JP7220134B2 - Facility construction inspection device, facility construction inspection method, and program - Google Patents

Description

Application of Article 30, Paragraph 2 of the Patent Act [1] [Forum] Dates October 25 to 26, 2018 Name of meeting ``Tsukuba Forum 2018'' <Publication> Explanatory material for ``AI in construction quality inspection'' [2] [Forum] Date: October 25, 2018 Meeting name: "MM Forum 21" <Publication> Lecture information on MM Forum 21 [3] [Forum] Date: November 21-22, 2018 Meeting Name "NTT COMWARE'S DAY 2018" <Publication> Explanatory materials for "Automation of photo inspection work with AI"・Meister's Cup 2018 in Fukuoka" <Publication> Explanatory materials for "AI measures in EG work (photo inspection work)" [5] [Forum] Date April 3-5, 2019 Meeting name 3rd AI/Artificial Intelligence EXPO” <Publication> Explanatory materials for “Maintenance Solutions for Infrastructure Facilities” [6] [Industry-Government Council] Date March 5, 2019 Name of Meeting Next-Generation Infrastructure 'Meeting (3rd)' <Publication> Agenda/Explanatory material for 'Smart Maintenance Efforts in the NTT Group' Name “Unstructured Data Processing PoC Final Report” <Publication> “Unstructured Data Processing PF Final Report” Explanatory Material [8] [Report Meeting/Discussion Meeting] Date June 5, 2019 “Discussion meeting on unstructured data processing PoC (AS Labs)” <Publication> Explanatory material for “AI photo inspection efforts” [9] [Report meeting / discussion meeting] Date August 7, 2019 Meeting Name ``Discussion meeting on unstructured data processing PoC (NTT EAST Equipment Planning Department)'' <Publication> Explanatory material for ``Access S department's smart maintenance efforts'' [10] [Report meeting/opinion exchange meeting] September 13, 2019 Name of Meeting: “Discussion Meeting on Unstructured Data Processing PoC (Docomo CS)” <Publication> “Smart Maintenance Initiatives” Explanatory material

Application of Article 30, Paragraph 2 of the Patent Law [11] to [14] [Report meeting/Opinion exchange meeting] July 18, 2019 Name of the meeting: “First Study Group on the Use of AI in Photographic Inspections”, “Demonstration of AI Utilization in Photographic Inspections” <Publication> “Discussion on Photographic Inspection AI”

The present invention relates to an inspection device for facility construction, an inspection method for facility construction, and a program.

2. Description of the Related Art Conventionally, techniques for detecting the state of utility poles are known. Techniques of this type are described, for example, in Patent Documents 1 to 3 below. The utility pole position inspection device described in Patent Document 1 includes a utility pole detection unit that detects a utility pole from an image taken by a camera, and determines whether or not the utility pole is outside the construction gauge based on the image taken by the camera. inspect whether In the inspection system described in Patent Document 2, an information terminal identifies the utility pole number of a utility pole that is about to pass by using the image of the 360-degree camera and the GPS function during patrol, and is installed on the utility pole. The equipment to be inspected is automatically identified from the image, and the presence or absence of abnormality is determined by the abnormality detection sensor. The support inclination abnormality determination device described in Patent Document 3 creates three-dimensional position information of a predetermined position of a support such as a utility pole, compares the three-dimensional position information over time, and uses the amount of change as a judgment value. If the determination value is equal to or greater than a predetermined value, it is determined that the support is tilted abnormally.

Furthermore, conventionally, there is known a technique of performing an inspection or the like based on an image captured at a construction site or the like. Techniques of this type are described, for example, in Patent Documents 4 to 6 below. The construction site photography system described in Patent Document 4 captures a construction site photograph showing the workmanship of the work with a mobile terminal when the work of each process is completed, and the mobile terminal transmits the construction site image data and the process It is transmitted to the server in association with work data indicating the result of each work. The construction status inspection system described in Patent Literature 5 specifies which construction inspection image is sent when image data indicating an inspection image of the construction status of a construction work is transmitted. The piping construction management system described in Patent Document 6 includes an inspection information input unit that inputs inspection information corresponding to inspection items, and an evaluation unit that evaluates whether or not jointing of joints has been properly performed based on the inspection information. and transmits the construction management information to the outside only when it is evaluated as appropriate. The piping construction management system further includes an image processing unit that evaluates whether or not jointing of the joints has been properly performed based on image data of the joints photographed by the imaging unit.

