Inspection support apparatus and inspection support method
The inspection support device and method address the burdensome and inaccurate nature of conventional voice-input inspection recording by using a processor to convert voice inputs to text, extract relevant phrases, and store inspection results, thereby enhancing efficiency and accuracy.
Patent Information
- Application Number
- JP2024101177
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2023-12-05
- Filing Date
- 2024-06-24
- Publication Date
- 2025-06-17
AI Technical Summary
Conventional methods for recording inspection results by voice input and text conversion are burdensome and prone to inaccuracies, especially when dealing with multiple inspection points, due to the reliance on accurate voice input and varying inspector skill levels.
An inspection support device and method that utilize a processor to receive voice inputs, convert them to text, extract phrases corresponding to registered words, and store inspection result information associated with each inspection point, thereby reducing the workload and improving accuracy.
The solution significantly reduces the workload of inputting inspection results for multiple points by allowing accurate and efficient recording of inspection results, regardless of the inspector's skill level, and facilitates easy confirmation of inspection progress and results.
Smart Images

Figure 2025090495000001_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to an inspection support device and an inspection support method, and more particularly to an inspection support device and an inspection support method used for inspections of a plurality of inspection points at a building construction site or inside a constructed building.
Background Art
[0002] When inspecting a plurality of inspection points at a building construction site or inside a constructed building, an inspector records inspection results for each inspection point. The methods for recording inspection results are diverse. As an example, there are a method of inputting and recording the content of inspection results on a terminal carried by the inspector, and a method of recording a photographed image of the inspection point together with the inspector's comments, etc.
[0003] Also, as another method for recording inspection results, there is a method in which the inspector speaks the content of the inspection results, inputs the voice into a terminal, converts the voice into text by the voice recognition function of the terminal, and records the text information (see, for example, Patent Document 1).
Prior Art Documents
Patent Documents
[0004]
Patent Document 1
Summary of the Invention
Problems to be Solved by the Invention
[0005] In a conventional method of recording inspection results by voice input and text conversion, the text information obtained by converting the inspector's voice into text may be stored as it is. In that case, since the inspector needs to speak the inspection results accurately, the burden of voice input increases, and the accuracy and reliability of the text information representing the inspection results may vary depending on the inspector's skill level, etc. Particularly, in the case of recording inspection results for each of a plurality of inspection points, the above problems may occur more prominently.
[0006] Therefore, the present invention has been made in view of the above problems, and an object thereof is to provide an inspection support device and an inspection support method for reducing the work burden involved in inputting inspection results for a plurality of inspection points at a building construction site or within a constructed building.
Means for Solving the Problems
[0007] According to the inspection support device of the present invention, the above problems are solved by an inspection support device used for inspections of a plurality of inspection points at a building construction site or within a constructed building, the device including a processor, and the processor executing, for each inspection point, a reception process for receiving an input of voice uttered by an inspector, a text conversion process for recognizing the voice and converting the voice into text, an extraction process for extracting, from the text information obtained by converting the voice into text, a phrase corresponding to a registered word when the registered word is included in the text information, and a storage process for associating and storing inspection result information based on the extracted phrase with the inspection point.
[0008] By using the inspection support device of the present invention configured as described above, inspection results can be accurately recorded based on the inspector's voice. As a result, regardless of the inspector's skill level, voice input of inspection results can be appropriately performed, and thus the work of inputting inspection results becomes easier.
[0009] In addition, in the inspection support device of the present invention, it is preferable that the processor executes a display process for displaying on the screen an image of a building indicating the positions of a plurality of inspection points, and in the image displayed in the display process, each of the plurality of inspection points is displayed in a display mode in which it is possible to identify whether or not it has been inspected. According to the above configuration, the positions of a plurality of inspection points, the progress status of inspections, etc. can be easily confirmed through the image of the building displayed on the screen.
[0010] Moreover, it is more preferable that in the display process, the processor displays information corresponding to the inspection result information stored in association with the inspection point in the region of the image corresponding to the inspected inspection point. According to the above configuration, the inspection results of the inspected inspection points can be easily confirmed.
[0011] In addition, in the inspection support device of the present invention, it is preferable that the processor executes a specifying process for specifying a designated inspection point specified from among a plurality of inspection points, and in the storage process, the processor stores the inspection result information based on the words and phrases extracted in the extraction process in association with the designated inspection point specified in the most recently executed specifying process. According to the above configuration, when storing the inspection result information, the inspection result information and the designated inspection point can be appropriately associated with each other.
[0012] In addition, in the inspection support device of the present invention, the processor executes an assignment process for assigning, to each of two or more words included in the text information, a label corresponding to the word from among a plurality of labels including an exclusion label, and in the extraction process, the processor extracts, from among the two or more words, a word having a predetermined order relationship with a word to which a label corresponding to a registered word is assigned and having a label other than the exclusion label assigned thereto, which is preferable. According to the above configuration, it is possible to appropriately extract a phrase corresponding to a registered word from the text information obtained by converting the voice input by the inspector into text.
[0013] In addition, in the inspection support device of the present invention, when a plurality of registered words are included in the text information, in the extraction process, the processor extracts, for each registered word, a phrase corresponding to the registered word from the text information, and in the storage process, it is preferable to store, in association with the inspection location, inspection result information based on the phrases extracted for each registered word. According to the above configuration, when a plurality of registered words are included in the text information obtained by converting the voice input by the inspector into text, it is possible to record inspection result information indicating the content of the inspection result corresponding to each registered word.
[0014] In addition, when inspections are carried out on one inspection location from two or more directions and the text information contains two or more registered words indicating the directions in which the inspections are carried out, in the extraction process, the processor extracts, for each direction, a phrase corresponding to the direction from the text information, and in the storage process, it is more preferable to store, in association with the inspection location, inspection result information based on the phrases extracted for each direction. According to the above configuration, when inspections are carried out on each inspection location from two or more directions, it is possible to record inspection result information indicating the content of the inspection result corresponding to each inspection direction.
[0015] In addition, in the inspection support device of the present invention, for specific inspection locations where inspections are carried out on two or more inspection positions among the plurality of inspection locations, it is preferable that the reception process, the text conversion process, the extraction process, and the storage process are performed for each of the two or more inspection positions. According to the above configuration, the inspector can focus on the inspection work for each inspection position each time, and if the inspection results are spoken at the end of each inspection work, the input of the inspection results regarding the inspection position can be efficiently performed. On the other hand, when the above configuration is not adopted, on the inspector side, it may be necessary to memo or memorize the inspection results of each of a series of inspection positions and speak them as a single sentence. However, by adopting the technology of the present invention, the situation of imposing such a burden on the inspector can be eliminated. As a result, the work burden for inputting inspection results for a plurality of inspection locations at a building construction site or within a constructed building can be further reduced.
[0016] Also, in the inspection support device of the present invention, if the processor does not execute the storage process for the phrase extracted in the immediately preceding extraction process or deletes the phrase already stored in the immediately preceding storage process when the text information obtained by the text conversion process contains a cancellation word, it is preferable. According to the above configuration, in an environment where the input of inspection results for each inspection position is possible, a simple and accurate voice cancellation response can be made for mis-inputs (mis-speeches) that may occur for each inspection position (each speaking opportunity). As a result, the work burden for inputting inspection results for a plurality of inspection locations at a building construction site or within a constructed building can be further reduced.
[0017] Also, in the inspection support device of the present invention, if the processor does not execute the storage process of the phrase or executes the deletion according to the cancellation word for a predetermined inspection position, and then re-executes the reception process, the text conversion process, the extraction process, and the storage process for the inspection position, it is preferable. According to the above configuration, after the cancellation response for mis-inputs or the like, the input (speech) of the inspection results regarding the inspection position can be smoothly resumed. As a result, the work burden for inputting inspection results for a plurality of inspection locations at a building construction site or within a constructed building can be further reduced.
[0018] In addition, in the inspection support device of the present invention, if the processor executes a switching process for the inspection location to be the target of the reception process when a specific word is included in the text information obtained by the text conversion process, it is preferable. According to the above configuration, for a plurality of inspection locations where inspections are sequentially performed in a series of processes, simple switching according to the intention of the inspector and the inspection situation becomes possible. As a result, it is possible to further reduce the work load involved in inputting inspection results for a plurality of inspection locations at a building construction site or within a constructed building.
[0019] In addition, in the inspection support device of the present invention, if the processor performs voice output for one or more of the phrases extracted by the extraction process, it is preferable. According to the above configuration, it becomes possible to confirm through repetition by the inspection support device whether the speech content by the inspector has been correctly recognized or whether the speech content itself by the inspector is not incorrect. As a result, it is possible to further reduce the work load involved in inputting inspection results for a plurality of inspection locations at a building construction site or within a constructed building.
