Data acquisition system and data acquisition method

Abstract

A data acquisition system (10) according to the present disclosure includes: an input unit (11) that acquires measurement data on the intensity of each prescribed wavelength of electromagnetic waves reflected by a region to be inspected of a surface of a concrete structure, where the measurement data are obtained by a hyperspectral camera capable of measuring the intensity of each wavelength of the electromagnetic waves; a detection unit (13) that detects, from a captured image obtained by imaging the region to be inspected by a camera that receives visible light to capture an image, a deteriorated / dirty region in the region to be inspected, in which deterioration or dirt is present; a deletion unit (15) that deletes, from the measurement data, data to be deleted, which is data corresponding to the detected deteriorated / dirty region; and a generation unit (18) that, on the basis of the measurement data from which the data to be deleted has been deleted, generates spectrum data indicating the intensity of each wavelengths of the electromagnetic waves reflected by the region to be inspected.

Description

Data acquisition system and method

[0001] The present disclosure relates to data acquisition systems and methods.

[0002] Concrete carbonation is a deterioration state that precedes visible deterioration of concrete structures, such as cracks, spalling, exposed reinforcement, etc. Detecting concrete carbonation makes it possible to accurately predict the deterioration of concrete structures, individually optimize inspection cycles, and create long-term facility renewal plans.

[0003] As shown in Figure 8, concrete neutralization occurs when carbon dioxide in the atmosphere penetrates into the interior of the concrete 1 from the surface 1a of the concrete 1. Specifically, calcium hydroxide and carbon dioxide in the concrete 1 react to form calcium carbonate (Ca(OH) ２ +CO ２ →CaCO ３ +H ２ O). As a result, the inside of the concrete 1, which had been kept alkaline, becomes neutralized, and a neutralized region forms toward the inside of the concrete 1. Deterioration inside the concrete 1 due to concrete neutralization cannot be detected visually. As concrete neutralization progresses and the neutralized region reaches the rebar 2 inside the concrete 1, the passive film near the rebar 2 is destroyed, causing corrosion of the rebar 2 and causing cracks and spalling of the concrete 1. It is only once cracks and spalling of the concrete 1 occur that deterioration of the concrete 1 can be visually observed.

[0004] Currently, the only method available for testing concrete neutralization is destructive testing, in which concrete cores are extracted by cutting the concrete structure. However, for the many concrete structures currently in use, non-destructive testing is desirable, as it allows for quick testing without damaging the structure.

[0005] As a non-destructive method for inspecting concrete carbonation, a method using a hyperspectral camera capable of measuring the intensity of electromagnetic waves at predetermined wavelengths has been considered (Non-Patent Document 1). In Non-Patent Document 1, a concrete structure (sound specimen) in which concrete carbonation had not occurred was carbonated, and the intensity (spectrum) of electromagnetic waves at predetermined wavelengths reflected from the surface of the concrete structure was measured every week using a hyperspectral camera. The measurement results confirmed that the reflectance (wavelength intensity) of a specific wavelength range decreases as concrete carbonation progresses. Utilizing this, Non-Patent Document 1 considers a method for evaluating the progress of concrete carbonation by comparing the spectrum obtained by imaging a concrete structure with the spectrum of a sound specimen.

[0006] Jun Arita et al., "Study on Evaluation Method of Concrete Deterioration Using Hyperspectral Remote Sensing," Production Research, 53(11 / 12), November 2001, pp. 615-618.

[0007] 9A to 9E, a description will be given of measurement of the intensity (wavelength intensity) of each predetermined wavelength of electromagnetic waves reflected from the surface of a concrete structure using the hyperspectral camera 3. There are two imaging methods for the hyperspectral camera 3: a line method and an area method.

[0008] 9A , in the line method, the concrete 1 is moved in one direction (vertical direction in the paper in FIG. 9A ) relative to the hyperspectral camera 3, while the hyperspectral camera 3 captures an image line by line in a direction intersecting the one direction (measuring the wavelength intensity of the electromagnetic waves reflected by the concrete 1). In the line method, the hyperspectral camera 3 simultaneously measures the intensity of multiple wavelengths.

[0009] 9B, in the area type, the hyperspectral camera 3 simultaneously captures an image of a certain area of ​​the concrete 1. Since the area type can only measure the intensity of one wavelength, the same area is captured multiple times while changing the wavelength.