JP 2015-116916 A JP 2019-032684 A JP 2018-087798 A JP 2017-091495 A JP 2014-197387 A JP 2012-123589 A

However, the technology for detecting the state of utility poles described in Patent Documents 1 to 3 described above detects the state of utility poles whose construction has already been completed, so there is a problem that the construction state of utility poles cannot be inspected. There is In addition, although the techniques described in Patent Documents 4 and 5 mentioned above transmit image data, there is a problem that manpower is required for construction inspection. Furthermore, the technique described in Patent Literature 6 may not be able to evaluate with high accuracy when applied to inspection of a process involving a large number of members.

The present invention has been made in view of the above problems, and provides an inspection device for facility construction, an inspection method for facility construction, and a program that can inspect with high accuracy the state of construction using a plurality of members. It is intended to

(1) One aspect of the present invention is an acquisition unit that acquires an inspection image obtained by imaging a support member of a utility pole and inspection item information indicating that an inspection item is the support member of the utility pole, from a terminal device; a certainty factor indicating the likelihood of the supporting member and the inspection object around the supporting member is calculated from the above, and based on the certainty factor, it is determined whether or not the supporting member and the inspection object are included in the inspection image. a recognizing unit for judging, and when the inspection image includes the supporting member and the inspection object, performing an inspection of equipment construction using the supporting member based on the positional relationship between the supporting member and the inspection object. and a notification unit for notifying the terminal device of the result of the inspection performed by the inspection unit.

(2) An aspect of the present invention is the inspection apparatus for facility construction described above, wherein the recognition unit is a recognition model learned using a learning image including a normal inspection object, A recognition model whose processing parameters have been learned may be used so that when an inspection image including an object is input, it is recognized that the object to be inspected exists.

(3) An aspect of the present invention is the inspection apparatus for facility construction, wherein the recognition unit recognizes, as the inspection object, a virtual cross section of a utility pole supported by the support member, and The inspection unit may perform the inspection based on the positional relationship between the position of the support member and a virtual point based on the area occupied by the virtual cross section.

(4) An aspect of the present invention is the above equipment construction inspection device, wherein the recognition unit recognizes an arrow indicating a pulling direction of the utility pole from the inspection image as the inspection object, and The inspection unit may perform the inspection based on the relationship between the position of the support member, the virtual point based on the area occupied by the virtual cross section, and the direction indicated by the arrow.

(5) An aspect of the present invention is the inspection apparatus for facility construction, wherein the inspection unit determines the center-of-gravity position of the support member and the center-of-gravity position of the utility pole based on the area occupied by the virtual cross section. , and the direction indicated by the arrow.

(6) An aspect of the present invention is the inspection apparatus for facility construction, wherein the inspection unit, when the utility pole is of the first type, detects the difference between the longitudinal direction of the support member and the direction of the arrow. determining whether the angle is within a first angle, and if the utility pole is of the second type, whether the angle between the longitudinal direction of the support member and the direction of the arrow is within a second angle; can be determined.

(7) An aspect of the present invention is a step of obtaining from a terminal device an inspection image obtained by imaging a support member of a utility pole and inspection item information indicating that the inspection item is the support member of the utility pole; calculating a certainty factor indicating the likelihood of the supporting member and the inspection object around the supporting member; and based on the certainty factor, whether or not the supporting member and the inspection object are included in the inspection image. and, when the inspection image includes the support member and the inspection object, performing an inspection of equipment construction using the support member based on the positional relationship between the support member and the inspection object. and notifying the terminal device of the result of the inspection.

(8) An aspect of the present invention is a step of obtaining, from a terminal device, an inspection image of a support member of a utility pole and inspection item information indicating that the inspection item is the support member of a utility pole, to a computer; calculating a certainty factor indicating the likelihood of the support member and the inspection object around the support member from the inspection image; and including the support member and the inspection object in the inspection image based on the certainty factor. a step of determining whether or not the support member and the inspection object are included in the inspection image, based on the positional relationship between the support member and the inspection object, the equipment construction using the support member A program causing execution of a step of performing an inspection and a step of notifying the terminal device of the result of the inspection.

ADVANTAGE OF THE INVENTION According to 1 aspect of this invention, the construction state using several members can be inspected|inspected with high precision.