[0020] Further, the above-described problem is solved by an inspection support method of the present invention, which is an inspection support method used for inspections of a plurality of inspection locations at a building construction site or within a constructed building, and includes a step of receiving an input of voice uttered by an inspector by a processor, a step of recognizing the voice and converting the voice into text by the processor, a step of extracting, by the processor, a phrase corresponding to a registered word from the text information when the registered word is included in the text information obtained by converting the voice into text, and a step of storing, by the processor, inspection result information based on the extracted phrase in association with the inspection location, for each inspection location. According to the above inspection support method, for a plurality of inspection locations at a building construction site or within a constructed building, it is possible to easily perform the work of inputting inspection results (voice input), so the work load is reduced.
Effect of the Invention
[0021] According to the inspection support device and inspection support method of the present invention, it is possible to reduce the work load involved in inputting inspection results for a plurality of inspection locations at a building construction site or within a constructed building.
Brief Description of the Drawings
[0022]
Figure 1
Figure 2
Figure 3
Figure 4
Figure 5A
Figure 5B
Figure 6
Figure 7
Figure 8
Figure 9
Figure 10
Figure 11
Figure 12
Figure 13
Figure 14
Figure 15
Figure 16
Figure 17
Figure 18
Figure 19
Figure 20
Figure 21
Mode for Carrying Out the Invention
[0023] <<Regarding One Embodiment of the Present Invention>> Hereinafter, one embodiment of the present invention (hereinafter, this embodiment) will be described with reference to the accompanying drawings. Note that basic data processing technologies (communication / transmission technology, data acquisition technology, data recording technology, data processing / analysis technology, speech recognition technology, natural language processing technology, image processing technology, visualization technology, etc.) for realizing the content of this embodiment are well-known technologies, and thus the description thereof will be omitted.
[0024] [Outline of the Inspection According to This Embodiment] First, the outline of the inspection according to this embodiment will be described with reference to FIG. 1. The inspection according to this embodiment is an inspection carried out on a plurality of inspection locations at a building construction site. Specifically, it is an inspection for diagnosing the appropriateness of the construction status of each inspection location in the building construction work. Here, the building includes single-family houses or apartment houses such as condominiums, and also includes buildings other than houses, specifically, commercial facilities such as offices, stores, movie theaters, and department stores, public facilities such as hospitals and schools, factories, or buildings for various purposes such as warehouses. Hereinafter, the inspection carried out at a housing construction site will be described as an example.
[0025] After a predetermined process is carried out in the construction work of a house, as shown in FIG. 1, the inspector conducts an inspection regarding management items for a plurality of inspection locations where work (construction) is carried out in the process. The process is various types of work carried out during the construction work, for example, building construction work, steel bar placement work, and interior finishing work, etc. are carried out. The plurality of inspection locations and management items are set according to the process. For example, in the case of building construction work, a plurality of columns built in the building construction work become the plurality of inspection locations, and the building state of the columns (specifically, the inclination of the columns) corresponds to the management item. In this case, the position of the core of the column to be inspected is determined in advance, and the inspection locations are set according to the position of the core.
[0026] When conducting the inspection, the inspector confirms the positions of a plurality of inspection locations at the construction site through the screen of the terminal 12 held by the inspector, and designates any one of the inspection locations (hereinafter also referred to as the designated inspection location). Then, the inspector moves towards the designated inspection location and conducts the inspection of the location after arriving at the designated inspection location. Note that the plurality of inspection locations are set and registered in advance by the inspector before the inspection is carried out. Also, regarding the inspection locations, at a stage after the registration of the inspection locations, for example, immediately before the inspection is carried out, they may be increased or decreased conversely according to the construction situation.
[0027] Also, there may be a case where a plurality of inspections for one inspection location are carried out by changing the inspection direction. For example, when inspecting the building state of a column, the inclination of the column defined by the amount of deviation from the normal position is inspected (measured) in each of two directions (hereinafter referred to as the X direction and the Y direction) that are orthogonal to each other in the horizontal direction. Note that the inspection direction is not limited to two, and may be only one, or three or more.
[0028] After the inspection at the specified inspection location is completed, as shown in Fig. 1, the inspector uses the terminal 12 in his / her possession to input the inspection results. The inspection results may be the numerical values (measurement values) of the items measured during the inspection, or may be words, phrases or their abbreviations indicating the suitability of the construction (the result of the pass / fail judgment), such as "OK" or "NG", or may be words, phrases or their abbreviations indicating the construction status of the inspection location, such as "The current position is off by a few millimeters from the specified position".
[0029] In this embodiment, as shown in Fig. 1, the inspector inputs the inspection results by voice, specifically, speaks the content corresponding to the inspection results. The voice spoken by the inspector when inputting the inspection results includes registered words.
[0030] The registered words are words that directly or indirectly represent the management items as the inspection content, are determined for each inspection content, and are pre-registered before the inspection is carried out. Specific examples of the registered words include "column inclination" which is the inspection content after the building construction process, "rising cover thickness" and "reinforcement specification" which are the inspection content after the reinforcement process, and "diagonal dimension" which is the inspection content after the building method process, etc. Also, when inspections are carried out from two or more directions for one inspection location, as words indirectly representing the inspection content, two or more registered words indicating the directions in which the inspections are carried out, for example, "X direction" and "Y direction", are used.
[0031] Note that the registered words may also be words corresponding to numerical values such as "1" and "2", or words corresponding to symbols such as "A" and "X". That is, when identification numerical values or symbols are assigned to the registered words to simplify them further, the identification numerical values or symbols may be used as the registered words.
[0032] When voice input is performed, the examiner usually speaks, immediately after the registered word, information corresponding to the inspection result of the management item indicated by the registered word. Also, when using two or more registered words indicating the direction in which the inspection is to be performed, the examiner speaks, immediately after each registered word (direction), information corresponding to the corresponding inspection result. Here, taking as an example the case where the inclination of the column is spoken as the inspection result for each of the X direction and the Y direction, the examiner speaks, for example, "1 mm in the X direction, 0 mm in the Y direction".
[0033] Thereafter, the voice input by the examiner is, as shown in FIG. 1, converted into text by the voice recognition function and converted into text information. Also, the text information is divided into two or more words by natural language processing, and the two or more words included in the text information are divided into phrases having a certain meaning and coherence by natural language processing. In the case where the voice "1 mm in the X direction, 0 mm in the Y direction" is input as described above, it is divided into words "X (ex)", "direction (houkou)", "1 (ichi)", "mm (milli)", "Y (wai)", "direction (houkou)", "0 (zero)", "mm (milli)", and divided into phrases "X direction", "1 mm", "Y direction", "0 mm".
[0034] And when a registered word is included in the text information, a phrase corresponding to the registered word is extracted from the text information as indicating the inspection result. The phrase corresponding to the registered word is usually the phrase immediately after the registered word. Also, when two or more registered words are included in the text information, for each registered word, a phrase corresponding to the registered word is extracted from the text information. In the case where the voice "1 mm in the X direction, 0 mm in the Y direction" is input as described above, "1 mm" is extracted as the phrase corresponding to the registered word "X direction", and "0 mm" is extracted as the phrase corresponding to the registered word "Y direction".
[0035] Thereafter, as shown in FIG. 1, inspection result information is generated based on the extracted phrases. To generate inspection result information based on the extracted phrases, the information represented by the extracted phrases (specifically, numerical values, determination results, etc.) may be used as the inspection result information as it is, or the calculated value obtained by substituting the numerical value indicated by the extracted phrases into a predetermined arithmetic expression may be used as the inspection result information. Further, information obtained by determining the magnitude relationship between the numerical value indicated by the extracted phrases and a reference value, or information obtained by determining whether the numerical value indicated by the extracted phrases is within a reference range may be used as the inspection result information. Note that the inspection result information may include information other than the above information. For example, it may include the sum of inspection results (numerical values) at each of a plurality of inspection locations, the difference between inspection locations, the trend of inspection results for the entire plurality of inspection locations, and the difference between the inspection result (numerical value) and the reference value.
[0036] The generated inspection result information is stored in association with the inspection location. More specifically, the inspection result information is stored in association with the designated inspection location specified at the time of performing the inspection from which the inspection result indicated by the information is obtained.
[0037] The series of operations described above are performed for each inspection location and repeated until inspection result information is stored for all of the plurality of inspection locations. Note that in the above case, each time an inspection is performed at each inspection location, the operations from voice input of the inspection result to storage of the inspection result information are performed for that inspection location, but the present invention is not limited to this. For example, after inspections for all of the plurality of inspection locations are completed, the operations from voice input of the inspection result to storage of the inspection result information may be sequentially performed for each of the plurality of inspection locations.
[0038] Further, the inspector can confirm the inspection result information stored in association with the inspection location. More specifically, the inspection result information can be displayed on the screen of the terminal 12 held by the inspector in association with the inspection location (see FIGS. 5A and 5B), and the inspector can visually recognize the inspection result information through the screen of the terminal 12. Note that the method of checking the inspection result information, that is, the output method of the inspection result information is not limited to the above method. For example, the voice indicating the inspection result information may be played and output on the terminal 12 side, or the inspection result information may be printed and output by a printer (not shown).
