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

Abstract

To more accurately identify the shape or distribution of a specific material inside a specimen.SOLUTION: An image acquisition unit acquires a stereoscopic image in which cross-sectional images of a specimen are stacked and which represents the structure of the specimen by brightness. A binarization processing unit binarizes the brightness of each pixel of the stereoscopic image, and generates a binary image indicating either brightness or darkness for each pixel. An index calculation unit calculates, for each pixel of the binary image, the ratio of bright portions in a nearby region within a predetermined stereoscopic range from the pixel as a bright portion distribution index. An analysis unit defines a characteristic region in which the bright portion distribution index has a value within a predetermined value range.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] Embodiments of the present invention relate to an image processing apparatus, an image processing system, an image processing method, and a program. [Background technology]

[0002] The main symptom of deterioration in concrete structures is cracking. Cracks are known to originate around reinforcing bars under tensile force and propagate along the transition zone between coarse aggregate and cement paste. This indicates that the distribution of coarse aggregate around reinforcing bars is important information for predicting the failure behavior of concrete structures. Therefore, visualizing the interior of concrete structures is highly anticipated.

[0003] X-ray CT (Computed Tomography) is a well-known method for non-destructively visualizing the interior of an object. X-ray CT not only allows for non-destructive visualization of the object's interior, but also enables quantitative analysis using numerical data such as CT values ​​(luminance values). Using X-ray CT, a two-dimensional cross-sectional image of the object can be obtained.

[0004] Patent Document 1 describes a non-destructive testing method in which a neutron beam is irradiated onto a test specimen, a transmission image is taken that is visualized as the beam passes through the specimen, and the moisture content of the specimen is evaluated from the captured transmission image. Patent Document 2 describes an evaluation method in which an X-ray CT method is applied to a cementitious hardened body to obtain an image showing the density distribution inside the cementitious hardened body, and the obtained image is analyzed to evaluate areas with relatively high density in the image as areas with a high probability of efflorescence occurring. [Prior art documents] [Patent Documents]

[0005] [Patent Document 1] Japanese Patent Publication No. 2019-174255 [Patent Document 2] Japanese Patent Publication No. 2021-165722 [Overview of the Initiative] [Problems that the invention aims to solve]

[0006] In cross-sectional images of concrete structures, the areas where coarse aggregate is present have relatively higher brightness than other areas. Therefore, it is conceivable to set a threshold for the cross-sectional image and separate the areas of coarse aggregate by binarizing the image using that value as a boundary. By sequentially superimposing cross-sectional images representing the areas of coarse aggregate, it is possible to observe those areas in three dimensions. However, simply binarizing the cross-sectional image may not sufficiently separate the areas of coarse aggregate from other areas. As a result, the shape and distribution of the coarse aggregate may not be accurately identified. [Means for solving the problem]

[0007] The image processing apparatus according to the first embodiment has the structure of the test specimen Obtained by radiation measurement An image acquisition unit acquires a stereoscopic image represented by brightness; a binarization processing unit binarizes the brightness of each pixel of the stereoscopic image and generates a binary image showing either light or dark for each pixel; an index calculation unit calculates the ratio of bright areas in the neighboring region within a predetermined three-dimensional range from the pixel of the binary image as a brightness distribution index; and the brightness distribution index is predetermined Above the threshold An analysis unit that defines the characteristic region, A parameter setting unit defines a pixel count distribution showing the number of pixels for each bright area distribution index in the stereoscopic image, and sets a bright area distribution index having a significant difference from the bright area distribution index at which the number of pixels is maximum in the pixel count distribution as the threshold. It is equipped with.

[0008] The image processing method according to the second embodiment is an image processing method in an image processing apparatus, The image processing device stacks cross-sectional images of the specimen and displays the structure of the specimen. Obtained by radiation measurement The process includes an image acquisition step to obtain a stereoscopic image represented by brightness, a binarization step to binarize the brightness of each pixel of the stereoscopic image and generate a binary image in which each pixel shows either light or dark, an index calculation step to calculate the ratio of bright areas in the neighboring region within a predetermined three-dimensional range from the pixel of the binary image as a bright area distribution index, and the bright area distribution index is predetermined Above the thresholdAn analysis step for determining a characteristic region, and A parameter setting step in which a pixel count distribution showing the number of pixels for each bright area distribution index in the stereoscopic image is defined, and a bright area distribution index having a significant difference from the bright area distribution index at which the number of pixels is maximum in the pixel count distribution is defined as the threshold, An image processing method for executing.

Effect of the Invention

[0009] According to the present embodiment, the shape or distribution of a specific material inside the specimen can be specified more accurately.

Brief Description of the Drawings

[0010] [Figure 1] It is an explanatory diagram for explaining the outline of the image processing system according to the present embodiment. [Figure 2] It is a schematic block diagram showing a functional configuration example of the image processing apparatus according to the present embodiment. [Figure 3] It is a diagram showing a first example of a cross-sectional image. [Figure 4] It is a diagram showing a second example of a cross-sectional image. [Figure 5] It is a diagram showing the luminance value distribution of a cross-sectional image. [Figure 6] It is a diagram showing a first example of a binary image. [Figure 7] It is a diagram showing an example of a binary image integrated in a three-dimensional space. [Figure 8] It is a plan view showing an example of a neighborhood area. [Figure 9] It is a perspective view showing an example of a neighborhood area. [Figure 10] It is a diagram showing a first example of a second image representing the two-dimensional distribution of ADI. [Figure 11] It is a diagram showing an example of the spatial distribution of ADI. [Figure 12] It is a flowchart showing an example of image processing according to the present embodiment. [Figure 13] It is a schematic block diagram showing a hardware configuration example of the image processing apparatus according to the present embodiment. [Figure 14] It is a cross-sectional view showing a configuration example of the specimen. [Figure 15] It is a table exemplifying a part of the shooting conditions. [Figure 16] It is a table exemplifying another part of the shooting conditions. [Figure 17] This figure shows a second example of the second image relating to the two-dimensional distribution of ADI. [Figure 18] This figure shows a third example of the second image relating to the two-dimensional distribution of ADI. [Figure 19] This figure illustrates a one-dimensional distribution of ADI. [Figure 20] This is a second example of a binary image. [Figure 21] This figure shows the first example of a feature image relating to a feature region. [Figure 22] This figure shows a second example of a feature image relating to a feature region. [Modes for carrying out the invention]