[0010] 9C is a diagram showing an example of measurement data of wavelength intensity by the hyperspectral camera 3. As shown in FIG. 9C, １ nm, λ ２ nm, λ ３ For each of the lines (1 line, 2 lines, 3 lines, ...), the wavelength intensity of each pixel (1 pixel, 2 pixels, 3 pixels, ...) of each line (1 line, 2 lines, 3 lines, ...) is measured.

[0011] Next, as shown in FIG. 9D , a two-dimensional image of x pixels by y pixels is generated from the measurement data of the hyperspectral camera 3, which indicates the intensity of electromagnetic waves of a predetermined wavelength reflected at each position in the concrete 1. The two-dimensional image is generated from the measurement data of the hyperspectral camera 3 at multiple wavelengths λ (wavelengths λ １ , λ ２ , λ ３ 9D, a data cube is created in which the two-dimensional images generated for each of the multiple wavelengths λ are stacked in order of wavelength.

[0012] After creating the data cube, the wavelength intensities are averaged for each two-dimensional image to create spectral data showing the intensities (average intensities J1, J2, J3, ...) for each wavelength of the electromagnetic waves reflected from the surface of the concrete specimen, as shown in Figure 9E.

[0013] When imaging concrete 1 using the hyperspectral camera 3, if the surface of the concrete 1 is deteriorated (cracked, peeled, exposed reinforcement) or dirty, as shown in Figure 10, the spectral data of materials other than the concrete 1 will be mixed in, making it difficult to obtain accurate spectral data of the concrete structure.

[0014] In consideration of the above-mentioned problems, an object of the present disclosure is to provide a data acquisition system and a data acquisition method that can acquire more accurate spectral data of concrete structures using a hyperspectral camera.

[0015] In order to solve the above problem, the data acquisition system disclosed herein is a data acquisition system that acquires spectral data indicating the intensity of each wavelength of electromagnetic waves reflected from an inspection target area on the surface of a concrete structure, and includes an acquisition unit that acquires measurement data of the intensity of each specified wavelength of electromagnetic waves reflected from the inspection target area using a hyperspectral camera that can measure the intensity of each specified wavelength of electromagnetic waves; a detection unit that detects deteriorated / dirty areas in the inspection target area of ​​the concrete structure from an image of the inspection target area captured by a camera that receives visible light and captures images; a deletion unit that deletes data to be deleted from the measurement data, which is data corresponding to the detected deteriorated / dirty areas; and a generation unit that generates the spectral data based on the measurement data from which the data to be deleted has been deleted.

[0016] In order to solve the above problem, the data acquisition method disclosed herein is a data acquisition method executed by a data acquisition system that acquires spectral data indicating the intensity of each wavelength of electromagnetic waves reflected from an inspection target area on the surface of a concrete structure, and includes the steps of acquiring measurement data of the intensity of each predetermined wavelength of electromagnetic waves reflected from the inspection target area using a hyperspectral camera capable of measuring the intensity of each predetermined wavelength of electromagnetic waves, detecting deteriorated / dirty areas in the inspection target area of ​​the concrete structure from an image captured of the inspection target area using a camera that receives visible light, deleting data to be deleted from the measurement data, which is data corresponding to the detected deteriorated / dirty areas, and generating the spectral data based on the measurement data from which the data to be deleted has been deleted.

[0017] According to the data acquisition system and data acquisition method disclosed herein, more accurate spectral data of a concrete structure can be acquired using a hyperspectral camera.

[0018] 1 is a diagram illustrating an example of a configuration of a data acquisition system according to an embodiment of the present disclosure. FIG. 1 is a diagram illustrating an overview of the operation of the data acquisition system shown in FIG. 1. FIG. 1 is a diagram illustrating an example of a configuration of an input unit shown in FIG. 1. FIG. 1 is a diagram illustrating an example of a configuration of a learning unit shown in FIG. 1. FIG. 1 is a diagram illustrating an example of a configuration of a detection unit shown in FIG. 1. A flowchart illustrating an example of the operation of the data acquisition system shown in FIG. 1. FIG. 1 is a diagram illustrating an example of a configuration of a computer that functions as a data acquisition system according to the present disclosure. FIG. 2 is a diagram illustrating concrete neutralization. FIG. 3 is a diagram illustrating measurement of the reflection intensity of a concrete structure by a hyperspectral camera. FIG. 4 is a diagram illustrating measurement of the reflection intensity of a concrete structure by a hyperspectral camera. FIG. 5 is a diagram illustrating measurement of the reflection intensity of a concrete structure by a hyperspectral camera. FIG. 6 is a diagram illustrating measurement of the reflection intensity of a concrete structure by a hyperspectral camera. FIG. 7 is a diagram illustrating an example of imaging of a concrete structure by a hyperspectral camera.