[0039] [Inspection Support System According to This Embodiment] Next, an inspection support system (hereinafter referred to as inspection support system S) according to this embodiment will be described. The inspection support system S is an information communication system used when an inspector conducts inspections on a plurality of inspection locations at a housing construction site. As shown in FIG. 2, the inspection support system S includes an inspection support device 10 and an inspector's terminal 12. The inspection support device 10 and the terminal 12 are connected to each other via a communication network N such as the Internet or a mobile communication line so as to be able to communicate with each other.
[0040] The terminal 12 is an information processing terminal used by an inspector and is composed of, for example, a smartphone, a mobile phone, a tablet terminal, a laptop computer, or a wearable terminal. The terminal 12 includes a display composed of a display, a recording device (sound collecting device) composed of a microphone, and a photographing device composed of a camera. Further, the terminal 12 transmits and receives data to and from the inspection support device 10 via the communication network N, and information transmitted from the inspection support device 10, for example, can be displayed on the display screen configured by the display. Note that the display may be a display provided in the terminal 12 itself or a display connected to the terminal 12 in a wired or wireless form. In addition to a general stationary display, the display connected to the terminal 12 may include an HMD (Head Mounted Display) such as a VR goggles.
[0041] An inspection application program (hereinafter referred to as the inspection app) is installed on the terminal 12. By launching the inspection app on the terminal 12, the inspector can specify inspection information and check a plurality of inspection locations where the inspection related to the specified inspection information is to be carried out. The inspection information is information indicating the construction site where the inspection is to be performed, the process to be inspected in the construction work, and the area where the inspection is to be performed at the construction site.
[0042] Also, after the inspection is carried out, the inspector speaks the content corresponding to the inspection result, and inputs (records) the voice to the terminal 12 in a state where the inspection app is launched. The terminal 12 can convert the input voice into data and transmit the voice data to the inspection support device 10 via the communication network N.
[0043] The inspection support device 10 is a device used for inspecting a plurality of inspection locations at the construction site of a house, and is composed of a computer, and is constituted by, for example, a personal computer (PC), a workstation, a server computer, or the like. The inspection support device 10 may be constituted by one computer, or may be constituted by a plurality of computers that are parallel-distributed.
[0044] Also, when the computer constituting the inspection support device 10 is a server computer, it may be a server computer for ASP (Application Service Provider), SaaS (Software as a Service), PaaS (Platform as a Service), or IaaS (Infrastructure as a Service). In this case, when necessary information is input at the client terminal, the above server computer performs various processes (calculations) based on the input information, and the calculation result is output on the client terminal side. That is, the function of the server computer that is the inspection support device 10 can be utilized on the client terminal side.
[0045] As shown in FIG. 2, the computer that constitutes the inspection support device 10 includes a processor 21, a memory 22, a storage 23, a communication interface 24, an input device 25, and an output device 26.
[0046] The processor 21 is constituted by, for example, a CPU (Central Processing Unit), an MPU (Micro-Processing Unit), an MCU (Micro Controller Unit), a GPU (Graphics Processing Unit), a DSP (Digital Signal Processor), a TPU (Tensor Processing Unit), or an ASIC (Application Specific Integrated Circuit). The memory 22 is constituted by, for example, semiconductor memories such as a ROM (Read Only Memory) and a RAM (Random Access Memory).
[0047] The storage 23 is constituted by, for example, a flash memory, an HDD (Hard Disc Drive), an SSD (Solid State Drive), an FD (Flexible Disc), an MO disk (Magneto-Optical disc), a CD (Compact Disc), a DVD (Digital Versatile Disc), an SD card (Secure Digital card), or a USB memory (Universal Serial Bus memory). The storage 23 may be built into the computer main body that constitutes the inspection support device 10, or may be attached to the computer main body in an external form. Further, the storage 23 may be constituted by a NAS (Network Attached Storage) or the like, or may be an external device that can communicate with one computer that constitutes the inspection support device 10 through a communication network N, for example, an online storage or a database server.
[0048] The communication interface 24 may be configured by, for example, a network interface card, a communication interface board, or the like. The computer constituting the inspection support device 10 can perform data communication with other devices connected to the communication network N via the communication interface 24.
[0049] The input device 25 is configured by, for example, a keyboard, a mouse, or a touch panel. The output device 26 is configured by, for example, a display and a speaker.
[0050] Also, programs for an operating system (OS) and programs for inspection are installed as software in the computer constituting the inspection support device 10. By these programs being read and executed by the processor 21, the computer constituting the inspection support device 10 exhibits its functions. Specifically, in cooperation with the terminal 12, it executes a series of data processing related to the inspection.
[0051] In addition, information related to the construction work of the house, various types of information necessary for the inspection, and inspection result information are accumulated in the storage 23 to construct a database. The information related to the construction work of the house includes information on the work name, work period, specifications of the house to be constructed, and construction location, and also includes an image of the house to be constructed. The image of the house may be a drawing of the house (design drawing or construction drawing) drawn by BIM (Building Information Modeling) or CAD (Computer Aided Design), or may be a photographed image of an existing house having the same structure as the house planned for construction.
[0052] Among the information stored in the storage 23, various information necessary for inspections includes information regarding a plurality of inspection locations at the construction site, information necessary for the assignment of labels described later, a speech recognition model, and a natural language processing model. The speech recognition model and the natural language processing model are mathematical models constructed by machine learning, specifically, those configured by AI (Artificial Intelligence).
[0053] The inspection result information is stored in association with the construction site of the house and the inspection locations at that construction site, and indicates the content of the inspection result corresponding to the registered word. Also, when inspections are carried out from two or more directions for one inspection location, the inspection result information indicates the inspection results for each direction. The inspection result information may be further associated with the date of inspection and the information of the inspector.
[0054] [Functions of the Terminal] Next, the functions of the terminal 12 will be described. The terminal 12, when operated by the inspector, acquires information regarding the content input, selected, and specified by the inspector. Specifically, at the start of the inspection, the terminal 12 displays the input screen shown in FIG. 3. The inspector inputs inspection information through this input screen. Specifically, the inspector inputs the construction site where the inspection is to be carried out, the process to be inspected in the construction work, and the area where the inspection is to be carried out at the construction site. The terminal 12 acquires these input information as inspection information. Then, the terminal 12 transmits the acquired inspection information towards the inspection support device 10.
[0055] Also, as shown in FIG. 4, the terminal 12 displays an image of a house showing the positions of a plurality of inspection locations at the construction site on the screen of the display. The display data for this image is provided by the inspection support device 10, and the terminal 12 displays the above image by expanding the display data received from the inspection support device 10. In the image displayed at this time, each of the plurality of inspection locations is displayed in a display mode in which it is possible to identify whether it has been inspected or not. Specifically, the plurality of inspection locations are represented by, for example, circular marks in the image, and the display color of the marks is different between the case where it has been inspected and the case where the inspection has not been carried out. In the example shown in FIG. 4, the black mark indicates the inspected inspection location, and the white mark indicates the inspection location where the inspection has not been carried out. Note that the display mode of the inspection location may be changed according to whether it has been inspected or not, and may be, for example, other than the display color, such as the shape or size of the mark, the presence or absence of a sign displayed near the inspected inspection location, etc.
[0056] The inspector designates any one of the plurality of inspection locations displayed on the screen shown in FIG. 4 (specifically, any one of the inspection locations where the inspection has not been carried out) as the designated inspection location, and for example, touches the designated inspection location displayed on the above screen with a finger or performs an operation of instructing with a pointer or the like on the terminal 12. Thereby, the terminal 12 acquires the information of the designated inspection location designated by the inspector. Then, the terminal 12 transmits the acquired information of the designated inspection location (specifically, the information of the coordinate position of the designated inspection location in the XY direction) toward the inspection support device 10.
[0057] Also, as described above, when the inspector speaks the content according to the inspection result, the terminal 12 records the voice, converts the recorded voice into data, and transmits the voice data toward the inspection support device 10 through the communication network N.
[0058] In addition, the terminal 12 displays inspection result information (strictly speaking, information corresponding to the inspection result information) on the screen of the display. The display data for the inspection result information is provided by the inspection support device 10, and the terminal 12 expands the display data received from the inspection support device 10 to display the inspection result information. As a display method for the inspection result information, for example, as shown in FIG. 5A, the inspection result information may be displayed in an image of a house showing the positions of a plurality of inspection points at the construction site described above. In this case, the inspection result information may be displayed in an area corresponding to the inspection point (specifically, the inspected inspection point) associated with the inspection result information among the above images. Here, the area corresponding to the inspection point in the image means an area near the inspection point or an area where the inspection point exists. When displaying the inspection result information in the area corresponding to the inspection point, the position of the inspection point in the image, specifically, the XY coordinates of the inspection point may be specified, and the inspection result information may be displayed according to the specified position of the inspection point.