[0011] (System Overview) Embodiments of the present invention will be described below with reference to the drawings. Figure 1 is an explanatory diagram illustrating the outline of the image processing system 1 according to this embodiment. The image processing system 1 is a system for non-destructively visualizing the internal structure of an object using the X-ray CT method. In the example of Figure 1, the image processing system 1 consists of an image processing device 10, a detection unit 20, a display unit 30, an X-ray source 40, and a base 50. The detection unit 20 is positioned facing the X-ray source 40, with the test specimen SP in between. The test specimen SP is placed on the base 50 with its axial direction perpendicular to the horizontal plane, and is positioned and oriented such that the height of its center is approximately the same as the height of the X-ray source 40. In this invention, the case in which the test specimen SP has the shape of an elongated cylinder and its axial length is longer than the diameter of its base will be mainly described, but it is not limited to this. The shape of the test specimen SP does not necessarily have to be an elongated cylinder, and its diameter may be longer than its axial length. The shape of the test specimen SP may be, for example, a rectangular parallelepiped. In that case, the cross-sectional shape of the specimen SP will be square or rectangular. It is sufficient to determine the area occupied by the specimen SP in three-dimensional space from the information captured from multiple directions using the detection unit 20.

[0012] The base 50 is equipped with a turntable and a motor. The motor can rotate the turntable around its axis of rotation. The test specimen SP is placed on the turntable. The axis of rotation is oriented perpendicular to the horizontal plane. By positioning the test specimen SP so that its central axis coincides with the axis of rotation, the test specimen SP can be rotated around its central axis. The X-ray source 40 converts the power supplied to it into X-rays and emits the resulting X-rays. The X-ray source 40 includes, for example, an X-ray tube.

[0013] The detection unit 20 detects X-rays that have passed through the rotating test specimen SP. The detection unit 20 is, for example, an X-ray detector comprising a substrate and multiple detection devices. Multiple detection points are arranged in two dimensions at regular intervals on the main surface of the substrate. Each detection device is installed at a corresponding detection point and converts the intensity of the irradiated X-rays into an electrical signal having a voltage corresponding to that signal. The detection devices include, for example, photodiodes. The detection unit 20 outputs a detection signal, represented by the electrical signals converted by each detection device, to the image processing device 10. The output detection signal indicates the intensity distribution of X-rays irradiated onto the substrate of the detection unit 20.

[0014] The image processing device 10 acquires a cross-sectional image showing the structure within the cross-section of the specimen SP using a known X-ray CT method based on the detection signal input from the detection unit 20. The normal direction of the cross-section corresponds to the axial direction of the specimen. The image processing device 10 stacks the cross-sectional images acquired for each cross-section in the axial direction to acquire a stereoscopic image representing the three-dimensional structure of the specimen in terms of brightness. The image processing device 10 binarizes the brightness of each pixel in the stereoscopic image and generates a binary image showing either light or dark for each pixel. For each pixel in the binary image, the image processing device 10 calculates the ratio of bright areas in the neighboring region within a predetermined three-dimensional range from that pixel as the brightness distribution index. The brightness distribution index will be described later.

[0015] The image processing device 10 defines a feature region in which the brightness distribution index is above a predetermined threshold. The image processing device 10 generates a display image in which the brightness of the defined feature region is represented with a higher brightness than other regions, and displays the generated display image on the display unit 30. The display image represents the regions occupied by materials with different densities using differences in brightness. Regions with higher brightness represent regions of materials with higher density. The specimen SP is, for example, a concrete column. Concrete is formed by mixing cement paste and aggregate. The cement paste acts as a binder that connects the individual aggregates. The aggregate is the material that forms the skeleton of the concrete. In this application, the case of coarse aggregate with a particle size of 5 mm or more is mainly illustrated. Natural gravel, artificial crushed stone, etc., can be used as coarse aggregate.

[0016] (Image processing device) Next, an example of the functional configuration of the image processing apparatus 10 according to this embodiment will be described. Figure 2 is a schematic block diagram showing an example of the functional configuration of the image processing apparatus 10 according to this embodiment. The image processing device 10 comprises a control unit 110, a storage unit 130, and an input / output unit 140. The control unit 110 performs processing to enable the image processing device 10 to function and controls that processing. The control unit 110 includes an image acquisition unit 112, a parameter setting unit 114, a binarization processing unit 116, an index calculation unit 118, an analysis unit 120, and a display processing unit 122.