[0019] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings.

[0020] Fig. 1 is a diagram illustrating an example of the configuration of a data acquisition system 10 according to an embodiment of the present disclosure. The data acquisition system 10 according to the present disclosure acquires spectral data indicating the intensity for each wavelength of electromagnetic waves reflected from an inspection target area on the surface of a concrete structure such as a manhole or a tunnel. Fig. 2 is a diagram illustrating an overview of the acquisition of spectral data by the data acquisition system 10 according to the present embodiment.

[0021] 1, a data acquisition system 10 according to this embodiment includes a data input unit 11 as an acquisition unit, a learning unit 12, a detection unit 13, a processing unit 14, a deletion unit 15, an averaging unit 16, and an output unit 17. The averaging unit 16 and the output unit 17 constitute a generation unit 18.

[0022] The data input unit 11 acquires measurement data of the intensity of electromagnetic waves at each predetermined wavelength reflected from an inspection target area of ​​a concrete structure using a hyperspectral camera capable of measuring the intensity of electromagnetic waves at each predetermined wavelength. From the input measurement data, the data input unit 11 creates a data cube in which two-dimensional images showing the intensity of electromagnetic waves reflected at each position in the inspection target area of ​​the concrete structure for each predetermined wavelength are stacked in wavelength order, as shown in Figure 2. The data input unit 11 outputs the created data cube to the processing unit 14.

[0023] 3 is a diagram showing an example of the configuration of the data input unit 11. As shown in FIG. 3, the data input unit 11 includes a measurement data input unit 111 and a data cube creation unit 112.

[0024] The measurement data input unit 111 receives measurement data of the intensity of electromagnetic waves reflected by an inspection target area of ​​a concrete structure for each predetermined wavelength, measured by a hyperspectral camera. In the case of an area type, the measurement data is expressed as an area of ​​width x pixels x height y pixels (for example, 1024 pixels x 1024 pixels) and a wavelength of the electromagnetic waves λ α nm to λ β It can be obtained by taking an image with a hyperspectral camera (measuring the intensity of the electromagnetic waves reflected in the area) while changing the wavelength from λ to 2218 nm in increments of t nm (for example, from 1295 nm to 2218 nm in increments of 9 nm). β -λ α ) / t pieces of measurement data are input to the measurement data input unit 111.

[0025] In addition, in the case of the line method, the measurement data is an area of ​​width x pixels (1024 pixels) x height 1 pixel, and the wavelength of the electromagnetic wave is λ α nm to λ β The line method captures images using a hyperspectral camera while changing the wavelength from y1 to y2 by t nm increments (for example, from 1295 nm to 2218 nm in 9 nm increments). In the line method, images are captured while moving the hyperspectral camera by y pixels (for example, 1024 pixels) in the height direction.

[0026] The measurement data input unit 111 outputs the input measurement data to the data cube creation unit 112 .

[0027] The data cube creation unit 112 generates a two-dimensional image showing the intensity of the electromagnetic wave reflected at each position (each pixel) in the x pixel × y pixel inspection area for each captured wavelength from the measurement data of the hyperspectral camera. The data cube creation unit 112 stacks the generated two-dimensional images in wavelength order to create a data cube of x pixel × y pixel × wavelength (λ β -λ α The data cube creating unit 112 outputs the created data cube to the processing unit 14.

[0028] Referring again to FIG. 1 , the learning unit 12 receives an input of a deteriorated / dirty area image, which is an image of the surface of a concrete structure, including deteriorated / dirty areas where deterioration or dirt such as rust, streaks, efflorescence, or cracks is present, captured by a camera (RGB camera) that receives visible light for capturing images. The learning unit 12 also receives an input of a mask image in which the deteriorated / dirty areas in the deteriorated / dirty area image are labeled. The learning unit 12 constructs a detection model for detecting deteriorated / dirty areas from the captured image of the surface of the concrete structure by machine learning of a dataset including the deteriorated / dirty area image and the mask image in which the deteriorated / dirty areas in the deteriorated / dirty area image are labeled. The learning unit 12 outputs the constructed detection model to the detection unit 13.