[0059] Note that when displaying the inspection result information in the image of the house, if the entire image of the house is displayed, the display of the inspection result information may be omitted, and when one inspected inspection point is enlarged and displayed, the inspection result information associated with that inspection point may be displayed. At this time, as shown in FIG. 5A, information indicating the position of the inspection point may be displayed. Specifically, information indicating the coordinate positions of the inspection point in each of the X direction and the Y direction may be displayed. In the case shown in FIG. 5A, the inspection point is a column, and as information indicating the position of the inspection point, information indicating the XY coordinates of the center of the column (denoted as X0, X1, X2, Y3, Y4 in the figure) is displayed.
[0060] In addition, as another method for displaying the inspection result information, for example, as shown in FIG. 5B, the inspection results of each inspection point may be displayed in a list form. In this case, each column (row) of the list shows the position of the inspection point, the specifications of the inspection point (for example, "type of column" etc. for inspection points in the building construction process), and the inspection results at the inspection point, etc.
[0061] [Functions of the Inspection Support Device] Next, the functions of the inspection support device 10 will be described. The functions of the inspection support device 10 are realized by the processor 21 of the inspection support device 10 executing an inspection program. In other words, the processor 21 executes various processes defined in the inspection program in order to exert the functions of the inspection support device 10. Specifically, when an inspector performs inspections on each of a plurality of inspection locations and inputs inspection results, the processor 21 repeatedly executes display processing, identification processing, reception processing, text conversion processing, allocation processing, extraction processing, and storage processing for each inspection location in a corresponding manner. Hereinafter, each process executed by the processor 21 will be described.
[0062] (Display Processing) The display processing is a process of displaying an image of a house indicating the positions of a plurality of inspection locations on the screen of the inspector's terminal 12. The display processing is executed based on the inspection information sent from the terminal 12. Specifically, the processor 21 identifies the construction site, process, and inspection area input by the inspector from the inspection information received from the terminal 12. Then, based on these identified information, the processor 21 selects an image to be displayed from the images of houses stored in the storage 23. Also, a plurality of inspection locations pre-registered for the above construction site, process, and inspection area are identified. Then, the processor 21 reflects the positions of the identified plurality of inspection locations in the image to be displayed, generates display data for an image showing the positions of the plurality of inspection locations, and transmits the generated display data to the terminal 12.
[0063] In the display processing, as described above, in the image of the house that is the display target, each of the plurality of inspection locations may be displayed in a display mode in which it is possible to identify whether it has been inspected. That is, the above display data may be generated so that it is possible to identify whether each inspection location has been inspected in a display mode.
[0064] Also, in the display process, as shown in FIG. 5A, the processor 21 may display information corresponding to the inspection result information stored in association with the inspected inspection location in the area corresponding to the inspected inspection location among the images of the houses to be displayed. That is, the above display data may be generated so that information corresponding to the inspection result information is displayed in the area corresponding to the inspected inspection location. Note that the information corresponding to the inspection result information may include characters and numerical values representing the content indicated by the inspection result information, figures and symbols corresponding to the content indicated by the inspection result information, and information related to the content indicated by the inspection result information (for example, arrows and signs indicating the inspection locations associated with the inspection result information).
[0065] (Specific process) The specific process is a process of specifying a designated inspection location designated by an inspector from among a plurality of inspection locations at a housing construction site. The specific process is executed based on the information on the designated inspection location sent from the terminal 12. Specifically, the processor 21 specifies the designated inspection location designated by the inspector from among a plurality of inspection locations (strictly speaking, inspection locations where inspections have not been performed) based on the information on the designated inspection location received from the terminal 12.
[0066] (Reception process) The reception process is a process of receiving the voice uttered by the inspector, specifically, the input of the voice when the inspector speaks the content corresponding to the inspection result. In the reception process, the processor 21 receives the voice data recorded by the terminal 12 from the terminal 12 via the communication network N, thereby receiving the input of the voice.
[0067] (Text conversion process) The text conversion process recognizes the voice received in the reception process, that is, the voice when the inspector speaks the content according to the inspection result, and converts the voice into text. In the text conversion process, the processor 21 uses a voice recognition model (AI) to convert the voice into text. That is, when the voice is input into the voice recognition model, text information can be obtained as its output. The text information is generated in the same language as the inspector's language. For example, when the inspector speaks in Japanese, Japanese text information is generated.
[0068] (Assignment process) The assignment process is a process of assigning a label corresponding to each of two or more words included in the text information from among a plurality of preset labels. In the assignment process, the processor 21 first uses a natural language processing model (AI) to divide the text information into two or more words. Then, the processor 21 assigns a label corresponding to the meaning of each of the two or more words included in the text information by means of a named entity recognition (NER) technique.
[0069] The plurality of labels are a plurality of labels including an exclusion label, specifically, BIO (Begin Inside Outside) tags. In the present embodiment, the correspondence relationship between the word and the label (that is, the label assignment rule) is determined in advance and stored in advance as table data as shown in FIG. 6, for example. The processor 21 refers to this table data and assigns a label to each of the two or more words included in the text information.
[0070] Here, the label assignment procedure will be specifically described by taking the case where text information such as "1 mm in the X direction and 0 mm in the Y direction" is obtained as an example. The words included in the above text information are "X", "direction", "1", "mm", "Y", "direction", "0", "mm". On the other hand, in the table data shown in FIG. 6, for the words "X" and "Y", the label "B-RESULT", which is the label of the direction of the inspection result, is associated. Also, for the word "direction", the label "I-DIRECTION", which indicates that it is the direction of the inspection result, is associated. Also, for words representing numerical values such as "one" and "zero", the label "I-RESULT", which is the label of the numerical value of the inspection result, is associated. Also, for the word "milli", the label "I-UNIT", which indicates that it is the unit of the inspection result, is associated. From the above relationships, the following labels are assigned to each word included in the text information "1 mm in the X direction and 0 mm in the Y direction". "X": B-RESULT "direction": I-DIRECTION "1": I-RESULT "mm": I-UNIT "Y": B-RESULT "direction": I-DIRECTION "0": I-RESULT "mm": I-UNIT
[0071] Also, when the inspector utters unnecessary languages such as "ah" and "eh", so-called fillers, during voice input, the text information obtained by converting the voice into text includes words corresponding to the fillers. On the other hand, in the table data shown in FIG. 6, for words other than the above four types of labels (B tags, I tags) associated with fillers, an exclusion label "O (Outside)" is associated. Therefore, when the text information includes a word corresponding to a filler, the exclusion label "O" is assigned to that word. Note that the exclusion label may be assigned to words other than fillers, specifically, words not related to the inspection result.
[0072] Note that the label is not limited to the BIO tag, and labels other than the BIO tag, such as the BILUO tag, may be used. Also, the assignment of labels is not limited to referring to the table data shown in FIG. 6, and a learning model that assigns appropriate labels based on the meaning of words, that is, AI, may be used to assign labels to each word.
[0073] (Extraction process) The extraction process is a process of extracting a phrase corresponding to a registered word from the text information when the registered word is included in the above text information. Also, when a plurality of registered words are included in the text information, the processor 21 extracts, in the extraction process, a phrase corresponding to the registered word for each registered word from the text information. More specifically, when inspections are performed from two or more directions for one inspection location and the text information includes two or more registered words indicating the directions in which the inspections are performed, the processor 21 extracts, in the extraction process, a phrase corresponding to the inspection direction for each direction from the text information.
[0074] In the extraction process, the processor 21 extracts a word (hereinafter referred to as a corresponding label word) to which a label corresponding to the registered word is assigned and that has a predetermined order relationship among two or more words included in the text information and to which a label other than the exclusion label is assigned.
[0075] The label corresponding to the registered word is a label associated with the word constituting the registered word. For example, when the registered word is "X direction", the label 『B-RESULT』 associated with the word "X (ex) " and the label 『I-DIRECTION』 associated with the word "direction" correspond to the label corresponding to the registered word. Note that the label corresponding to the registered word may be one or two or more.
[0076] A word that has a predetermined order relationship with the word of the corresponding label is a word that appears after the word of the corresponding label in the text information. A word to which a label other than the exclusion label is assigned is a word to which a label other than the label 'O' is assigned. In short, it is a word other than a word not related to the inspection including fillers.
[0077] Then, a word that has a predetermined order relationship with the word of the corresponding label and to which a label other than the exclusion label is assigned (hereinafter referred to as the corresponding word) is extracted as a phrase corresponding to the registered word. Also, when there are two or more corresponding words, a corresponding word (strictly speaking, a group of corresponding words) having a certain meaning and coherence is extracted as a phrase corresponding to the registered word.