[0017] The image acquisition unit 112 acquires a cross-sectional image representing the structure of the cross-section of the specimen SP based on the detection signal input from the detection unit 20 using a known X-ray CT method. The cross-sectional image is also called a tomographic image. Generally, the higher the density of the transmitted area, or the longer the distance transmitted through the material, the lower the intensity of the X-rays detected by the detection unit 20. The detection signal contains information on the intensity of X-rays irradiated from all directions, depending on the rotation of the specimen SP at each detection point. The image acquisition unit 112 can use a known method to compensate for the effect of distance attenuation for intensities with different irradiation directions and estimate the relative intensity of X-rays at each of a plurality of pixels arranged within the cross-section of the specimen SP at a certain height. The plurality of pixels are arranged two-dimensionally at regular intervals within the cross-section. Each individual pixel corresponds to one pixel. The image acquisition unit 112 can generate a tomographic image in which the relative intensity estimated for each pixel is scaled to obtain a value that corresponds to the brightness value of the pixel. In scaling, the brightness value is converted to a value that falls within a predetermined range based on the relative intensity. For example, if the bit depth for each pixel is 8 bits, the brightness value will take any integer value in the range of 0 to 255.

[0018] The image acquisition unit 112 generates cross-sectional images for each cross-section with different heights, and stacks the generated cross-sectional images sequentially along the height direction to obtain a three-dimensional image showing the three-dimensional structure inside the specimen SP. The height direction corresponds to the normal direction of each cross-section. The image acquisition unit 112 stores the acquired three-dimensional image data in the storage unit 130. The first three-dimensional image data is configured as volume data showing the brightness value of each pixel distributed in three-dimensional space.

[0019] The parameter setting unit 114 sets various parameters used in the processing of the image processing device 10. The parameter setting unit 114 displays a parameter setting screen on the display unit 30 in response to parameter setting instructions indicated by operation signals input from, for example, an operation input unit (not shown). The operation input unit includes a device that receives user operations, generates operation signals according to the received operations, and outputs the generated operation signals. The operation input unit may include, for example, a pointing device such as a mouse, keyboard, or touch sensor. The parameter setting unit 114 may determine the values ​​of individual parameters according to the operation signals input from the operation input unit. The parameters to be set may include a threshold used for binarization of luminance (hereinafter referred to as the "first threshold") and a threshold used for binarization of the highlight distribution index (hereinafter referred to as the "second threshold").

[0020] The parameter setting unit 114 can determine a first threshold using one or more cross-sectional images shown in the first stereoscopic image data using the following method. For each luminance value within a predetermined range, the parameter setting unit 114 counts the number of pixels representing that luminance value (hereinafter referred to as "pixel count") in part or all of the tomographic image to be analyzed as a frequency, and generates a histogram showing the pixel count for each luminance value. The generated histogram shows the luminance distribution of the entire area to be analyzed. The parameter setting unit 114 determines the luminance value with the maximum frequency in the generated histogram as the first feature value. The position of the pixel that gives the first feature value as a luminance value (hereinafter referred to as "pixel position") in the tomographic image is estimated as a representative position where the principal components constituting the specimen SP exist. The parameter setting unit 114 can determine a luminance value significantly smaller than the first feature value as the first threshold.

[0021] The histogram may have a characteristic frequency distribution for each component of the specimen SP. Therefore, the parameter setting unit 114 may define a second feature value as a value that has a significant difference from the first feature value among the frequencies that give rise to the feature points of the histogram. Examples of feature points include the maximum luminance point and the maximum point of the derivative of luminance. In the tomographic image, the pixel position of the pixel that gives the second feature value as a luminance value is estimated as a representative position where the second principal component of the specimen SP exists. The parameter setting unit 114 can define a value between the first feature value and the second feature value as the first threshold.

[0022] The binarization processing unit 116 generates binary image data by binarizing the brightness values ​​of each pixel shown in the first three-dimensional image data using a preset first threshold. The binary image data has a binary value (hereinafter referred to as the "first binary value") that indicates brightness or darkness for each pixel. The binarization processing unit 116 determines for each pixel whether the brightness value is equal to or greater than the first threshold, sets the first binary value of pixels whose brightness value is equal to or greater than the first threshold to 1, and sets the first binary value of pixels whose brightness value is less than the first threshold to 0. Brightness or darkness is represented by whether the first binary value of each pixel is 1 or 0. In three-dimensional space, the area where pixels with a first binary value of 1 are distributed corresponds to the bright area, and the area where pixels with a first binary value of 0 are distributed corresponds to the dark area. The binarization processing unit 116 stores the generated binary image data in the storage unit 130.

[0023] The index calculation unit 118 reads binary image data from the storage unit 130, and, referring to the read binary image data, calculates the ratio of bright areas in a predetermined three-dimensional area, the neighboring region, for each pixel as the bright area distribution index. That is, the neighboring region is a three-dimensional area with the position of the pixel as the reference point. The index calculation unit 118 defines the data showing the bright area distribution index calculated for each pixel as the second three-dimensional image data. The second three-dimensional image data is volume data showing the spatial distribution of the bright area distribution index in the same spatial area as the first three-dimensional image data. The index calculation unit 118 stores the generated second three-dimensional image data in the storage unit 130.

[0024] The analysis unit 120 reads the second stereoscopic image data from the storage unit 130 and determines feature regions where the highlight distribution index is equal to or greater than the second threshold based on the read second stereoscopic image data. The analysis unit 120 may use a pre-set second threshold to binarize the highlight distribution index for each pixel shown in the second stereoscopic image data and determine a binary value (hereinafter referred to as the "second binary value") for each pixel indicating whether or not it belongs to a feature region. Feature region data is generated as volume data indicating the second binary value for each pixel. The analysis unit 120 stores the generated feature region data in the storage unit 130.