[0029] Fig. 4 is a diagram showing an example of the configuration of the learning unit 12. As shown in Fig. 4, the learning unit 12 includes a deterioration / stain area image input unit 121, a rust mask image input unit 122, a dew line mask image input unit 123, an efflorescence mask image input unit 124, a crack mask image input unit 125, a stain mask image input unit 126, a data set creation unit 127, and a model learning unit 128.

[0030] The deteriorated / soiled area image input unit 121 receives an image of a deteriorated / soiled area captured by a camera (RGB camera) that receives visible light to capture an image of the surface of a concrete structure, including deteriorated / soiled areas such as rust, exposed lines, efflorescence, cracks, or dirt. The deteriorated / soiled area image is preferably an image captured over an area of ​​the same size as the imaging range of the hyperspectral camera, with the same number of pixels as the hyperspectral camera. The deteriorated / soiled area image input unit 121 outputs the input deteriorated / soiled area image to the dataset creation unit 127.

[0031] The rust mask image input unit 122 receives a rust mask image in which rust portions included in the deteriorated / stained region image are labeled. The rust mask image is a binary image that indicates the rust portions included in the deteriorated / stained region image and other portions using two values. The rust mask image input unit 122 outputs the received rust mask image to the data set creation unit 127.

[0032] The dew line mask image input unit 123 receives a dew line mask image in which dew line portions included in the deteriorated / soiled region image are labeled. The dew line mask image is a binary image that indicates, using two values, the dew line portions included in the deteriorated / soiled region image and other portions. The dew line mask image input unit 123 outputs the input dew line mask image to the data set creation unit 127.

[0033] The efflorescence mask image input unit 124 receives an efflorescence mask image in which efflorescence portions included in the image of the deteriorated / soiled region are labeled. The efflorescence mask image is a binary image that indicates, by binary values, the efflorescence portions included in the image of the deteriorated / soiled region and other portions. The efflorescence mask image input unit 124 outputs the received efflorescence mask image to the data set creation unit 127.

[0034] The crack mask image input unit 125 receives a crack mask image in which crack portions included in the deteriorated / dirty area image are labeled. The crack mask image is a binary image that uses two values ​​to indicate the crack portions included in the deteriorated / dirty area image and other portions. The crack mask image input unit 125 outputs the input crack mask image to the data set creation unit 127.

[0035] The dirt mask image input unit 126 receives a dirt mask image in which dirt portions included in the deteriorated / dirty region image are labeled. The dirt mask image is a binary image that indicates, using two values, the dirt portions included in the deteriorated / dirty region image and other portions. The dirt mask image input unit 126 outputs the input dirt mask image to the data set creation unit 127.

[0036] For one deteriorated / stained region image, a rust mask image, a dew line mask image, an efflorescence mask image, and a crack and stain mask image are prepared and input to deteriorated / stained region image input unit 121, rust mask image input unit 122, dew line mask image input unit 123, efflorescence mask image input unit 124, crack mask image input unit 125, and stain mask image input unit 126. Therefore, when n deteriorated / stained region images are input to deteriorated / stained region image input unit 121, n rust mask images are input to rust mask image input unit 122, n dew line mask images are input to dew line mask image input unit 123, n efflorescence mask images are input to efflorescence mask image input unit 124, n crack mask images are input to crack mask image input unit 125, and n stain mask images are input to stain mask image input unit 126.

[0037] The dataset creation unit 127 creates a dataset that includes a deteriorated / soiled region image and a rust mask image, a dew streak mask image, an efflorescence mask image, a crack mask image, and a dirt mask image that correspond to the deteriorated / soiled region image. Therefore, when n deteriorated / soiled region images are input, the dataset creation unit 127 creates n datasets. The dataset creation unit 127 outputs the created datasets to the model learning unit 128.

