale marks from a partial image including a part of a test object in an input image of a target object and a test object having a plurality of scale marks whose positional relationship is known, and derives actual size conversion information based on the size of the pair of scale marks in the image and the real-world size of the pair of scale marks.

2. The image processing apparatus of claim 1, wherein the input image is a radiographic image for non-destructive testing, and the test object is at least one of a multi-linear image quality meter and a perforated penetrameter.

3. The image processing device according to claim 1, wherein the processor detects, as the pair of scale marks, a pair of scale marks that is not the pair of scale marks that is closest in distance on the image, from among the scale marks included in the partial image.

4. An image processing device according to claim 1, wherein the processor derives the actual size conversion information of a specified position by interpolating the positions of the pair of scale marks and the actual size conversion information derived from each of the plurality of partial images.

5. The image processing device according to claim 1, wherein the processor derives the size on the image and the real-world size of an object whose real-world size is known and which is larger than the size on the image of the pair of scales as real-world size conversion information.

6. The image processing device according to claim 1, wherein the processor detects distortion of the test object, and corrects the positions of the pair of scale marks on the image based on the distortion.

7. The image processing device according to claim 6, wherein said processor selects a pair of scales excluding the pair with the furthest distance on the image when it detects distortion of said test object.

8. The image processing device according to claim 6, wherein the processor calculates the distortion correction information based on changes in the distance between the scale marks.

9. The image processing device according to claim 6, wherein the processor defines the position of the scale on the image as a known variable, the position of the light source relative to the light-receiving surface in the real world, the position of the test object relative to the light-receiving surface in the real world, and the rotation angle of the test object relative to the light-receiving surface in the real world as unknown variables, and defines the unknown variables as the distortion correction information.

10. The image processing device of claim 1, wherein the processor detects a portion of the test object from the input image.

11. The image processing device according to claim 1, wherein the processor calculates the real-world size of the target object from the actual size conversion information and the size of the target object on the image.

12. The image processing device according to claim 1, wherein the processor calculates the real-world size of the target object from the actual size conversion information and the size of the target object on the image, and superimposes the real-world size of the target object and the actual size conversion information on the input image.

13. An image processing device comprising a processor that detects the scales from a partial image including a part of a test object in an input image including the test object having a plurality of scales whose positional relationships are known, and calculates distortion correction information that corrects distortion of the test object from the positions of the detected scales on the image.

14. An image processing method comprising: detecting a pair of scale marks from a partial image including a part of a test object in an input image of a target object and a test object having a plurality of scale marks whose positional relationship is known; and deriving actual size conversion information based on the size of the pair of scale marks in the image and the real-world size of the pair of scale marks.

15. An image processing program for causing a computer to execute processing including: detecting a pair of scale marks from a partial image including a part of a test object in an input image of a target object and a test object having a plurality of scale marks whose positional relationship is known; and deriving actual size conversion information based on the size of the pair of scale marks in the image and the real-world size of the pair of scale marks.

Images