[0078] Regarding the procedure for extracting a phrase corresponding to the registered word, taking the text information of "1 mm in the X direction, 0 mm in the Y direction" as an example, it will be specifically described with reference to FIG. 7. The registered words are "X direction" and "Y direction", and the words of the corresponding labels are "X (ex) " and "direction (hōkō)", and "Y (wai)" and "direction (hōkō)". In this case, as shown in FIG. 7, the words in a predetermined order with the words "X (ex)" and "direction (hōkō)" of the corresponding label are "1 (ichi)" and "mm (milli)". Since these are words to which a label other than the exclusion label is assigned, they correspond to the corresponding words. Also, "1mm (ichimilli)" which combines "1 (ichi)" and "mm (milli)" corresponds to a group of corresponding words having a certain meaning and coherence. Therefore, in the extraction process, as shown in FIG. 7, "1mm" is extracted as the phrase corresponding to the registered word "X direction".
[0079] Similarly, the words "Y (wai)" and "direction (houkou)" which are the words of the corresponding label and the words in a predetermined order are, as shown in FIG. 7, "0 (zero)" and "mm (milli)", and since these are the words to which labels other than the exclusion labels are assigned, they correspond to the corresponding words. Further, "0mm (zero milli)" which is a combination of "0 (zero)" and "mm (milli)" corresponds to a group of corresponding words having a certain meaning and coherence. Therefore, in the extraction process, as shown in FIG. 7, "0mm" is extracted as the phrase corresponding to the registered word "Y direction".
[0080] (Memory process) The memory process is a process of associating and storing the inspection result information based on the phrase extracted in the extraction process, that is, the phrase corresponding to the registered word, with the inspection location. The inspection result information based on the extracted phrase is information representing the extracted phrase (specifically, a numerical value, a determination result, etc.) as it is, information of a calculated value obtained by substituting the numerical value indicated by the extracted phrase into a predetermined arithmetic expression, information indicating the magnitude relationship between the numerical value indicated by the extracted phrase and a reference value, or information indicating whether the numerical value indicated by the extracted phrase is within a reference range.
[0081] In the memory process, when the processor 21 associates and stores the inspection result information with the inspection location, the processor 21 associates and stores the inspection result information with the designated inspection location specified in the specific process executed most recently.
[0082] Also, when a plurality of registered words are included in the text information, in the extraction process, in order to extract the phrase corresponding to the registered word from the text information for each registered word, accordingly, in the memory process, the processor 21 associates and stores the inspection result information based on the phrase extracted for each registered word with the inspection location (designated inspection location). More specifically, when the text information includes two or more registered words indicating the direction in which the inspection is performed, in the extraction process, the phrase corresponding to the inspection direction is extracted from the text information for each direction, and in the memory process, the inspection result information based on the phrase extracted for each direction is associated and stored with the inspection location (designated inspection location).
[0083] [Inspection Support Flow According to This Embodiment] Next, the inspection support flow, which is an information processing flow using the inspection support device 10, will be described with reference to FIG. 8. The inspection support flow according to this embodiment employs the inspection support method of the present invention and proceeds along the flow shown in FIG. 8. That is, each step in the flow illustrated in FIG. 8 corresponds to each element constituting the inspection support method of the present invention. Note that the inspection support flow shown in FIG. 8 is merely an example, and unnecessary steps may be deleted, new steps may be added, or the execution order of the steps may be changed within the scope not departing from the gist of the present invention.
[0084] The inspection support flow is started by using, as a trigger, the fact that the inspector operates the terminal 12 to start the inspection application. In each step of the inspection support flow, the processor 21 of the computer constituting the inspection support device 10 executes the processing corresponding to each step.
[0085] When the inspection support flow is started, first, the processor 21 receives inspection information from the terminal 12 held by the inspector (S001). Then, based on the received inspection information, the processor 21 selects an image to be displayed from the images of the houses stored in the storage 23 and displays the selected image on the screen of the display provided in the terminal 12 (S002).
[0086] In step S002, the positions of a plurality of inspection points are shown in the image of the house displayed on the screen. By looking at the image of the house displayed on the screen, the inspector can grasp the positions of the plurality of inspection points shown in the image. Also, in step S002, each inspection point is displayed in a display mode in which it is possible to identify whether it has been inspected. Therefore, the inspector can easily recognize which inspection points have been inspected.
[0087] Then, among the multiple inspection locations, the inspector designates any of the uninspected inspection locations as the designated inspection location, and the terminal 12 transmits the information of the designated inspection location designated by the inspector to the inspection support device 10. On the inspection support device 10 side, the processor 21 receives the information of the designated inspection location sent from the terminal 12 (S003), and based on the received information, identifies the designated inspection location designated by the inspector (S004).
[0088] After that, the inspector conducts an inspection on the designated inspection location, speaks the content according to the inspection result, and inputs the voice at the terminal 12. The input voice data is transmitted from the terminal 12 to the inspection support device 10. On the inspection support device 10 side, the processor 21 receives the above voice data and thereby accepts the input of the voice (S005).
[0089] Next, the processor 21 recognizes the voice accepted in S005 and converts the voice into text (S006). When text information is obtained in this step S006, the processor 21 divides the text information into two or more words, and assigns a label corresponding to each word from among a plurality of labels to each word (S007).
[0090] And when a registered word is included in the text information obtained in step S006, the processor 21 extracts, from the text information, a phrase corresponding to the registered word based on the labels assigned to each word in step S007 (S008). Specifically, among the two or more words included in the text information, the corresponding word that has a predetermined order relationship with the word of the corresponding label and to which a label other than the exclusion label is assigned is extracted as the phrase corresponding to the registered word. When there are two or more corresponding words, a group of corresponding words having a certain meaning and coherence is extracted as the phrase corresponding to the registered word.
[0091] In addition, when two or more registered words are included in the text information, in step S008, the processor 21 extracts, for each registered word, a phrase corresponding to the registered word from the text information.
[0092] Next, the processor 21 stores, in association with the inspection location, inspection result information based on the phrases extracted in step S008 (S009). In this step S009, the inspection result information is stored in association with the most recently specified designated inspection location. Also, in step S008, when phrases corresponding to the registered words are extracted from the text information for each registered word, in step S009, the inspection result information based on the phrases extracted for each registered word is stored in association with the inspection location (strictly speaking, the designated inspection location).
[0093] Among the series of steps described above, steps S003 to S009 are repeatedly executed (S010) until the inspection result information is stored for all of the plurality of inspection locations. And when the inspection result information is stored for all of the plurality of inspection locations, the inspection support flow ends. Also, when the construction site where the inspection is performed, the process and / or area to be inspected changes, and the inspection is performed at a newly set plurality of inspection locations, a new inspection support flow will be executed according to the above procedure.
[0094] <<Regarding the effectiveness of this embodiment>> By using the inspection support device 10 and the inspection support method of this embodiment, it is possible to facilitate the input of inspection results for a plurality of inspection locations at a housing construction site and reduce the burden of the input work. More specifically, in this embodiment, when the inspector speaks the content of the inspection result, the inspector speaks the registered word and then speaks the content of the inspection result in sequence. As a result, the text information obtained by converting the inspector's voice into text will include the registered word.
[0095] Then, identify the registered words included in the text information, and extract from the text information the phrases corresponding to the registered words, for example, the phrases immediately following the registered words. The phrases extracted in this way are likely to be the phrases corresponding to the content of the inspection results. As a result, information regarding the inspection results can be accurately obtained from the text information of the voice input by the inspector, and the inspection results can be accurately recorded. Due to such an effect, even an inspector with little inspection experience can appropriately perform voice input so that the inspection results are accurately recorded. As a result, voice input regarding the inspection results becomes easier, and the work burden of the input is reduced.
[0096] Moreover, according to the present embodiment, the versatility of inputting inspection results is improved, and it is possible to flexibly respond even when the place where the inspection is carried out (construction site) is different, the inspection content is different, and the inspection location is different, and in any case, the inspection results can be appropriately input.
[0097] Also, in the present embodiment, when extracting the phrases corresponding to the registered words from the text information, for each of two or more words included in the text information, a label corresponding to each word is assigned from among a plurality of labels. Then, based on the labels assigned to each word, the phrases corresponding to the registered words are extracted from the text information. Thereby, the phrases corresponding to the registered words can be appropriately and easily extracted from the text information. Moreover, even if the word spoken by the inspector does not exactly match the registered word, by assigning a label corresponding to the registered word to that word, the phrase corresponding to that word can be appropriately extracted as indicating the detection result. Furthermore, even if two or more inspection results are spoken continuously, based on the positions of the registered words (the order of speech), the phrases indicating each inspection result can be appropriately extracted.
[0098] In addition, among the plurality of labels, there is an exclusion label. Even if the text information contains inappropriate words (specifically, filler words or words not related to inspection) as the extracted phrases, the exclusion label is assigned to such words, so that the situation where the above words are extracted can be avoided. As a result, the phrases corresponding to the registered words can be more appropriately extracted from the text information.
