JP7453848B2 - Face image processing device, face image processing method - Google Patents

Description

The present invention relates to a face image processing device and a face image processing method.

There is a system in which a camera is installed around the face of a mountain tunnel to monitor the face (for example, Patent Document 1). In such a system, the working face is monitored with a camera to detect cracks in mirror shotcrete, which are a sign of face collapse or skin fall that occurs at the working face. In addition, in such a system, when attempting to detect using AI (artificial intelligence), images of cracks that occur in the mirror shotcrete of the face are required, and in order to improve the detection accuracy. , it is necessary to collect a large amount of training data that includes various crack occurrence patterns and noise from heavy machinery and people.

Japanese Patent Application Publication No. 2018-207194

However, since cracks rarely occur in mirror shotcrete in actual tunnel faces, it is extremely difficult to collect a large amount of diverse training data.
Furthermore, although it is possible to experimentally generate cracks in mirror shotcrete using heavy equipment such as breakers, there is a problem in that the crack patterns that can be obtained are limited.

The present invention has been made in view of the above circumstances, and an object thereof is to provide a face image processing device and a face image processing method that can efficiently collect images of cracks occurring in mirror shotcrete of the face. It is about providing.

In order to solve the above-mentioned problems, one aspect of the present invention includes an acquisition unit that acquires a face image of a face on which shotcrete has been applied, and a crack type input unit that inputs an instruction to select the type of crack. , a data processing unit that generates a composite image by compositing a crack image indicating that there is a crack in the shotcrete with the face image, and a processed data output unit that outputs the composite image , The crack type input section indicates the crack width that specifies the width of the crack, the crack density that specifies the degree of darkness of the shadow of the crack, and the crack in the image of the crack that actually occurred. The data processing section acquires at least one of the crack colors corresponding to the pixel values of the pixels constituting the image, and the data processing section converts the crack image into the crack image according to the type of crack selected by the crack type input section. This is a face image processing device that synthesizes a face image .

Further, in one aspect of the present invention, the acquisition unit acquires a face image obtained by capturing a face on which shotcrete has been applied, the crack type input unit inputs an instruction to select the type of crack, and the data processing unit inputs an instruction to select the type of crack. , a composite image is generated by combining a crack image indicating that there is a crack in the shotcrete with the face image, and a processed data output unit outputs the composite image and selects the type of the crack. Inputting instructions includes crack width to specify the width of the crack, crack density to specify the degree of darkness of the shadow of the crack, and an image showing the crack in which the crack actually occurred. Generating the composite image includes obtaining at least one of the crack colors corresponding to the pixel values of the pixels constituting the image, and generating the composite image includes obtaining at least one crack color corresponding to the pixel value of the pixels constituting the image, and generating the composite image includes obtaining the crack color according to the crack type selected by the crack type input section. This is a face image processing method including composing a corresponding crack image with the face image .

As described above, according to the present invention, images of cracks occurring in the mirror shotcrete of the face can be efficiently collected.

1 is a schematic block diagram showing the configuration of a face monitoring system 1 according to an embodiment of the present invention. It is a figure which shows an example of the case where a crack image is synthesized with the face image which imaged the face on which mirror shotcrete construction was performed. It is a flow chart explaining operation of face image processing device 40.

EMBODIMENT OF THE INVENTION Hereinafter, a face monitoring system using a face image processing apparatus according to an embodiment of the present invention will be described with reference to the drawings.
FIG. 1 is a schematic block diagram showing the configuration of a face monitoring system 1 according to an embodiment of the present invention.
The face monitoring system 1 includes a camera 10, terminal devices 20 (terminal devices 20a, terminal devices 20b), an operation terminal 30, a face image processing device 40, and a crack detection device 50.

The camera 10 is installed on the side wall of the tunnel at a position a predetermined distance (for example, about 10 meters) behind the tunnel face, and images the face F. The camera 10 outputs the imaging result to the crack detection device 50 as a face image. The camera 10 captures images at a timing ranging from several frames to more than ten frames per second, and sequentially outputs face images to the crack detection device 50.
Here, for the face F, "mirror shotcrete" is constructed, in which concrete is sprayed to a thickness of several centimeters (centimeters) onto the ground that is exposed when the tunnel is excavated. Ru.
The camera 10 images a face on which mirror shotcrete has been applied, and outputs a face image. Furthermore, the camera 10 may capture face images for each type of face construction cycle, such as earth excavation, shotcrete, and steel shoring construction.