[0025] The display processing unit 122 acquires display data for displaying various display information on the display unit 30, and outputs the acquired display data to the display unit 30 via the input / output unit 140. The display processing unit 122 identifies the display information indicated by the operation signal input from the operation input unit, for example. The operation signal indicates, for example, one of the following: a setting screen, a first stereoscopic image, a binary image, a second stereoscopic image, or a feature region image. For the first stereoscopic image, binary image, second stereoscopic image, and feature region image, the operation signal may also indicate the height of the cross-section to be displayed. When the settings screen is requested, the settings screen data stored in advance is read from the storage unit 130, and the read settings screen data is output to the display unit 30 as display data. The display unit 30 displays the settings screen based on the display data input from the image processing device 10.

[0026] When a first stereoscopic image is requested, the first stereoscopic image data, which has been stored in advance, is read from the storage unit 130, and display data indicating the brightness value of each pixel located in the cross section corresponding to the requested height is output to the display unit 30 from the read first stereoscopic image data. The display unit 30, which is the output destination, displays a cross-sectional image of the cross section of the test specimen SP at the requested height. When a binary image is requested, the storage unit 130 reads pre-stored binary image data and outputs display data to the display unit 30, which shows the first binary value for each pixel located in the cross section corresponding to the requested height from the read binary image data. The display unit 30, which is the output destination, displays a cross-sectional image representing the distribution of light and dark areas in the cross section of the test specimen SP at the requested height.

[0027] When a second stereoscopic image is requested, the second stereoscopic image data, which has been stored in advance, is read from the storage unit 130, and display data showing the brightness distribution index for each pixel located in the cross section corresponding to the requested height is output to the display unit 30 from the read second stereoscopic image data. The display unit 30, which is the output destination, displays a cross-sectional image representing the distribution of the brightness distribution index of the cross section at the requested height of the test specimen SP.

[0028] When a feature region image is requested, the memory unit 130 reads pre-stored feature region data and outputs display data to the display unit 30 that shows the second binary value for each pixel located in the cross section corresponding to the requested height, based on the read feature region data. The display unit 30, which is the output destination, displays a cross-sectional image representing the distribution of feature regions in the cross section of the test specimen SP at the requested height.

[0029] The display processing unit 122 may perform processing such as enlarging, reducing, translating, and rotating the display information in accordance with the operation signals input to it, and output display data showing the processed display information to the display unit 30.

[0030] The storage unit 130 temporarily or permanently stores various data (including parameters) used by the control unit 110, and various data acquired by the control unit 110. The storage unit 130 includes the main storage medium and the auxiliary storage medium described above. The input / output unit 140 outputs input data received from another external device that serves as the input source to the control unit 110, and outputs output data from the control unit 110 to a device that serves as the output destination. The input / output unit 140 connects to other devices via wired or wireless connections to enable input and output of various types of data. The input / output unit 140 may also be connected to other devices via a communication network. The input / output unit may include, for example, one or both of an input / output interface and a communication interface.

[0031] The detection unit 20 may also include an image acquisition unit that has the function of generating a cross-sectional image showing the structure of the cross-section of the test specimen SP from the acquired detection signal. In that case, the image acquisition unit 112 of the image processing device 10 has the function of acquiring cross-sectional image data showing the cross-sectional image from the detection unit 20, and the function of generating a cross-sectional image from the detection signal may be omitted.

[0032] (Cross-sectional image) Next, an example of an acquired cross-sectional image will be described. Figure 3 shows the first example of a cross-sectional image. The illustrated cross-sectional image shows a cross-section of a concrete column, referred to as specimen SP. Specimen SP has an elongated cylindrical shape, as illustrated in Figure 14, with a high-tensile threaded steel bar Sb embedded in the longitudinal direction along its central axis. The shape and size of specimen SP are determined using the diameter R2 and length SL of the concrete column, the diameter R1 and length BL of the high-tensile threaded steel bar Sb, and the embedment depth EL.

[0033] Generally, in images acquired using X-ray CT, areas with higher density materials are represented by higher brightness. In the example in Figure 3, the bright circular area in the center represents the area occupied by high-tensile threaded steel bars Sb (hereinafter referred to as the "steel bar area"). The darker circular area surrounding the center represents the area occupied by concrete (hereinafter referred to as the "concrete area"). The darker area surrounding the outermost part represents the area where the specimen SP does not exist. To observe the concrete composition in more detail, the binarization processing unit 116 may exclude the steel bar area from the cross-sectional image and leave the concrete area as the area to be analyzed. For example, the binarization processing unit 116 can identify the area where pixels with a brightness value equal to or greater than a predetermined reference brightness (e.g., the upper limit of the brightness value) are located as the steel bar area. Figure 4 is an example of the steel bar area identified from the cross-sectional image in Figure 3. In the examples in Figures 3 and 4, the bit depth of the brightness value is 8 bits. That is, the range of brightness values ​​for each pixel is from 0 to 255.

[0034] (histogram) Next, an example of a histogram obtained from a cross-sectional image will be described. A histogram can be viewed as a luminance distribution, showing the number of pixels for each luminance value in the cross-sectional image, or the ratio of such values ​​as frequency. The parameter setting unit 114 can count the number of pixels for each luminance value in the concrete region to be analyzed within the cross-sectional image and generate a histogram.