[0099] <<Application Example of the Present Embodiment>> As an application example of the present embodiment, as described above, in the construction work of a house, a case where a plurality of columns built in the building process are inspection locations and the construction state of each column (specifically, the inclination of each column) is inspected can be cited. In such a case, each column is inspected from two directions, the X direction and the Y direction. When inputting the voice of the inspection result, the registered words are set as "X direction" and "Y direction", and after speaking each registered word, the inclination of the column in each direction (the amount of deviation from the normal position) is spoken. Specifically, it is spoken like "1 mm in the X direction, 0 mm in the Y direction". Note that there is no particular limitation on how the inspection results are spoken. In the above case, the numerical values indicating the inclination of each column in the X direction and the Y direction are stored for each column as inspection result information (see FIG. 5B).
[0100] By the way, as application examples of the present embodiment, in addition to the above case, other examples can be considered. For example, three examples described later can be cited. Hereinafter, each of the three application examples will be described. (First Application Example) The first application example is a case where, during the construction of a house, for a plurality of inspection points in a mat foundation constructed in the foundation work, the position of the steel bars arranged in the foundation, specifically, the rising cover thickness is inspected. In such a case, at each inspection point, the dimensions of three different parts (rising cover thickness) are inspected. Also, when inputting the voice of the inspection result, the registration word is set as "rising cover thickness", and after uttering the registration word, the rising cover thickness of each of the three parts is uttered. Specifically, it is uttered as "rising cover thickness 42mm, 44mm, 45mm", etc. Note that there is no particular limitation on how the inspection result is uttered. In the above case, as shown in FIG. 9, for each inspection point in the mat foundation, the numerical values indicating the rising cover thicknesses of the three parts are stored for each inspection point as inspection result information.
[0101] (The second application example) The second application example is a case where, during the construction of a house, for a plurality of inspection points in an exterior wall panel arranged in the exterior wall work, the alignment of the exterior wall is inspected. In such a case, as shown in FIG. 10, the plurality of inspection points are arranged substantially in a straight line. Also, when inputting the voice of the inspection result, the registration word is set as "alignment", and after uttering the registration word, the value of the alignment is uttered. Specifically, it is uttered as "alignment +1mm", etc. Note that there is no particular limitation on how the inspection result is uttered. In the above case, as shown in FIG. 10, for each inspection point in the exterior wall panel, the numerical value indicating the alignment of the exterior wall panel is stored for each inspection point as inspection result information.
[0102] (The third application example) The third application example is a case where, in the construction work of a house, a plurality of sashes arranged in the building process are used as a plurality of inspection points, and the diagonal dimensions of each sash are inspected. In such a case, for each sash, two diagonal dimensions are inspected. Here, let one of the two diagonal dimensions be diagonal dimension a, and the other be diagonal dimension b. Also, when inputting the voice of the inspection result, the registration words are set as "diagonal dimension a" and "diagonal dimension b", and after uttering each registration word, the value of each diagonal dimension is uttered. Specifically, it is uttered as "diagonal dimension a 1365 mm, diagonal dimension b 1367 mm", etc. Note that there is no particular limitation on how the inspection result is uttered. In the above case, as shown in FIG. 11, for each of the plurality of inspection points (sashes), numerical values indicating two diagonal dimensions a and b are stored for each inspection point as inspection result information.
[0103] <<Regarding Other Embodiments>> So far, one embodiment of the inspection support device 10 and the inspection support method of the present invention has been described. However, the above embodiment is merely an example for facilitating the understanding of the present invention and does not limit the present invention. That is, the present invention can be changed and improved without departing from its gist. Also, it goes without saying that equivalents of the present invention are included.
[0104] Also, the screen examples shown in the drawings referred to in the above description are merely examples. The configuration example of the screen, the content of the information displayed, the GUI (Graphical User Interface), etc. can be freely designed according to the system design specifications, the user's preferences, etc., and can also be changed as appropriate.
[0105] Also, in the above embodiment, it is assumed that the inspection support device 10 is configured by a computer that can communicate with the terminal 12 of the inspector. However, it is not limited to this, and the terminal 12 alone, or in cooperation with a computer that can communicate with the terminal 12, may configure the inspection support device 10.
[0106] Also, in the above-described embodiment, when designating a specified inspection location from among a plurality of inspection locations at a housing construction site, an image of the housing showing the plurality of inspection locations is displayed on the screen of the inspector's terminal 12, and the inspector designates any one of the plurality of inspection locations on the display screen. However, the present invention is not limited to this, and the inspector may designate the specified inspection location by speaking the position of the specified inspection location and inputting the voice at that time. In this case, the inspector may speak (input) the position of the specified inspection location and the inspection result at the specified inspection location at once, such as "X4, Y5, 1 mm in the X direction, 0 mm in the Y direction". In this example, "X4, Y5" corresponds to a phrase indicating the position of the specified inspection location (coordinate position in the XY direction), and "1 mm in the X direction, 0 mm in the Y direction" corresponds to a phrase indicating the inspection result at the specified inspection location, specifically, the inspection result for each direction.
[0107] Also, when the inspector speaks and inputs both the position of the specified inspection location and the inspection result at the specified inspection location as described above, the processor 21 of the inspection support device 10 receives the input of the voice indicating the position of the specified inspection location and the inspection result at the specified inspection location in the reception process. Further, the processor 21 converts the voice into text to obtain text information and assigns a label to each of the plurality of words included in the text information in the same procedure as in the above-described embodiment. At this time, labels for the position of the specified inspection location (in the case of BIO tags, 'B-XSTREET', 'B-YSTREET') are assigned to "X (ex) " and "Y (why) " in "X4, Y5". Also, labels for numerical values related to the position of the specified inspection location (in the case of BIO tags, 'I-XSTREET', 'I-YSTREET') are assigned to "4 (yon) " and "5 (go) " in "X4, Y5". Then, in the identification process, the specified inspection location is identified based on the words to which the above-described labels for the position of the specified inspection location are assigned and the words to which the labels for numerical values related to the position of the specified inspection location are assigned.
[0108] In addition, when the inspector announces the content of the inspection result, the registered word may be included, and this registered word may also be used as a word for designating the specified inspection location. Here, taking the second application example described above, that is, the case of inspecting along the outer wall for a plurality of inspection locations on the outer wall panel, as an example for specific explanation, assume that the positions of the plurality of inspection locations are, for example, "X3, X5, X9". In this case, when the inspector designates each inspection location as the specified inspection location and announces the inspection result (specifically, along the wall) at each inspection location, the inspector announces something like "along the wall X3 +1mm, X5 -1mm, X9 0mm". In this case, in addition to "along the wall", "X3", "X5", and "X9" are used as registered words and are used as clues when extracting the sentences indicating the inspection result in the extraction process. Also, "X3", "X5", and "X9" are also used as words for specifying the positions of the specified inspection locations.
[0109] In the above-described embodiment, it is assumed that the inspector designates the specified inspection location through the terminal 12. However, the present invention is not limited to this, and the inspection support device 10 may automatically determine the specified inspection location without the inspector performing the operation of designating the specified inspection location.
[0110] In the above-described embodiment, the inspection of a plurality of inspection locations at the construction site of a building is taken as an example for explanation, but the present invention is not limited to this. The present invention can also be applied to the inspection of a plurality of inspection locations in a constructed building. That is, the inspection support device 10 and the inspection support method described in the above-described embodiment can also be used, for example, in the case of periodically inspecting the deterioration state of building materials or equipment installed in a completed (completed) house.
[0111] In addition to the above-described embodiments, embodiments that correspond to the situation of performing inspections on two or more inspection positions at one inspection location and that aim to reduce the labor and improve the efficiency of voice input of the inspection results can also be adopted. In this embodiment, on the side of the inspector who performs the above inspection, in order to prepare for voice input, measures such as memorizing or storing each inspection result of a series of inspection positions are no longer necessary. To memorize or store each inspection result of a series of inspection positions, the inspector needs to make a memo or memory for each inspection, which is cumbersome. In addition, there is also a concern that the inspector may be distracted by the repeated memo operation and it will be difficult to concentrate on the important inspection operation. Also, when speaking a series of inspection results obtained from a memo or the like as a single sentence, since the single sentence contains a plurality of numerical values and information, it is easy to make mistakes or omissions in speech. In that case, there is also a possibility that the labor and time for rephrasing may increase instead.
[0112] Therefore, in the inspection support device 10, for a specific inspection location where inspections are performed on two or more inspection positions, each of the reception process, text conversion process, extraction process, and storage process described with reference to FIG. 8 and the like is to be performed for each inspection position. Therefore, the inspection support device 10 is to hold at least the word information 110 in its memory 22 or storage 23. FIG. 12 shows a configuration example of this word information 110.