[0038] The model learning unit 128 constructs a detection model for detecting deteriorated / stained areas from captured images of the surface of a concrete structure through machine learning of the dataset output from the dataset creation unit 127 (a dataset consisting of deteriorated / stained area images and mask images in which the deteriorated / stained areas are labeled). The detection model has, for example, a fully convolutional neural network (FCN) with an encoder-decoder structure as its model structure. With this structure, deteriorated / stained areas can be detected pixel by pixel by performing semantic segmentation in the detection of deteriorated / stained areas by the detection unit 13, which will be described later. Furthermore, by learning typical deteriorations found on the surface of a concrete structure using multi-class semantic segmentation, accurate classification of the type of deterioration for each pixel becomes possible.

[0039] The model learning unit 128 outputs the constructed learning model to the detection unit 13 .

[0040] 1 , the detection unit 13 receives an image of a concrete structure captured by a camera (RGB camera) that captures images using visible light. The captured image is preferably an image of the inspection target area of ​​the concrete structure captured by the hyperspectral camera, captured with the same number of pixels as the hyperspectral camera. Therefore, if the number of pixels of the hyperspectral camera is x pixels wide by y pixels high (e.g., 1024 pixels wide by 1024 pixels high), the number of pixels of the captured image is preferably x pixels wide by y pixels high.

[0041] The detection unit 13 detects deteriorated or dirty areas in the inspection target area from the input captured image. Specifically, the detection unit 13 inputs the input captured image from the camera (RGB camera) into the detection model constructed by the learning unit 12, and detects deteriorated or dirty areas in the inspection target area. The detection unit 13 outputs the detection result to the processing unit 14.

[0042] 5 is a diagram showing an example of the configuration of the detection unit 13. As shown in FIG. 5, the detection unit 13 includes a deteriorated / soiled area detection unit 131 and a result output unit 132.

[0043] The deteriorated / dirty area detection unit 131 inputs the input captured image into a detection model to detect deteriorated / dirty areas in the inspection target area of ​​the concrete structure. The deteriorated / dirty area detection unit 131 outputs the detection result of the deteriorated / dirty areas to the result output unit 132.

[0044] 2 , the result output unit 132 generates a binary image that indicates the detection result of the deteriorated / stained area by the deteriorated / stained area detection unit 131. Specifically, the result output unit 132 generates a binary image with a width of x pixels and a height of y pixels that indicates the detected deteriorated / stained area and other areas in binary form. The result output unit 132 outputs the generated binary image to the processing unit 14.

[0045] 1 , the processing unit 14 receives a data cube (measurement data from the hyperspectral camera) from the data input unit 11 and a binary image indicating the detection results of deteriorated / stained areas in the inspection target area of ​​the concrete structure from the detection unit 13. The processing unit 14 creates a data set that is a pair of the input data cube and the binary image. The processing unit 14 outputs the created data set to the deletion unit 15.

[0046] The deletion unit 15 receives a data set consisting of a data cube (measurement data from the hyperspectral camera) and a binary image showing the detection results of deteriorated / contaminated areas from the processing unit 14. The deletion unit 15 removes data to be deleted, which is data corresponding to the detected deteriorated / contaminated areas, from the measurement data. Specifically, as shown in FIG. 2 , the deletion unit 15 deletes portions corresponding to the detected deteriorated / contaminated areas from each of the multiple two-dimensional images that make up the data cube. By removing data to be deleted, which is data corresponding to the deteriorated / contaminated areas, from the measurement data, it is possible to obtain measurement data of only the concrete portions necessary for diagnosing concrete carbonation.

[0047] The deletion unit 15 outputs to the averaging unit 16 a data cube from which the portions corresponding to the deteriorated / stained regions have been deleted (measurement data from which the data to be deleted has been deleted).

[0048] As shown in FIG. 2, the averaging unit 16 calculates the data for each predetermined wavelength (λ １ , λ ２ , λ ３ For each of the two-dimensional images, the average value of the reflection intensity (average wavelength intensity J １ , J ２ , J ３ , ...) is calculated for each predetermined wavelength from the data cube after the degradation and contamination areas have been deleted (measurement data after the deletion of data to be deleted). １ , J ２ , J ３ , ..., it is possible to average the intensity of the electromagnetic waves reflected by the concrete portion excluding the deteriorated or dirty area, and to grasp the tendency and characteristics of the entire concrete portion. １ , J ２ , J ３ , . . . are output to the output unit 17.