A plurality of terminal devices 20 (for example, a terminal device 20a, a terminal device 20b, etc., simply referred to as the terminal device 20 if not specifically identified) are used at a site where excavation work is performed for a face. As this terminal device 20, for example, a smartphone, a tablet, or the like can be used. The terminal device 20 is carried by a worker who performs work such as excavation. The terminal device 20 is connected to the crack detection device 50 wirelessly or by wire, and has a function of communicating with the crack detection device 50. Further, the terminal device 20 includes a speaker, a touch panel, and the like.
Although a case will be described here in which there are two terminal devices 20 (terminal device 20a, terminal device 20b), there may be only one, or three or more terminal devices.

The operation terminal 30 has an input device such as a mouse and a keyboard, and a display device such as a display. Various operations can be accepted from users such as engineers. A computer may be used as this operating terminal 30.

The face image processing device 40 includes an image acquisition unit 400, a storage unit 401, a face image selection unit 402, a crack type input unit 403, a color extraction unit 404, a data processing unit 405, a labeling unit 406, a teacher data generation unit 407, and an output unit. 408.

Here, in the face image processing device 40, there are roughly two types of teacher data. The first training data is training data using an image taken of a state in which cracks have occurred in a face where mirror shotcrete has been constructed, and the second training data is training data that uses an image taken of a state in which cracks have occurred in a face where mirror shotcrete has been constructed. This training data uses a composite image in which a crack image, which is an image representing a crack, is composited with a portion where no crack has occurred in an image of a face that has been captured. The shape of such a crack is generally linear.

As the first training data, a face image obtained by capturing an image of a face in which cracks have occurred in mirror shotcrete at an actual face construction site can be used.
The second training data is an image of a face where mirror shotcrete has been constructed, and is based on technical knowledge and experience to determine what the condition would be like if cracks occurred. A composite image drawn on the face image can be used by a person who operates the crack image on the operating terminal 30.
FIG. 2 is a diagram showing an example of a case where a crack image is synthesized with a face image obtained by capturing an image of a face where mirror shotcrete has been applied. Here, a plurality of crack images as shown in crack images 201, 202, 203, 204, 205, 206, 207, 208, and 209 are synthesized on the surface of mirror shotcrete. Although a case is shown in which a plurality of crack images are combined with one face image, only one may be combined.

When using the first teacher data, an image of a crack that actually occurs in mirror shotcrete is used, so there is an advantage that the accuracy of crack determination by the trained model obtained through learning is high. On the other hand, it may not be easy to prepare the necessary number of images for learning.
When using the second teacher data, although it is not an image of an actually generated crack, various crack patterns can be drawn, and there is an advantage that it is easy to prepare the number of images necessary for learning. Here, by drawing the crack image in the composite image after reflecting the experience and knowledge of the engineer, it is also possible to draw it so as to reproduce the crack that actually occurred with high precision. Further, the image representing the crack is drawn by an engineer inputting an operation from the operation terminal 30 to an image of a face where mirror shotcrete has been applied. The crack image is expressed by shape, color, etc., taking into consideration the condition of the face and the environment of the construction site.

The image acquisition unit 400 acquires a face image of a face on which shotcrete has been applied. For example, the image acquisition unit 400 acquires a face image generated by the camera 10.

The storage unit 401 stores various data. For example, the storage unit 401 stores a plurality of face images captured by the camera 10 and acquired by the image acquisition unit 400, a composite image to be described later, and the like.

The face image selection unit 402 inputs an instruction to select a face image to be processed from among the acquired face images.
A crack type input section 403 inputs an instruction to select the type of crack.

The color extracting unit 404 extracts the color of an image area represented by a shadow caused by a crack in a face image obtained by capturing an image of a face where a crack has actually occurred in the shotcrete.
Here, if a crack occurs, a shadow will appear in that area. Using this, it is possible to extract the color of the shadow part of the crack in an image of an actually generated crack, depict the crack with the extracted color, and generate a crack image.

The data processing unit 405 generates a composite image by combining a crack image indicating that there is a crack in the shotcrete with the face image. The composite image obtained in this manner can be used as teacher data necessary for learning in the learning section 501 of the crack detection device 50.
The data processing unit 405 synthesizes an image representing a crack with the face image according to the operation details input from the keyboard, mouse, etc. of the operating terminal 30. The operation details may include conditions for synthesizing crack images, operations for drawing the crack itself, and the like.
The function of the data processing unit 405 to generate a composite image may be realized, for example, by executing image processing software. Photoshop (registered trademark) from Adobe Systems Incorporated may be used as the image processing software that performs the process of synthesizing the crack image with the face image.