[0113] The word information 110 illustrated in FIG. 12 is information that designates, as “registered words”, phrases similar to the “words” in the table data (a table defining labels corresponding to respective prescribed words) shown in FIG. 6, and defines operations to be performed when such phrases are extracted from the voice of the examiner's speech. In the illustrated word information 110, it is a collection of records associating the registered words with the operations, with an ID that uniquely identifies a record defining the operation for each registered word as a key. Among these, examples of registered words include phrases such as “mm”, “cancel”, “next”, and “back”. Also, examples of operations include “extract and organize inspection results and display the results” for “mm”, “delete the currently displayed inspection results in the reverse order of speech (display) made” for “cancel”, “save the inspection results and transition to the next inspection location” for “next”, and “move to the previous inspection result without saving the inspection results” for “back”.
[0114] In this case, in the inspection support apparatus 10, the processor 21 uses a model (AI) for speech recognition to convert the voice of the examiner who desires voice input of inspection results into text and obtains text information. Also, the processor 21 uses a model (AI) for natural language processing to split the text information into words. The processor 21 collates the words obtained here with the word information 110 to specify and execute the control content for display and storage of the inspection results regarding one inspection location included in the text information.
[0115] Hereinafter, the inspection support method in the above-described embodiment will be described with reference to the drawings. FIG. 13 is a diagram showing an inspection support flow according to another embodiment of the present invention. As already described, it is assumed that an input screen (distributed from the inspection support apparatus 10) corresponding to inspection information (for example, the construction site or property where the inspection is performed, the process to be inspected, and the area where the inspection is performed at the construction site) is displayed on the terminal 12 operated by the examiner (S100).
[0116] Such input screens correspond to screens D10 to D15 shown in FIGS. 17 to 19, for example. Taking the screen D10 in FIG. 17 as an example, the input screen has a configuration including a construction drawing D101 that displays the inspection location in an overhead view, an object D102 indicating the inspection location, a display column D103 for inspection items and the like, and a display column D104 for inspection results. Note that, as exemplified in FIG. 4, the object D102 indicating the inspection location in such an input screen is displayed in a display mode in which it is possible to identify whether it has been inspected or not. Specifically, it is preferable to subject it to appropriate display control such as making the display color of the object different between the case where it has been inspected and the case where the inspection has not been carried out.
[0117] On the other hand, the inspector views the above input screens on the terminal 12 and recognizes a plurality of inspection locations D102 in the construction drawing, the inspection items to be carried out for each of the inspection locations (in the example in the drawing, "the rising cover thickness of the fabric foundation reinforcement"), and the inspection positions (in the example in the drawing, "A" to "C"). The inspector carries out inspections for each of the plurality of inspection locations, for example, in the order of the numbers indicated by the object. The terminal 12 or the inspection support device 10 identifies the uncompleted status of such inspections based on the status of the storage process of the inspection results, and targets the inspection locations that have not been completed for inspection implementation and inspection result input. Therefore, the inspector speaks the inspection results for each of the inspection positions of the inspection locations that have not been completed toward the terminal 12.
[0118] Also, when the inspector speaks the content corresponding to the inspection result, the terminal 12 records the voice, converts the recorded voice into data, and transmits the voice data toward the inspection support device 10 through the communication network N. The inspection support device 10 receives the voice data and acquires text information by applying it to a model (AI) for speech recognition (S101). At this time, as described above, the inspection support device 10 uses a model (AI) for natural language processing to split the above text information into words. Suppose the word group obtained here is, for example, "A, 42mm".
[0119] The inspection support device 10 collates the word group obtained in S101 with the word information 110, and determines, for example, whether it includes the registered word "mm" with ID "1" (S102). As shown in FIG. 12, this registered word "mm" indicates an opportunity to extract the word spoken immediately before the registered word, that is, the value of the inspection result. As a result of the above determination, when it is found that the registered word "mm" is included (S102: Yes), the inspection support device 10 acquires the inspection item label and the inspection result, and displays this on the input screen (S103), and returns the process to S101. The inspection item label is the word "A" spoken immediately before the value "42" of the inspection result among the above word group, that is, the word indicating the inspection position. Therefore, the pair "A 42mm" of the inspection position and the inspection result obtained there is specified by the inspection support device 10, and the information is displayed on the input screen D10 etc. of the terminal 12 (see FIG. 18). On the other hand, as a result of the above determination, when it is found that the registered word "mm" is not included (S102: No), the inspection support device 10 transfers the process to S104.
[0120] Note that as described above, when the inspection position, the value of the inspection result, and a specific registered word are spoken in order, for example, when the inspection position and the value of the inspection result are obtained in text form, the inspection support device 10 holds the information obtained so far as the information to be displayed, that is, the information being displayed 111, in the memory 22, and causes the terminal 12 to display this (see the screen D10 in FIG. 17). An example of the information being displayed 111 held in the memory 22 in the above situation is shown in FIG. 14. In the example of FIG. 14, based on the drawing "Z1" of the property "P001" under construction, for the inspection item "reinforcement cover thickness", at the inspection location "3", at the inspection position "A", the inspection result "42mm" is obtained, indicating that the information is being displayed.
[0121] Also, following the inspection position and the value of the inspection result, when a specific registered word "mm" is spoken and the word is texturized, the inspection support device 10 stores the information obtained so far as the information to be subject to memory processing, that is, the stored information 112, in the memory 22 or the storage 23. As such memory processing is completed for each inspection position of a certain inspection location, for example, in ascending order of the numbers assigned to each inspection location, voice input of the inspection result regarding the next inspection location will be accepted. In the example of the stored information 112 shown in FIG. 15, it shows that the memory processing has been completed up to the "2"nd among the inspection locations "1" to "3" where the inspection item "reinforcement cover thickness" should be inspected. Also, for the inspection location "3", it shows that the memory processing has been completed only for the inspection position "A".
[0122] In addition, the status of each process of display and memory in the case where the inspection result and the specific registered word "mm" are included in the speech by the inspector, which has been described so far, is conceptually shown as the first stage in FIG. 16. Hereafter, FIG. 16 will be appropriately referred to and explained according to the progress of the flow.
[0123] As a result of the determination in step S102, when it is determined that the registered word "mm" is not included (S102: No), the inspection support device 10 collates the word group obtained in S101 with the word information 110, and determines, for example, whether it includes the registered word "cancel" or "delete" of ID "2" (S104). As shown in FIG. 12, this registered word "cancel" or "delete" indicates an opportunity not to execute memory processing for the word spoken immediately before the registered word, that is, the value of the inspection result, or to delete the value that has been stored in the most recent memory processing.
[0124] As a result of the above determination, if it is found that the registered word "cancel" or "delete" is included (S104: Yes), the inspection support device 10 deletes (without executing the storage process) the display on the input screen for the word spoken immediately before the registered word, that is, the value of the inspection result (S105), and returns the process to S101. In this case, in the display information 111, the value is deleted. If the utterance of the registered word "mm" is recognized, that is, after the storage process of the inspection result is executed, if the utterance of the registered word "cancel" or "delete" is recognized, the inspection support device 10 deletes the value of the inspection result stored in the storage information 112 in the most recent storage process. Such a situation corresponds to the situation of transitioning from the second stage to the fourth stage in FIG. 16. That is, in the example of FIG. 16, it corresponds to the situation where after the utterance of "A 42mm B 44mm", the inspector utters "cancel". In this case, the inspection result of "B 44mm" that was most recently stored is deleted from the display on the input screen (including deletion from the display information 111) and is also deleted from the storage information 112.
[0125] Note that, in response to the result of the determination in S104 above (the presence or absence of the cancel word "cancel" or "delete"), after the inspection support device 10 executes step S105 and transitions to S101, this corresponds to re-executing the reception process, text conversion process, extraction process, and storage process for the relevant inspection position. Such a situation is shown, for example, by the transition from the second stage to the fourth stage or from the fifth stage to the sixth stage in FIG. 16, where the inspection results for the inspection positions "A" and "B" are obtained and displayed from the inspector's utterance content, but for the inspection position "B" that is the target of "cancel", the display and storage are deleted. After that, the processes after the reception process are executed again for the inspection position "B" or another inspection position "C". The screen transition corresponding to the transition from the fifth stage to the sixth stage is shown in screens D14 and D15 of FIG. 8. This figure shows the situation where after the display for the inspection position "B" is deleted, the inspection result "43mm" is displayed for the inspection position "C".
[0126] On the other hand, if it is determined as a result of the above determination that the registered word "cancel" or "delete" is not included (S104: No), the inspection support device 10 collates the word group obtained in S101 with the word information 110, and determines, for example, whether it includes the registered word "next" with ID "3" (S106). As shown in FIG. 12, this registered word "next" indicates an opportunity for a switching operation to the next inspection location where the inspection result has not been stored (or instructed by the inspector) from the inspection location where the inspection result has been stored up to immediately before the registered word. By performing such control, it becomes possible to perform a simple switch according to the intention of the inspector and the inspection situation for a plurality of inspection locations where inspections are sequentially performed in a series of processes.