[0049] The output unit 17 outputs the average wavelength intensity J １ , J ２ , J ３ , ..., the output unit 17 generates spectral data indicating the intensity of each wavelength of the electromagnetic wave reflected from the inspection target area of ​​the concrete structure. Specifically, as shown in FIG. 2, the output unit 17 outputs the wavelength λ (wavelength λ α From wavelength λ β The spectral data is generated to indicate the wavelength λ (up to the wavelength λ) and the intensity (average average intensity J) of the electromagnetic wave in the two-dimensional image corresponding to the wavelength λ.

[0050] As described above, the averaging unit 16 and the output unit 17 constitute the generating unit 18. Therefore, the generating unit 18 generates spectrum data indicating the intensity for each wavelength of electromagnetic waves reflected from the inspection target area of ​​the concrete structure based on the measurement data from which the deletion target data has been deleted. More specifically, the generating unit 18 generates spectrum data by averaging, for each predetermined wavelength, the intensity of electromagnetic waves shown in the two-dimensional image corresponding to that wavelength from which the deletion target data has been deleted.

[0051] Next, an operation of the data acquisition system 10 according to this embodiment will be described. Fig. 6 is a flowchart showing an example of the operation of the data acquisition system according to this embodiment, and is a diagram for explaining a data acquisition method executed by the data acquisition system 10.

[0052] The data input unit 11 acquires measurement data of the intensity of electromagnetic waves at each specified wavelength reflected from the inspection target area of ​​the concrete structure using a hyperspectral camera capable of measuring the intensity of electromagnetic waves at each specified wavelength (step S11).

[0053] The detection unit 13 detects deteriorated and dirty areas in the target area of ​​the prefecture from the captured image of the target area of ​​the concrete structure, which is captured using a camera (RGB camera) that receives visible light and captures images (step S12).

[0054] The deletion unit 15 deletes the deletion target data, which is data corresponding to the detected deteriorated / dirty area, from the measurement data of the hyperspectral camera (step S13). Specifically, as shown in Fig. 2, the deletion unit 15 deletes the portion corresponding to the detected deteriorated / dirty area from the two-dimensional image for each predetermined wavelength that constitutes the cube data created from the measurement data of the hyperspectral camera.

[0055] The generation unit 18 generates spectrum data based on the measurement data from which the data to be deleted has been deleted (step S14). Specifically, as shown in Fig. 2, the generation unit 18 averages the intensities of the electromagnetic waves shown in the two-dimensional image from which the data to be deleted has been deleted for each predetermined wavelength to generate spectrum data.

[0056] As described above, the data acquisition system 10 according to this embodiment includes a data input unit 11 as an acquisition unit, a detection unit 13, a deletion unit 15, and a generation unit 18. The data input unit 11 acquires measurement data of the intensity of electromagnetic waves at each predetermined wavelength reflected from an inspection target area on the surface of a concrete structure using a hyperspectral camera capable of measuring the intensity of electromagnetic waves at each predetermined wavelength. The detection unit 13 detects deteriorated / dirty areas in the inspection target area of ​​the concrete structure from an image of the inspection target area captured by a camera that receives visible light. The deletion unit 15 deletes data to be deleted, which is data corresponding to the detected deteriorated / dirty areas, from the measurement data of the hyperspectral camera. The generation unit 18 generates spectral data based on the measurement data from which the data to be deleted has been deleted.

[0057] By detecting deteriorated or dirty areas on the surface of a concrete structure from images captured by the camera and deleting data to be deleted, which is data corresponding to the detected deteriorated or dirty areas, from the measurement data of the hyperspectral camera, it is possible to obtain measurement data of only the concrete portion. Therefore, the data acquisition system 10 and data acquisition method according to this embodiment make it possible to acquire more accurate spectral data of a concrete structure using a hyperspectral camera.

[0058] The data acquisition system 10 described above can be realized by a computer 20 shown in FIG. 7. A program for causing the computer 20 to function as the data acquisition system 10 may be provided. The program may be stored on a storage medium or provided via a network. FIG. 7 is a block diagram showing a schematic configuration of the computer 20 functioning as the data acquisition system 10. The computer 20 may be a general-purpose computer, a dedicated computer, a workstation, a PC (Personal Computer), an electronic notepad, or the like. The program instructions may be program code, code segments, or the like for executing necessary tasks.