[0035] Figure 5 illustrates a histogram relating to the concrete region identified from the cross-sectional image in Figure 3. In Figure 5, the vertical axis and horizontal axis represent frequency and luminance value, respectively. The parameter setting unit 114 can define the luminance value that gives the histogram's maximum point as the first feature value. This first feature value gives the maximum frequency. Pixels with luminance values ​​within a predetermined range from the first feature value can be estimated to represent pixels of coarse aggregate forming the first principal component. In other words, luminance values ​​included within a predetermined range from the first feature value represent the luminance distribution of coarse aggregate as the first principal component.

[0036] The derivative of the histogram exemplified in Figure 5 has a maximum value and is significantly smaller than the first feature value. The parameter setting unit 114 can set this maximum value of the derivative as the second feature value. Pixels with a luminance value within a predetermined range from the second feature value can be estimated to represent the concrete paste that constitutes the second principal component of the concrete. In other words, the luminance values ​​included within a predetermined range from the second feature value are estimated as the luminance distribution of the concrete paste as the second principal component. In the exemplified histogram, an inflection point exists between the first and second feature values. The parameter setting unit 114 can set this inflection point as the first threshold. Furthermore, if multiple frequency maxima are detected, the parameter setting unit 114 may determine the luminance value that gives the second highest frequency maxima among the multiple maxima as the second feature value.

[0037] (brightness distribution index) However, using only luminance values ​​can make it difficult to distinguish between the first region representing the first principal component and other regions, or to determine the first threshold for that distinction. In the binary image exemplified in Figure 6, although the detection of coarse aggregate with relatively large particle sizes is expected in the first region, fine bright areas appear on the surface of the particles constituting the coarse aggregate. Furthermore, although it is expected that multiple bright areas representing adjacent coarse aggregates in the first region will be detected spatially separated, they were sometimes detected as linked. Possible reasons for this include the relatively small density difference between different types of materials, resulting in insufficient difference between the first threshold and the first feature value, and the luminance of the first principal component and other components being similar. In addition, when known filtering processes are applied to the luminance values ​​and the processed first image is binarized, the shape of the bright areas sometimes changes significantly. As a result, it sometimes becomes impossible to accurately identify the shape and distribution of coarse aggregate appearing in the bright areas.

[0038] Therefore, the index calculation unit 118 according to this embodiment determines a brightness distribution index for each pixel based on binary image data. When observing the region where aggregate is distributed from the specimen SP, the index calculation unit 118 can calculate an aggregate distribution index (ADI) as an example of a brightness distribution index using the following method. The index calculation unit 118 identifies a cubic region centered on the pixel of interest to be processed as the neighboring region R in the three-dimensional space represented by the binary image (see Figure 7). In the examples of Figures 8 and 9, the length of one side of the neighboring region R is 2n+1 points. n represents a predetermined natural number of 1 or more.

[0039] The index calculation unit 118 refers to the binary image data and determines the (2n+1) values ​​belonging to the neighboring region. 3 From among the individual pixels, the pixels whose first binary value indicates a bright area are identified, and the number of identified pixels is counted. The brightness of the part representing the aggregate, and the number of counted pixels, indicate the size of the bright area in the neighboring region. The index calculation unit 118 calculates, for example, the number of bright pixels in the neighboring region as shown in equation (1), which is the total number of pixels (in the example of Figures 8 and 9, (2n+1)). 3 The ADI can be calculated by dividing by ).

[0040]

number

[0041] ADI is an index that indicates the degree to which a pixel of interest belongs to the aggregate region. This is based on the fact that individual particles making up aggregate are made of a certain material and occupy a spatially continuous region. On the other hand, when observing the structure of concrete as a specimen SP, the measurement system is set so that the pixel spacing (pixel pitch) is sufficiently smaller than the aggregate particles contained in the concrete, especially the coarse aggregate particles. If a pixel of interest belongs to a region representing the interior of the aggregate, there is a possibility that pixels in the vicinity of the pixel of interest also belong to the interior of the aggregate. If the first threshold is set so that the aggregate region is represented by a bright area, the ratio of the number of bright pixels in the neighboring region indicates the degree to which the pixel of interest also belongs to the aggregate region represented by a bright area. Since regions that are misdetected as different bright areas in each cross-sectional image have low correlation with each other, the influence of misdetection can be reduced by using neighboring regions that span multiple cross-sections with different heights.

[0042] The above n is a parameter that indicates the size of the neighboring region. The parameter setting unit 114 may increase n as the resolution of the cross-sectional image increases. More specifically, the parameter setting unit 114 may set n so that it is inversely proportional to the pixel spacing. The parameter setting unit 114 may set n such that the product of n and the pixel spacing is smaller than the dimensions of at least the main component members. As the dimensions of the main component members, for example, the particle size of individual particles that make up the coarse aggregate can be applied. In that case, the parameter setting unit 114 may identify the feature regions described later as spatially continuous regions, each representing an individual particle, and set the diameter of the identified particle region as the particle size. The parameter setting unit 114 may set the average value, mode, or median of the diameters of each of the multiple particle regions identified from the second image (described later) as the particle size.

[0043] The index calculation unit 118 generates a second stereoscopic image data showing the ADI calculated for each pixel, and stores the generated second stereoscopic image data in the storage unit 130. The analysis unit 120 refers to the second stereoscopic image data and determines whether the ADI for each pixel is equal to or greater than a second threshold, and can determine that pixels whose ADI is equal to or greater than the second threshold belong to a feature region.

[0044] Figure 11 shows an example of the spatial distribution of ADI as an example of a brightness distribution index. Figure 11 shows the ADI calculated for each pixel on the A-B line in the binary image shown in Figure 10. The ADI initially increases as it approaches point B (right end) from point A (left end), reaches a maximum value (peak), and then decreases. At the position where the ADI peaks, the ADI of the coarse aggregate that constitutes the first principal component reaches a minimum value, then increases again, reaches a new maximum value, and then decreases again. In Figure 11, the horizontal line indicates the second threshold. As the second threshold, the ADI at the outer edge of the region representing the coarse aggregate that constitutes the first principal component may be used.