BRIEF DESCRIPTION OF THE DRAWINGS It is a block diagram which shows one structural example of the equipment construction inspection system of embodiment. 4 is a block diagram showing an example of an inspection object recognition unit; FIG. It is a figure which shows an example of the construction location in construction of a utility pole. It is a figure which shows an example of the operation screen for construction inspection in a user terminal device. FIG. 4 is a diagram showing an example of a guide image; FIG. FIG. 10 is a diagram showing another example of a guide image; It is a flow chart which shows an example of inspection processing of equipment construction in a construction inspection server device. FIG. 10 is a diagram showing the relationship between an inspection image, construction members a to e included in the inspection image, and reliability. FIG. 10 is a diagram showing an example of an inspection image including metal fittings for segmentation; It is a figure which shows the construction member, tag name, and recognition conditions in the installation work of the metal fittings for a branch line. It is a figure which shows an example of the partial area|region in the hardware for a dividing line. FIG. 10 is a diagram showing an example of an inspection image including an anchorage; It is a figure which shows the construction member, the tag name, and recognition conditions in the installation work of an anchor. FIG. 10 is a diagram showing a process of detecting the height of a horizontal bar or horizontal measure in inspection of installation work of an embankment; It is a figure which shows the positional relationship of the utility pole cross section, the shackle, and the drawing board which shows a tension|tensile_strength direction. FIG. 10 is a diagram showing an example of an inspection image including a lower portion of a branch line; FIG. 10 is a diagram showing construction members, tag names, and recognition conditions in installation work for the lower part of the branch line; FIG. 10 is a diagram showing a process of calculating the burying depth of a branch line in inspection of the lower part of the branch line; FIG. 10 is a diagram for explaining an example of calculating an embedding depth under a branch line; It is a figure which shows an example of the inspection image containing a utility pole lower part. FIG. 4 is a diagram showing construction members, tag names, and recognition conditions in the construction of the lower part of a utility pole; FIG. 10 is a diagram for explaining the process of inspecting the height of scaffolding; FIG. 10 is a diagram for explaining a process of inspecting the height of a sign plate of a utility pole; FIG. 4 is a diagram showing an example of an inspection image including insulators; FIG. 4 is a diagram showing construction members, tag names, and recognition conditions in insulator installation work; It is a figure for demonstrating the process which inspects the height of an insulator. FIG. 4 is a diagram showing an example of a photographed image including branch lines and a protractor for measuring branch line angles; FIG. 4 is a diagram showing member names, tag names, and recognition conditions of a protractor; FIG. 10 is a diagram for explaining processing for inspecting branch line angles; FIG. 10 is a diagram showing an example of a photographed image including a wrapped grip; FIG. 4 is a diagram showing member names, tag names, and recognition conditions in installation work of the wrapped grip. It is a figure which shows an example of the test|inspection image containing clogs. FIG. 10 is a diagram showing member names, tag names, and recognition conditions in the installation work of clogs. FIG. 10 shows an example of an inspection image including spur guards and insulators; FIG. 10 is a diagram showing member names, tag names, and recognition conditions in branch line protection and insulator installation work; FIG. 4 is a diagram showing an example of an inspection image including insulators; It is a figure which shows the member name, tag name, and recognition conditions in the installation work of an insulator. FIG. 10 is a diagram showing an example of an inspection image including fasteners; It is a figure which shows the member name, tag name, and recognition conditions in installation work of a fastener.

An inspection device for facility construction, an inspection method for facility construction, and a program to which the present invention is applied will be described below with reference to the drawings.

The facility construction inspection system according to the embodiment uses images captured by site workers for utility pole facility construction to inspect the facility construction. The facility construction inspection system of the embodiment determines whether an image captured by a field worker is suitable for inspection, whether or not a construction member to be inspected exists in the image, and whether or not the construction member is appropriate. By determining , the construction state using a plurality of construction members can be inspected with high accuracy. In addition, the facility construction inspection system of the embodiment presents the determination results (image determination results, construction member determination results, and inspection results) to the site worker, thereby reducing the time and effort of the site worker to take images and It is possible to reduce the trouble of inspection by a construction inspector and the trouble of re-construction when there is a defect in the equipment construction.

FIG. 1 is a block diagram showing a configuration example of a facility construction inspection system according to an embodiment. The facility construction inspection system includes, for example, one or more user terminal devices 100, a construction inspection server device 200, and a management terminal 300. The user terminal device 100, the construction inspection server device 200, and the management terminal 300 are connected to the communication network NW, for example. Each device connected to the communication network NW has a communication interface such as a NIC (Network Interface Card) or a wireless communication module (not shown in FIG. 1). The communication network NW includes, for example, the Internet, WAN (Wide Area Network), LAN (Local Area Network), cellular network, and the like.

The user terminal device 100 is, for example, a mobile terminal device equipped with a camera device such as a smart phone or a tablet terminal. The user terminal device 100 starts a UA (User Agent) such as a browser or an application program. UA is, for example, an application for communicating with the construction inspection server device 200 . The user terminal device 100 uses a construction inspection application as a UA, and performs display processing, operation reception processing, and the like using the content received from the construction inspection server device 200 . In addition, the user terminal device 100 provides the construction inspection server device 200 with operation information based on the user's operation, captured images, and the like using the construction inspection application.