[0059] 7, the computer 20 includes a processor 21, a read-only memory (ROM) 22, a random access memory (RAM) 23, a storage 24, an input unit 25, a display unit 26, and a communication interface (I / F) 27. Each component is communicably connected to one another via a bus 29. The processor 21 is specifically a central processing unit (CPU), a micro processing unit (MPU), a graphics processing unit (GPU), a digital signal processor (DSP), a system on a chip (SoC), or the like, and may be configured with multiple processors of the same type or different types.

[0060] The processor 21 is a control unit that controls each component and performs various arithmetic processing. That is, the processor 21 reads a program from the ROM 22 or the storage 24 and executes the program using the RAM 23 as a work area. The processor 21 controls each component and performs various arithmetic processing in accordance with the program stored in the ROM 22 or the storage 24. In this embodiment, the ROM 22 or the storage 24 stores a program for operating the computer 20 as the data acquisition system 10 according to the present disclosure. The program is read and executed by the processor 21 to realize each component of the data acquisition system 10.

[0061] The program may be provided in a form stored on a non-transitory storage medium such as a CD-ROM (Compact Disk Read Only Memory), a DVD-ROM (Digital Versatile Disk Read Only Memory), a USB (Universal Serial Bus) memory, etc. The program may also be provided in a form downloaded from an external device via a network.

[0062] The ROM 22 stores various programs and various data. The RAM 23 temporarily stores programs or data as a working area. The storage 24 is configured with an HDD (Hard Disk Drive) or an SSD (Solid State Drive) and stores various programs including the operating system and various data.

[0063] The input unit 25 includes a pointing device such as a mouse and a keyboard, and is used to input various types of information.

[0064] The display unit 26 is, for example, a liquid crystal display, and displays various information. The display unit 26 may be a touch panel type and function as the input unit 25. The display unit 26 displays, for example, the spectral data generated by the output unit 17.

[0065] The communication interface 27 is an interface for communicating with other devices, for example, an interface for a LAN.

[0066] The following additional notes are provided regarding the above-described embodiments.

[0067] [Supplementary Item 1] A data acquisition system for acquiring spectral data indicating the intensity for each wavelength of electromagnetic waves reflected from an inspection target area on the surface of a concrete structure, comprising a control unit, wherein the control unit is configured to: acquire measurement data of the intensity for each predetermined wavelength of electromagnetic waves reflected from the inspection target area using a hyperspectral camera capable of measuring the intensity for each predetermined wavelength of electromagnetic waves; detect deteriorated / dirty areas in the inspection target area of ​​the concrete structure from an image of the inspection target area captured by a camera that receives and captures visible light; delete data to be deleted, which is data corresponding to the detected deteriorated / dirty areas, from the measurement data; and generate the spectral data based on the measurement data from which the data to be deleted has been deleted.

[0068] [Supplementary Item 2] In the data acquisition system described in Supplementary Item 1, the control unit constructs a detection model that detects deteriorated / dirty areas from an image of the surface of a concrete structure by machine learning of a dataset consisting of deteriorated / dirty area images of the surface of a concrete structure including deteriorated / dirty areas and mask images in which the deteriorated / dirty areas in the deteriorated / dirty area images are labeled, and inputs the image taken by the camera into the detection model to detect deteriorated / dirty areas in the area to be inspected.

[0069] [Supplementary Item 3] In the data acquisition system described in Supplementary Item 1 or 2, the control unit generates, from the measurement data of the hyperspectral camera, two-dimensional images indicating the intensity of electromagnetic waves reflected at each position in the inspection target area for each of the predetermined wavelengths; deletes, as the data to be deleted, data of areas corresponding to the detected deteriorated / soiled areas in each of the two-dimensional images generated for each of the predetermined wavelengths; and generates the spectral data by averaging, for each of the predetermined wavelengths, the intensities of the electromagnetic waves shown in the two-dimensional images corresponding to the wavelengths from which the data to be deleted has been deleted.

[0070] [Supplementary Item 4] A data acquisition method executed by a data acquisition system that acquires spectral data indicating the intensity for each wavelength of electromagnetic waves reflected from an inspection target area on the surface of a concrete structure, the data acquisition method comprising: acquiring measurement data of the intensity for each predetermined wavelength of electromagnetic waves reflected from the inspection target area using a hyperspectral camera that can measure the intensity for each predetermined wavelength of electromagnetic waves; detecting deteriorated / dirty areas in the inspection target area of ​​the concrete structure from an image of the inspection target area captured by a camera that receives and captures visible light; deleting data to be deleted, which is data corresponding to the detected deteriorated / dirty areas, from the measurement data; and generating the spectral data based on the measurement data from which the data to be deleted has been deleted.