[0045] The parameter setting unit 114 may set a second threshold according to an operation signal while displaying a first image representing the cross-section of the specimen SP and a second image showing the distribution of ADI on the display unit 30. This allows the user to visually inspect the first and second images to understand the distribution of coarse aggregate, which is the first principal component, and set the second threshold. The parameter setting unit 114 may also determine the second threshold based on the second stereoscopic image data using a method similar to the method for determining the first threshold based on the first stereoscopic image data.

[0046] Next, an example of image processing according to this embodiment will be described. Figure 12 is a flowchart showing an example of image processing according to this embodiment. Figure 12 shows an example where the first principal component is coarse aggregate and the brightness distribution index is ADI. (Step S102) The image acquisition unit 112 acquires a cross-sectional image representing the cross-sectional structure of the specimen SP based on the detection signal from the detection unit 20. The image acquisition unit 112 stacks the acquired cross-sectional images in the height direction to construct a three-dimensional image. (Step S104) The parameter setting unit 114 generates a histogram representing the brightness distribution from one of the cross-sectional images. The parameter setting unit 114 determines a first threshold based on a first feature value that gives the maximum value of the generated histogram. (Step S106) The binarization processing unit 116 binarizes the stereoscopic image using a first threshold and generates a binary image that indicates whether each pixel is a bright area or not.

[0047] (Step S108) The index calculation unit 118 calculates the ratio of bright areas within the peripheral region set for each pixel as ADI and generates a second stereoscopic image showing the ADI for each pixel. (Step S110) The parameter setting unit 114 generates a histogram representing the distribution of ADI in any cross-section. The parameter setting unit 114 determines a second threshold based on a first feature value that gives the maximum value of the generated histogram.

[0048] (Step S112) The analysis unit 120 estimates the feature region where the ADI is greater than or equal to the second threshold from the second stereoscopic image as the region of coarse aggregate. (Step S114) The display processing unit 122 outputs display data representing the estimated coarse aggregate area to the display unit 30 and displays a display image representing the coarse aggregate area. After that, the process shown in Figure 12 is terminated.

[0049] The parameter setting unit 114 may pre-set one or both of the first and second thresholds offline before executing the process shown in Figure 12. In that case, one or both of steps S104 and S110 may be omitted in the process shown in Figure 12.

[0050] (Hardware configuration) Next, an example of the hardware configuration of the image processing apparatus 10 according to this embodiment will be described. The image processing apparatus 10 may be configured to include dedicated components (for example, integrated circuits) that form sets of one or more functional parts as shown in Figure 2, but it may also be configured as a computer system in part or in whole.

[0051] Figure 13 is a schematic block diagram showing an example of the hardware configuration of the image processing apparatus 10 according to this embodiment. Each part of the image processing apparatus 10 may be configured using dedicated components, or it may include a general-purpose computer system. The image processing device 10 is configured to include, for example, a processor 152, a drive unit 156, an input unit 158, an output unit 160, a ROM (Read Only Memory) 162, a RAM (Random Access Memory) 164, an auxiliary storage unit 166, and an interface unit 168. The processor 152, drive unit 156, input unit 158, output unit 160, ROM 162, RAM 164, auxiliary storage unit 166, and interface unit 168 are interconnected using a bus BS (baseline).

[0052] The processor 152, for example, reads programs and various data stored in the ROM 162, executes the programs, and controls the operation of the image processing device 10. The processor 152 may include, for example, one or more CPUs (Central Processing Units). The processor 152 may also include one or more GPUs (Graphic Processing Units) or coprocessors.

[0053] The processor 152 may execute a predetermined program to implement all or some of the functions of the above-mentioned functional units, for example, some or all of the functions of the image acquisition unit 112, parameter setting unit 114, binarization processing unit 116, index calculation unit 118, analysis unit 120, and display processing unit 122 of the control unit 110. The processor 152 may also work in cooperation with any or any combination of the ROM 162, RAM 164, and auxiliary storage unit 166 to implement the functions of the storage unit 130. The processor 152 may also work in cooperation with any or any combination of the input unit 158, output unit 160, and interface unit 168 to implement the functions of the input / output unit 140. Note that "executing a program" means executing the processing instructed by various commands written in the program.

[0054] The storage medium 154 stores various types of data. The storage medium 154 is a portable storage medium such as a magneto-optical disk, a flexible disk, or flash memory. The drive unit 156 is a device that performs, for example, either reading various data from the storage medium 154 or writing various data to the storage medium 154, or both.

[0055] The input unit 158 ​​receives input data from various devices that serve as input sources and outputs the input data to the processor 152. The output unit 160 outputs the output data input from the processor 152 to various devices that serve as output destinations.

[0056] ROM162 stores, for example, a program for the processor 152 to execute. RAM164 is used, for example, as a main memory medium that functions as a working area for temporarily storing various data and programs used by processor 152. The auxiliary storage unit 166 is a storage medium such as an HDD (Hard Disk Drive) or flash memory.

[0057] The interface unit 168 connects to other devices and allows for the input and output of various types of data. The interface unit 168 includes, for example, a communication module that connects to a network via wired or wireless connection. The input unit 158, output unit 160, and interface unit 168 correspond to the input / output unit 140 described above. The input unit 158 ​​can receive input data such as operation signals from the operation input unit and data from other devices. The output unit 160 can output display data to the display unit 30 as output data.