[0127] If it is determined as a result of the above determination that the registered word "next" is included (S106: Yes), the inspection support device 10 executes a storage process (if necessary) for the words spoken up to immediately before the registered word, that is, the values of the inspection results at each inspection position, and executes a transition of the input screen or the like with the next inspection location as the reception target of the inspection result (S107). Such a situation corresponds to the situation where the transition is made from the second stage to the third stage in FIG. 16. That is, in the example of FIG. 16, it corresponds to the situation where after the utterance of "A 42mm B 44mm", the inspector utters "next" (the situation where "next" is uttered on the input screen D11 in FIG. 17). In this case, the inspection result of "A 42mm B 44mm" that has been stored up to this point is the target of the storage process, and a screen transition to the new inspection location "4" is to be made.
[0128] Subsequent to the above step S107, the inspection support apparatus 10 determines whether there is an inspection item or an inspection location for which the storage process of the inspection result is incomplete, or whether an end instruction has been received from the terminal 12 (S110). For those for which the storage process of the inspection result is incomplete, in the stored information 112, it is sufficient to search for the existence of inspection items or inspection locations whose inspection result columns contain blanks. As a result of the above determination, if there is no inspection item or inspection location for which the storage process of the inspection result is incomplete, or if an end instruction has been received from the terminal 12 (S110: Yes), the inspection support apparatus 10 ends this flow. On the other hand, as a result of the above determination, if there is an inspection item or an inspection location for which the storage process of the inspection result is incomplete, or if an end instruction has not been received from the terminal 12 (S110: No), the inspection support apparatus 10 returns the process to S110.
[0129] Incidentally, as a result of the determination in S106, if it is found that the registered word "next" is not included (S106: No), the inspection support apparatus 10 collates the word group obtained in S101 with the word information 110, and determines, for example, whether it includes the registered word "return" with ID "4" (S108). As shown in FIG. 12, this registered word "return" indicates the opportunity for the switching (return) operation to the inspection location where the most recent display process or storage process was performed without storing the inspection result displayed up to immediately before the registered word. Therefore, the inspection support apparatus 10 does not execute the storage process for the words spoken up to immediately before the registered word, that is, the values of the inspection results at each inspection position, and executes the transition of the input screen, etc. with the previous inspection location as the reception target of the inspection result (S109), and returns the process to S100. Thereafter, in the above step S110, the above respective processes are repeatedly executed until there are no uninspected items or the like, or until an end instruction is received from the terminal 12.
[0130] Incidentally, the inspection support apparatus 10 can recognize the speech content of the inspection result by the inspector as intended by the inspector, and can provide the inspector with an opportunity for verification by repetition as to whether it can be extracted as the inspection result. Based on FIGS. 20 and 21, the above process of repetition will be described. In this case, the inspection support apparatus 10 acquires a word group such as "A +1mm" based on the speech data of the speech obtained through the input screen displayed on the terminal 12 (S200, S201).
[0131] The inspection support device 10 collates the word group obtained in S201 with the word information 110, and determines, for example, whether it includes the registered word "mm" with ID "1" (S202). As a result of the above determination, if it is found that the registered word "mm" is included (S202: Yes), the inspection support device 10 acquires the inspection item label and the inspection result, and displays them on the input screen (S203). Along with this, the inspection support device 10 inputs the word group "A +1mm" to be displayed into a predetermined speech synthesis engine, and outputs speech from the output device 26 (S204). The form of the information corresponding to the processing so far is shown in the row of "First stage" in FIG. 20. Note that the speech synthesis engine is a program that generates and outputs speech corresponding to the input text, and can be held by the inspection support device 10 itself or called from an external system via the communication network N for use.
[0132] The inspector listens to the speech output from the output device 26 and checks whether it correctly corresponds to the inspection result he / she spoke. As a result of this check, if it cannot be determined that the speech correctly indicates the inspection result, the inspector speaks a registered word such as "Cancel". In this case (S205: No), the inspection support device 10 deletes the display of the word spoken immediately before the registered word, that is, the value of the inspection result (S206), and returns the process to S200. In the case of the example in FIG. 20, if the inspection support device 10 repeats "A +1mm" and the inspector who heard the repetition speaks "Cancel", the display of the inspection result of "+1mm" for "A" is deleted and also deleted from the stored information 112.
[0133] Thereafter, while collaborating with the terminal 12, the inspection support device 10 repeatedly executes the same processing as described above in the second and third stages in FIG. 20, and transitions to processing related to the next inspection position or the end of the flow. By performing such processing, it becomes possible to confirm through repetition whether the speech content by the inspector was correctly recognized or whether the speech content itself by the inspector was not incorrect. By not only displaying the speech content on the screen but also prompting confirmation by voice, it becomes possible to further reduce the workload involved in inputting inspection results for a plurality of inspection locations at the construction site of a building or within the constructed building.
Explanation of Signs
[0134] 10 Inspection support device 12 Terminal 21 Processor 21 22 Memory 23 Storage 24 Communication interface 25 Input device 26 Output device N Communication network S Inspection support system 111 Word information 112 Displayed information 113 Stored information
Claims
1. An inspection support device used for inspecting a plurality of inspection points at a building construction site or within a constructed building, a processor, the processor comprising: A reception process for receiving voice input from an examiner; A text conversion process for recognizing the speech and converting the speech into text; an extraction process for extracting a phrase corresponding to the registered word from the text information obtained by converting the speech into text when the registered word is included in the text information; and storing test result information based on the extracted words and phrases in association with the test location for each of the test locations.
2. The processor executes a display process to display an image of the building showing the positions of the plurality of inspection points on a screen; The inspection support device according to claim 1 , wherein in the image displayed in the display process, each of the plurality of inspection locations is displayed in a display manner that makes it possible to identify whether or not the inspection location has been inspected.
3. 3. The inspection support device according to claim 2, wherein the processor, in the display processing, displays information corresponding to the inspection result information stored in association with the inspection location in an area of the image corresponding to the inspection location that has been inspected.
4. The processor executes a process of identifying a designated inspection point from among the plurality of inspection points; The inspection support device according to claim 1 , wherein in the storage process, the processor stores the inspection result information based on the words extracted in the extraction process in association with the designated inspection location identified in the most recently executed identification process.
5. the processor executes an assignment process of assigning, to each of two or more words included in the text information, a label corresponding to the word from among a plurality of labels including an exclusion label; 5. The inspection support device according to claim 1, wherein in the extraction process, the processor extracts, from among the two or more words, words that are in a predetermined order relationship with a word to which the label corresponding to the registered word is assigned and that are assigned the label other than the exclusion label.
6. 5. The testing support device according to claim 1, wherein when the text information contains a plurality of the registered words, the processor extracts, in the extraction process, the phrase corresponding to the registered word from the text information for each of the registered words, and in the storage process, stores the test result information based on the phrase extracted for each of the registered words in association with the test location.
7. 7. The inspection support device of claim 6, wherein when an inspection is performed from two or more directions on one inspection location and the text information includes two or more registered words indicating the directions in which the inspection is performed, the processor extracts, in the extraction process, the words corresponding to the directions from the text information for each direction, and, in the storage process, stores the inspection result information based on the words extracted for each direction in association with the inspection location.
8. 2. The inspection support device according to claim 1, wherein the processor performs the reception process, the text conversion process, the extraction process, and the storage process for a specific inspection location where inspection is performed at two or more inspection positions among the multiple inspection locations.
9. 9. The inspection support device according to claim 8, wherein when a cancel word is included in the text information obtained by the text conversion process, the processor does not execute the storage process for the word extracted in the most recent extraction process, or deletes the word that has already been stored in the most recent storage process.
10. 10. The inspection support device of claim 9, wherein when the storage process of the word is not performed or the deletion is performed in response to the cancellation word for a specified inspection position, the processor executes the reception process, the text conversion process, the extraction process, and the storage process again for the inspection position.
11. The test support device according to claim 1 , wherein the processor executes a process of switching the test location to be subjected to the reception process when a specific word is included in the text information obtained by the text conversion process.
12. The examination support device according to claim 1 , wherein the processor performs voice output for the one or more words extracted by the extraction process.
13. An inspection support method used for inspecting a plurality of inspection points at a building construction site or within a constructed building, comprising: receiving, by a processor, an input of a voice uttered by an inspector; recognizing the speech with a processor and converting the speech into text; When a registered word is included in text information obtained by converting the speech into text, a processor extracts a phrase corresponding to the registered word from the text information; and storing, by a processor, test result information based on the extracted words and phrases in association with the test location, for each test location.
Citation Information
Patent Citations
Information processing system, terminal device, server, information processing method and program
JP2017084050A