[0071] [Supplementary Item 5] A non-transitory storage medium storing a program executable by a computer, the non-transitory storage medium storing the program causing the computer to operate as the data acquisition system described in any one of Supplementary Items 1 to 3.

[0072] Although the above-described embodiments have been described as typical examples, it will be apparent to those skilled in the art that many modifications and substitutions can be made within the spirit and scope of the present disclosure. Therefore, the present invention should not be interpreted as being limited by the above-described embodiments, and various modifications and alterations are possible without departing from the scope of the claims. For example, multiple building blocks shown in the block diagrams of the embodiments can be combined into one, or one building block can be divided.

[0073] REFERENCE SIGNS LIST 1 Concrete 2 Reinforcing bar 3 Hyperspectral camera 10 Data acquisition system 11 Data input unit 12 Learning unit 13 Detection unit 14 Processing unit 15 Deletion unit 16 Averaging unit 17 Output unit 111 Measurement data input unit 112 Data cube creation unit 121 Deterioration / stain area image input unit 122 Rust mask image input unit 123 Dew line mask image input unit 124 Efflorescence mask image input unit 125 Crack mask image input unit 126 Stain mask image input unit 127 Data set creation unit 128 Model learning unit 131 Deterioration / stain area detection unit 132 Result output unit 20 Computer 21 Processor 22 ROM 23 RAM 24 Storage 25 Input unit 26 Display unit 27 Communication I / F 29 Path

Claims

1. A data acquisition system for acquiring spectral data indicating the intensity of each wavelength of electromagnetic waves reflected from an inspection target area on the surface of a concrete structure, comprising: an acquisition unit that acquires measurement data of the intensity of each specified wavelength of electromagnetic waves reflected from the inspection target area using a hyperspectral camera that can measure the intensity of each specified wavelength of electromagnetic waves; a detection unit that detects deteriorated / dirty areas in the inspection target area of ​​the concrete structure from an image of the inspection target area captured by a camera that receives and captures visible light; a deletion unit that deletes data to be deleted, which is data corresponding to the detected deteriorated / dirty areas, from the measurement data; and a generation unit that generates the spectral data based on the measurement data from which the data to be deleted has been deleted.

2. A data acquisition system as described in claim 1, further comprising a learning unit that constructs a detection model that detects deteriorated / dirty areas from an image of the surface of a concrete structure by machine learning of a dataset consisting of deteriorated / dirty area images of the surface of a concrete structure including deteriorated / dirty areas and mask images in which the deteriorated / dirty areas in the deteriorated / dirty area images are labeled, and the detection unit inputs the image taken by the camera into the detection model to detect deteriorated / dirty areas in the area to be inspected.

3. A data acquisition system as described in claim 1, wherein the acquisition unit generates, from the measurement data of the hyperspectral camera, two-dimensional images showing the intensity of electromagnetic waves reflected at each position in the inspection area for each of the predetermined wavelengths; the deletion unit deletes, as the data to be deleted, data of areas corresponding to the detected deteriorated / soiled areas in each of the two-dimensional images generated for each of the predetermined wavelengths; and the generation unit generates the spectral data by averaging, for each of the predetermined wavelengths, the intensity of electromagnetic waves shown in the two-dimensional images corresponding to the wavelengths from which the data to be deleted has been deleted.

4. A data acquisition method executed by a data acquisition system that acquires spectral data indicating the intensity of each wavelength of electromagnetic waves reflected from an inspection target area on the surface of a concrete structure, comprising the steps of: acquiring measurement data of the intensity of each predetermined wavelength of electromagnetic waves reflected from the inspection target area using a hyperspectral camera that can measure the intensity of each predetermined wavelength of electromagnetic waves; detecting deteriorated / dirty areas in the inspection target area of ​​the concrete structure from an image of the inspection target area captured by a camera that receives and captures visible light; deleting data to be deleted, which is data corresponding to the detected deteriorated / dirty areas, from the measurement data; and generating the spectral data based on the measurement data from which the data to be deleted has been deleted.

Images