[0058] (Example of experiment) Next, an experimental example according to this embodiment will be described. As the specimen SP, the concrete column shown in FIG. 14 was used. The diameter R2 and length SL of the concrete column are 56.7 mm and 250 mm, respectively. The diameter R1, length BL, and embedment depth EL of the high-tensile threaded bar steel Sb are 13 mm, 260 mm, and 183 mm, respectively. As materials for the concrete, water, cement, fine aggregate, and coarse aggregate were used. Ordinary Portland cement, river sand, and limestone crushed stone were used as the cement, fine aggregate, and coarse aggregate, respectively. The densities of the cement, fine aggregate, and coarse aggregate are 3.15 g / cm 3 , 2.71 g / cm 3 , 2.68 g / cm 3 , respectively (see FIG. 15). Also, the maximum size of the coarse aggregate was set to 15 mm (see FIG. 16). The weight ratios of water, cement, fine aggregate, and coarse aggregate were 175 kg / m 3 , 350 kg / m 3 , 720 kg / m 3 , 1072 kg / m 3 , respectively. Also, the air content was set to 5%. Therefore, the water-cement ratio W / C and the fine aggregate ratio were 50% and 0.4%, respectively. The slump value was 8 cm.

[0059] An X-ray tube was used as the X-ray source 40. The tube voltage and tube current applied to the X-ray tube were 200 (kV) and 100 (μA), respectively. The image size of a cross-sectional image of one frame was 1024×1024 pixels. One pixel was regarded as one pixel. The pixel pitch was 0.057 mm. The bit depth was 8 bits. Therefore, the value range of the luminance value of each pixel is 0 or more and 255 or less. The number of photographed images was 3680. The number of photographed images corresponds to the number of frames of the cross-sectional images constituting one three-dimensional image.

[0060] Next, Figures 17 and 18 show examples of second images obtained from cross-sectional images acquired under the above conditions. Figures 17 and 18 show second images generated with n=1 and n=5, respectively. Figure 19 shows the distribution of ADI along the horizontal line in Figures 17 and 18. Figure 19 confirms that in the case of n=1, the ADI changes abruptly at the surface of the particles, clearly separating the regions of coarse aggregate and concrete paste. In the case of n=5, the peak value of ADI is located at the center of the particles, and the ADI fluctuates relatively smoothly. This confirms that the surface of individual particles making up the coarse aggregate is smoothed.

[0061] Next, examples of binary images and feature images obtained from cross-sectional images are shown. Figure 20 shows a second example of a binary image. However, the first threshold for the cross-sectional image was set to 198. Figures 21 and 22 show examples of feature images that show the distribution of coarse aggregate estimated as feature regions, respectively. The examples in Figures 21 and 22 are based on the second images with n=1 and n=5. Here, the second threshold for ADI was set to 0.5. In the cross-sectional image exemplified in Figure 20, the individual particles making up the coarse aggregate and the concrete paste are not clearly separated, and there are significant fine irregularities on the particle surface. In the feature image exemplified in Figure 21, the individual particles making up the coarse aggregate are separated from the concrete paste, and the irregularities on the particle surface are mitigated by smoothing. In the feature image exemplified in Figure 22, the irregularities on the particle surface are further smoothed.

[0062] In the above explanation, the image processing device 10 is described as being configured with the detection unit 20 and the display unit 30 as separate devices, but it is not limited to this. The image processing device 10 may be configured as a single device that integrates one or both of the detection unit 20 and the display unit 30. Furthermore, the display forming the display unit 30 may be integrated with a touch sensor, which is an example of an operation input unit, and configured as a touch panel.

[0063] The shape of the neighboring region is not limited to a cube centered on the pixel of interest; it may be a rectangular prism, a sphere, or other different shapes. If the shape of the neighboring region is a rectangular prism, its size can be indicated using the lengths of three mutually orthogonal sides in three-dimensional space. If the shape of the neighboring region is a sphere, its size can be indicated using its radius or diameter.

[0064] As described above, the image processing apparatus 10 according to the above embodiment includes: an image acquisition unit 112 that stacks cross-sectional images of a test specimen to acquire a three-dimensional image in which the structure of the test specimen is represented by brightness; a binarization processing unit 116 that binarizes the brightness of each pixel of the three-dimensional image and generates a binary image in which each pixel shows either light or dark; an index calculation unit 118 that calculates the ratio of bright areas in a neighboring region within a predetermined three-dimensional range from the pixel to the pixel of the binary image as a brightness distribution index; and an analysis unit 120 that defines a feature region in which the brightness distribution index is a value within a predetermined range (for example, above the second threshold). According to this configuration, the ratio of bright areas in a neighboring region is used as the bright area distribution index, which quantifies the likelihood that a particular material is represented as a principal component in the bright areas and occupies a continuum. A characteristic region where the value of the bright area distribution index falls within a predetermined range is clearly separated from other regions as the region where that material is distributed. As a result, the shape, distribution, and size of a specific material within the specimen are more accurately identified. In terms of shape, for example, the shape of a spatially continuous series of continuum or individual particles is determined. In terms of distribution, for example, the region occupied by the continuum or the position or arrangement of individual particles is determined. In terms of size, for example, the volume, length, diameter, etc., of the continuum or individual particles are determined.