The data processing unit 405 synthesizes a crack image with the face image selected by the face image selection unit 402. For example, among the plurality of face images stored in the storage unit 401, the face image selected using the mouse or the like of the operating terminal 30 is determined as the face image to be combined with the crack image, and the combination is performed (processed). target selection function).
The data processing unit 405 synthesizes a crack image corresponding to the type of crack selected by the crack type input unit 403 with the face image. For example, options for the mode of the crack image are stored in advance in the storage unit 401 for a plurality of patterns, and from among the modes of the crack image, the crack image of the pattern selected using the mouse or the like of the operation terminal 30 is used. Composes the face image (crack pattern selection function).

The aspects of the crack image include, for example, the width of the crack, the length of the crack, the shape of the crack, the darkness (opacity) of the shadow of the crack, and the color of the crack. In the case of cracks experimentally generated in mirror shotcrete, the maximum width of the crack was about 4 pixels (actual size of about 25 mm). For this reason, the data processing unit 405 accepts an operation input from the operation terminal 30 specifying which value to use between 1 pixel and 4 pixels, with 4 pixels being the maximum width, and converts the crack image according to the specified width. generate. Here, the maximum width of the crack is about 4 pixels (actual size is about 25 mm), but other values may be used in consideration of cracks that occur in mirror shotcrete.
Regarding the length of the crack, the minimum and maximum lengths are predetermined based on experimental results, and it is possible to decide which value to use within the length range by operating the operation terminal 30. It accepts input and generates a crack image according to the specified length.
Moreover, the shape of the crack may be classified according to the characteristics of the shape, such as a straight line or a W-shape (zigzag shape). The data processing unit 405 can receive an operation input from the operating terminal 30 regarding which of these classifications to use, and can generate a crack image according to the specified length.
Moreover, the density of the shadow of the crack varied in various cases as a result of the experiment. Therefore, the data processing unit 405 uses the operating terminal 30 to determine which of a plurality of density patterns to use, such as 70%, 50%, and 30%, regarding the density (opacity) of the shadow of the crack. It accepts operation input and generates a crack image based on the density of the specified crack shadow.
The color extracted by the color extraction unit 404 can be applied as the color of the crack. Here, the data processing unit 405 acquires the pixel values of pixels forming an image showing the crack in the image of the crack that actually occurred, and generates a crack image according to this pixel value. In this case, a crack image can be generated based on the color of cracks that may actually occur.

The data processing unit 405 can determine the width of the crack, the length of the crack, the shape of the crack, the darkness of the shadow of the crack, and the color of the crack according to the operation details input from the operating terminal 30. In addition, at least one of the crack width, crack length, crack shape, crack shadow density, and crack color should be specified, and items that were not specified should be specified in the past. The value of the generated crack image may be used. Alternatively, the selection may be made from among the options for the crack width, crack length, crack shape, crack shadow density, and crack color, or the operation terminal 30 may manually The image may be drawn arbitrarily.
Further, the position in the face image at which the crack image generated in this manner is to be placed may be determined in accordance with an operation input input from the operation terminal 30 specifying the position to be placed. .

The composite image thus generated by the data processing unit 405 can be used as teacher data.

Furthermore, when synthesizing a crack image with the face image, the data processing unit 405 rewrites the pixel values (in the case of an RGB image, each RGB value) of pixels included in the area corresponding to the crack in the face image. Alternatively, the crack image may be drawn on a layer different from the face image layer, and the composite image may be generated by overlapping these layers. Here, if the crack image is generated as a layer separate from the face image, coordinate data indicating the area of the crack can also be associated based on the layer corresponding to the crack image.

The labeling unit 406 gives a label (annotation) indicating a crack to a pixel corresponding to the cracked image, based on the combined image. For example, the labeling unit 406 determines whether or not each pixel included in the composite image corresponds to a cracked image, and if the pixel is a cracked image, assigns a label to the pixel. Whether or not the image is a cracked image may be determined by determining whether or not the pixels are added by the data processing unit 405.
The label assigning unit 406 may assign a label indicating "crack" to each pixel that corresponds to a crack image in the composite image, or may assign a label indicating "crack" to each pixel that corresponds to a crack image in the composite image. A label may be added.

Note that when labeling a face image obtained by capturing an image of a face where cracks have actually occurred in mirror shotcrete, it is necessary to specify which part is the crack image. Here, for example, the crack image may be specified by the engineer operating the operating terminal 30 and specifying that the part corresponding to the crack in the face image is a crack. .

In this way, the labeling unit 406 adds a label indicating "crack" to each pixel, so that in a face image or a composite image, an area of the image where pixels attached with a label indicating "crack" are adjacent to each other. It becomes possible to identify the region where cracks exist.

The teacher data generation unit 407 obtains data in which the labeling unit 406 adds labels to pixels of the composite image as teacher data. For example, the teacher data generation unit 407 generates the teacher data by obtaining a labeled face image or a composite image as the teacher data.

The output unit 408 outputs a composite image. As an output destination, it may be written in the storage unit 401 or output to the learning unit 501 of the crack detection device 50.

Next, the crack detection device 50 will be explained.
The crack detection device 50 includes a teacher data acquisition section 500, a learning section 501, a storage section 502, and a detection section 503.

The teacher data acquisition unit 500 acquires the teacher data generated by the face image processing device 40.
The learning unit 501 acquires training data in which a label representing a crack is assigned to a pixel corresponding to a crack existing in the shotcrete among pixels in a face image in which a face on which mirror shotcrete has been constructed is imaged. This method is used to learn the relationship between a face image and pixels corresponding to cracks included in the face image. As this teacher data, teacher data obtained from the face image processing device 40 can be used by the teacher data acquisition unit 500.
This learning uses AI (artificial intelligence) technology. For example, this learning uses semantic segmentation. Semantic segmentation is an algorithm using deep learning that assigns labels to all pixels in an image.
Since this semantic segmentation can detect objects pixel by pixel, it has a high recognition resolution even for cracks that occur in many linear shapes. It also has high discrimination performance. Therefore, it is possible to efficiently identify cracks even under conditions where the face under construction is photographed as a moving image using the camera 10.
In semantic segmentation, for example, the relationship between information on a pixel corresponding to a crack and information on pixels surrounding that pixel is learned.
The learning unit 501 performs learning to generate a learned model that has learned the relationship between a face image and a pixel corresponding to a crack included in the face image.

The storage unit 502 stores the trained model obtained by the learning unit 501. The storage unit 502 includes a storage medium such as an HDD (Hard Disk Drive), a flash memory, an EEPROM (Electrically Erasable Programmable Read Only Memory), or a RAM (Random Access read/write). e Memory), ROM (Read Only Memory), or these Constructed by any combination of storage media.
This storage unit 502 can use, for example, a nonvolatile memory.

The detection unit 503 uses training data in which a label indicating a crack is assigned to a pixel corresponding to a crack existing in the shotcrete among pixels in a face image in which a face on which shotcrete has been applied is captured. Then, by inputting the face image to be judged to the trained model that has learned the relationship between the face image and the pixels corresponding to the cracks included in the face image, and performing segmentation on the face image to be judged. Detecting whether there is a region corresponding to a crack.
When the detection unit 503 detects that there is a region corresponding to a crack in the face image, the output unit 504 outputs information indicating that a crack has been detected. This output may be, for example, the terminal device 20a, the terminal device 20b, etc., or may be output to a terminal device of a manager who manages the face construction site. As a result, these terminal devices can output an alarm sound indicating that a crack has occurred, or display a warning message on the screen indicating that a crack has occurred. The output unit 408 can urge workers to evacuate before a face collapse occurs by transmitting the result that a crack has been detected to the terminal device 20.

In this crack detection device 50 described above, an image acquisition section 400, a face image selection section 402, a crack type input section 403, a color extraction section 404, a data processing section 405, a labeling section 406, a teacher data generation section 407, and an output section 408. may be configured with a processing device such as a CPU (central processing unit) or a dedicated electronic circuit, for example.
Furthermore, in the crack detection device 50, the teacher data acquisition section 500, the learning section 501, the detection section 503, and the output section 504 may be configured with a processing device such as a CPU or a dedicated electronic circuit.