[0065] Furthermore, the image processing device 10 may include a parameter setting unit 114 that defines the three-dimensional range of the nearby region based on the dimensions (e.g., particle size) of the material represented in the bright areas. This configuration determines the three-dimensional range of the neighboring region to be referenced for calculating the brightness distribution index, according to the dimensions of the material. In calculating the brightness distribution index, pixels within the neighboring region corresponding to the dimensions of the material are referenced, allowing for clear differentiation from other surrounding areas while avoiding deformation of the detected material's shape.

[0066] Furthermore, the parameter setting unit 114 may determine the three-dimensional range of the neighboring region based on the resolution of the stereoscopic image. With this configuration, the three-dimensional range of the neighboring region to be referenced for calculating the brightness distribution index is determined according to the resolution of the stereoscopic image. In calculating the brightness distribution index, by referencing pixels within the neighboring region corresponding to the resolution, it is possible to clearly distinguish it from other surrounding areas without causing an effective decrease in resolution.

[0067] Furthermore, the image processing device 10 may include a parameter setting unit 114 that defines a pixel count distribution (e.g., a histogram) showing the number of pixels for each bright area distribution index in a stereoscopic image, and defines a bright area distribution index that has a significant difference from the bright area distribution index at which the number of pixels is maximum in the pixel count distribution as a threshold value (e.g., a second threshold value) indicating a predetermined range. According to this configuration, the range of values ​​for defining the feature region is determined based on the range in which the highlight distribution index related to the principal components of the specimen, represented in the highlight areas of the binary image, is mainly distributed. Therefore, the feature region is clearly separated from other components.

[0068] Furthermore, the acquired cross-sectional images may be tomographic images that indicate a material with higher density, where higher brightness indicates higher density. With this configuration, the internal structure of the test specimen is accurately detected based on a cross-sectional image, which is given as a distribution of materials with densities corresponding to luminance, according to the luminance distribution.

[0069] Furthermore, the specimen may be concrete, the light distribution index may be the aggregate distribution index, and the analysis unit 120 may identify the characteristic region as the aggregate region where aggregate is represented. With this configuration, in concrete, the aggregate region, which represents the main component of the aggregate, can be clearly separated from other materials, such as concrete paste, and its distribution can be detected.

[0070] Although this embodiment has been described in detail with reference to the drawings, the specific configuration is not limited to the configurations described above, and includes designs and the like that do not depart from the gist of this embodiment. The above configurations can be combined arbitrarily, and some of them may be omitted. [Explanation of symbols]

[0071] 1…Image processing system, 10…Image processing device, 20…Detection unit, 30…Display unit, 110…Control unit, 112…Image acquisition unit, 114…Parameter setting unit, 116…Binarization processing unit, 118…Exponential calculation unit, 120…Analysis unit, 122…Display processing unit, 130…Storage unit, 140…Input / Output unit, 152…Processor, 156…Drive unit, 158…Input unit, 160…Output unit, 162…ROM, 164…RAM, 166…Auxiliary storage unit, 168…Interface unit

Claims

1. An image acquisition unit stacks cross-sectional images of a test specimen and acquires a three-dimensional image in which the structure of the test specimen is represented by brightness obtained by radiation measurement, A binarization processing unit that binarizes the brightness of each pixel in the aforementioned stereoscopic image and generates a binary image in which each pixel shows either light or dark, For each pixel of the binary image, an index calculation unit calculates the ratio of bright areas in the neighboring region within a predetermined three-dimensional range from that pixel as a bright area distribution index. An analysis unit that defines a feature region in which the brightness distribution index is equal to or greater than a predetermined threshold, A pixel count distribution is defined that shows the number of pixels for each of the bright area distribution indices in the stereoscopic image. The parameter setting unit includes a parameter setting unit that sets a threshold value for a bright area distribution index that has a significant difference from the bright area distribution index at which the number of pixels is maximized in the aforementioned pixel distribution. Image processing device.

2. The system includes a parameter setting unit that defines the three-dimensional range of the vicinity region based on the dimensions of the material of the test specimen shown in the bright area. The image processing apparatus according to claim 1.

3. The parameter setting unit is, Based on the resolution of the aforementioned stereoscopic image, the three-dimensional extent of the neighboring region is determined. The image processing apparatus according to claim 2.

4. The aforementioned cross-sectional image is a radiographic tomography image, The higher the brightness, the denser the material of the test specimen. The image processing apparatus according to claim 1.

5. The aforementioned test specimen is made of concrete. The aforementioned light distribution index is the aggregate distribution index. The analysis unit identifies the characteristic region as an aggregate region in which aggregate is represented. The image processing apparatus according to claim 4.

6. The image processing apparatus according to claim 4, The system includes a detector for detecting radiation representing the cross-sectional image. Image processing system.

7. On the computer, A program for causing the image processing device to function as described in claim 1.

8. An image processing method in an image processing device, The aforementioned image processing device The image acquisition step involves stacking cross-sectional images of the specimen to obtain a three-dimensional image in which the structure of the specimen is represented by brightness obtained by radiation measurement, A binarization step involves binarizing the brightness of each pixel in the aforementioned stereoscopic image to generate a binary image in which each pixel shows either light or dark. For each pixel of the binary image, an index calculation step is performed to calculate the ratio of bright areas in the neighboring region within a predetermined three-dimensional range from that pixel as a bright area distribution index. An analysis step to determine a feature region in which the brightness distribution index is equal to or greater than a predetermined threshold, A pixel count distribution is defined that shows the number of pixels for each of the bright area distribution indices in the stereoscopic image. The parameter setting step involves defining a bright area distribution index that has a significant difference from the bright area distribution index at which the number of pixels is maximized in the aforementioned pixel count distribution, as the threshold. Image processing methods.

Images