Next, the operation of the face image processing device 40 will be explained. FIG. 3 is a flowchart illustrating the operation of the face image processing device 40.
The image acquisition unit 400 of the face image processing device 40 acquires the face image captured by the camera 10 (step S101), and stores it in the storage unit 401. The face photographed here is a face image taken of a state in which no cracks have occurred in the mirror shotcrete that has been constructed. Note that the face image may be an image of a state in which cracks have occurred in mirror shotcrete. Further, the image acquisition unit 400 acquires a face image every time a photograph is taken by the camera 10, and stores it in the storage unit 401.

Next, the face image selection unit 402 reads a face image to be processed from among the face images stored in the storage unit 401 based on the operation input from the operating terminal 30 (step S102). The face image selected here is either a face image that captures mirror shotcrete without cracks or a face image that captures mirror shotcrete with cracks. However, if a face image of mirror shotcrete with no cracks is selected as the target, the construction condition of the mirror shotcrete with no cracks should be taken into consideration. Teacher data can be obtained by combining crack images obtained based on the results of examining what kind of cracks may occur under certain conditions.
On the other hand, when a face image of mirror shotcrete with cracks is selected as the target, the state of the cracks that have already occurred is taken into consideration, and if further cracks occur, the Training data can be obtained by combining crack images obtained based on the results of examining what kind of cracks may occur.

The color extraction unit 404 acquires pixel values indicating colors from pixels indicating cracks in the face image stored in the storage unit 401, in which mirror shotcrete with cracks is imaged (step S103). The target face image from which color is extracted is a face image different from the face image selected in step S102, and the face image in which the mirror shotcrete in which cracks have occurred is selected by the operation terminal 30. , the color of the crack may be extracted from this face image. Further, the face image whose color is to be extracted may be selected according to a selection input from the operation terminal 30.

Next, the crack type input unit 403 receives an input specifying the type of crack from the operating terminal 30 (step S104). Here, the crack type input unit 403 receives input specifying the width of the crack, the length of the crack, the shape of the crack, the density of the shadow of the crack (opacity), and the like.
When the color of the crack is extracted and the specified input of the form of the crack is received, the data processing unit 405 generates a crack image according to the extracted color and the specified inputted form of the crack (step S105). The generated crack image is combined with the face image (step S106). Note that the position in the face image at which the crack image is to be placed is specified from the operating terminal 30.

When the composite image is generated, the labeling unit 406 assigns a label to the crack image in the composite image (step S107). Once the label is assigned, the teacher data generation unit 407 acquires the labeled composite image as the teacher data, and outputs it to the crack detection device 50 as the teacher data (step S108).

In addition, in the above-described process, a plurality of types of composite images with different crack images may be generated for one face image.

In the embodiment described above, even if the face image is taken of mirror-shot concrete with no cracks, the crack image is synthesized with the face image by artificially creating pseudo-cracks through photo synthesis. Therefore, even if there is no face image taken of mirror-shot concrete in which cracks have actually occurred, a face image containing cracks can be prepared, and this can be used as training data. This makes it possible to efficiently create a large number of face images showing cracks in mirror shotcrete.

In actual tunnel faces, cracks occur less frequently in mirror shotcrete, and the crack patterns that can be obtained through experiments are limited. Therefore, it is not easy to prepare a large number of images of faces where cracks have actually occurred, but according to this embodiment, a composite image in which cracks are artificially expressed by photo composition etc. is used. A large amount of crack patterns and noise patterns can be created as learning data, making it possible to improve AI learning efficiency and judgment accuracy.

In addition, since it is possible to arbitrarily select the face image on which the crack image is to be synthesized and the aspect of the crack to be synthesized, it is possible to efficiently create a large amount of diverse training data according to the selected content. can.

In addition, even for cracks that cannot be obtained unless they actually occur, by generating and compositing crack images as described above, it is possible to create training data, greatly increasing the number of patterns that AI can detect. can be increased. That is, even if it is not possible to prepare a face image actually taken of a type of crack that may actually occur, it can be prepared by generating a composite image by utilizing the knowledge and experience of an engineer.
Further, according to the above-described embodiment, since the labeling unit 406 adds the labels, the originally necessary annotation work (selecting the cracked part using image processing software and adding the correct label) is performed manually. Since this is not necessary, the time required for annotation work can be significantly reduced.
Further, according to the embodiment described above, the time required from creating teacher data to creating a trained model can be significantly reduced.

Furthermore, in the embodiment described above, a case has been described in which the crack detection device 50 has the teacher data acquisition section 500 and the learning section 501. The learning device may be provided in a separate device, and the separate device may be configured as a learning device. Then, the learned model obtained by learning in the learning device may be written into the storage unit 502 of the crack detection device 50.

The face image processing device 40 and the crack detection device 50 in the embodiment described above may each be realized by a computer. In that case, a program for realizing this function may be recorded on a computer-readable recording medium, and the program recorded on the recording medium may be read into a computer system and executed. Note that the "computer system" herein includes hardware such as an OS and peripheral devices. Furthermore, the term "computer-readable recording medium" refers to portable media such as flexible disks, magneto-optical disks, ROMs, and CD-ROMs, and storage devices such as hard disks built into computer systems. Furthermore, a "computer-readable recording medium" refers to a storage medium that dynamically stores a program for a short period of time, such as a communication line when transmitting a program via a network such as the Internet or a communication line such as a telephone line. It may also include a device that retains a program for a certain period of time, such as a volatile memory inside a computer system that is a server or client in that case. Further, the above-mentioned program may be one for realizing a part of the above-mentioned functions, or may be one that can realize the above-mentioned functions in combination with a program already recorded in the computer system. It may be realized using a programmable logic device such as an FPGA (Field Programmable Gate Array).

Although the embodiments of the present invention have been described above in detail with reference to the drawings, the specific configuration is not limited to these embodiments, and includes designs within the scope of the gist of the present invention.

1... Face monitoring system, 10... Camera, 20, 20a, 20b... Terminal device, 30... Operation terminal, 40... Face image processing device, 50... Crack detection device, 400... Image acquisition section, 401... Storage section, 402... Face image selection section, 403...Crack type input section, 404...Color extraction section, 405...Data processing section, 406...Labeling section, 407...Teacher data generation section, 408...Output section, 500...Teacher data acquisition section, 501 ...learning section, 502...storage section, 503...detection section, 504...output section

Claims (7)

an acquisition unit that acquires a face image of a face on which shotcrete has been applied;
a crack type input section for inputting instructions for selecting a crack type;
a data processing unit that generates a composite image by combining a crack image indicating that there is a crack in the shotcrete with the face image;
a processed data output unit that outputs the composite image;
has
The crack type input section is
crack width, which specifies the width of the crack;
Crack density, which specifies the degree of darkness of the crack shadow;
A crack color corresponding to the pixel value of the pixels that make up the image showing the crack in the image of the crack that actually occurred;
Obtain at least one of the following,
The data processing unit synthesizes a crack image according to the type of crack selected by the crack type input unit with the face image.
Face image processing device. The crack type input section is
the crack width;
the crack density;
said crack color;
crack length, which specifies the length of the crack;
Crack shape, which specifies which classification it is in according to the characteristics of the crack shape;
Get at least one of
The face image processing device according to claim 1 . The data processing section is
Among the plurality of items that can be input using the crack type input section, the crack type specified by the crack type input section and the values of crack images generated in the past for items not specified by the crack type input section. A crack image is synthesized with the face image using
The face image processing device according to claim 1 or 2 . a face image selection unit for inputting an instruction to select a face image to be processed from among the acquired face images;
The face image processing device according to claim 1, wherein the data processing unit synthesizes a crack image with the face image selected by the face image selection unit. The face image processing device according to any one of claims 1 to 4 , further comprising a labeling section that adds a label indicating a crack to the crack image processed by the data processing section. a color extraction unit that extracts the color of an image area represented by a shadow caused by the crack in a face image taken of a face where a crack has actually occurred in the shotcrete;
The face image according to any one of claims 1 to 5 , wherein the data processing unit synthesizes a crack image representing a crack with the color extracted by the color extraction unit, with the face image. Processing equipment. The acquisition unit acquires a face image of a face on which shotcrete has been applied,
The crack type input section inputs an instruction to select the type of crack,
a data processing unit generates a composite image by combining a crack image indicating that there is a crack in the shotcrete with the face image;
a processed data output unit outputs the composite image;
Entering instructions to select the type of crack is
crack width, which specifies the width of the crack;
Crack density, which specifies the degree of darkness of the crack shadow;
A crack color corresponding to the pixel value of the pixels that make up the image showing the crack in the image of the crack that actually occurred;
including acquiring at least one of the following;
Generating the composite image includes:
The method includes composing a crack image corresponding to the type of crack selected by the crack type input section with the face image.
Face image processing method.

Images