The construction inspection server device 200 is an information processing device that receives inspection requests for facility construction work and provides inspection results for the facility construction work. The construction inspection server device 200 includes, for example, a cooperative API (Application Programming Interface) section 210 and an inspection target recognition section 220 . Functional units such as the cooperative API unit 210 and the inspection target recognition unit 220 are realized by executing a program stored in a program memory by a processor such as a CPU (Central Processing Unit). Some or all of these functional units may be realized by hardware such as LSI (Large Scale Integration), ASIC (Application Specific Integrated Circuit), or FPGA (Field-Programmable Gate Array), It may be realized by cooperation of software and hardware.

The cooperative API unit 210 includes an API unit 212, an image processing unit 214, an inspection unit 216, and a control unit 218, for example. The API unit 212 performs processing for receiving various types of information from the user terminal device 100, processing for providing content to the user terminal device 100, and the like. The image processing unit 214 performs predetermined image processing on image data acquired from the user terminal device 100 . An image used for inspection is hereinafter referred to as an inspection image. The inspection unit 216 uses the inspection image acquired by the API unit 212 and/or the inspection image processed by the image processing unit 214 to perform inspection processing for facility construction. The control unit 218 performs control such as activation of the image recognition engine in the inspection object recognition unit 220 .

The inspection target recognition unit 220 includes, for example, a plurality of image recognition engines 222A, 222B, . . . 222N (N is a natural number). In addition, when collectively referring to a plurality of image recognition engines, they are simply referred to as the image recognition engine 222 . Each image recognition engine 222 corresponds to each inspection item of the plurality of inspection items. The image recognition engine 222 performs recognition processing for recognizing construction members as inspection targets in inspection items.

FIG. 2 is a block diagram showing an example of an inspection object recognition unit. The inspection object recognition unit 220 includes, for example, a learning image database 230, a learning processing unit 232, a recognition processing unit 234, and a result output unit 236.

The learning image database 230 is a database that stores, for example, images acquired from the management terminal 300 as learning images. The learning image database 230 stores, for example, positive example images as learning images. A positive image is an image of a target object in a normal state. A positive example image is, for example, an image of a construction member used in construction of a utility pole (a utility pole, an anchor, a metal fitting for a branch line, a tape measure, etc.). A positive example image is associated with tag information including a construction member name. Information indicating, for example, a process name and an inspection item name may be associated with the positive example image. Note that the learning image database 230 may store not only positive images but also images of objects different from construction members (abnormal objects and defective products) as learning images.

A learning image is supplied from the learning image database 230 to the learning processing unit 232 . The learning processing unit 232 performs recognition processing using the learning image and obtains a determination result. The learning processing unit 232 learns the processing parameters of the recognition processing unit 234 (recognition model 234A) so that the result of determination is that the construction member is detected, and generates learning result data. The processing parameter of the recognition processing unit 234 (recognition model 234A) is, for example, a filter (also called weight or bias) included in the neural network. The learning result data is stored in the learning result data storage unit 2341 .

The recognition processing unit 234 has, for example, a learning result data storage unit 2341 and a determination unit 2342 . The learning result data storage unit 2341 accumulates processing parameters of the recognition processing unit 234 as learning results, and the processing parameters of the recognition processing unit 234 are updated by the learning processing unit 232 . When the inspection image is acquired, the recognition processing unit 234 extracts feature amounts using a recognition model 234A based on the processing parameters of the recognition processing unit 234 . The determination unit 2342 detects a construction member based on the extracted feature amount. The recognition processing unit 234 outputs the determination result by the determination unit 2342 as the recognition result.

In the following, inspection processing in the construction of utility poles in the facility construction inspection system described above will be described.
FIG. 3 is a diagram showing an example of a construction site in construction of a utility pole. Equipment work to be inspected in utility pole construction includes, for example, branching work using branching hardware (A1), utility pole burial work (A2), branch line installation work (A3), and branch line installation work. Includes insulator installation work (A4), protective cover installation work for branch lines (A5), branch line burying work (A6), and utility pole sign plate installation work (A7). Note that A1 to A7 in FIG. 3 indicate areas captured for inspection of each construction work.

When requesting inspection of facility construction, the user terminal device 100 activates a construction inspection application and displays an operation screen for facility inspection. FIG. 4 is a diagram showing an example of a construction inspection operation screen on the user terminal device. As shown in FIG. 4A, the construction inspection operation screen includes icon images corresponding to inspection items in facility construction. The user terminal device 100 activates the camera device when the icon image is selected based on the user's operation. When the icon image of "root" is selected, the user terminal device 100 displays a screen including an image captured by the camera device and a guidance message designating the construction member to be captured, as shown in FIG. 4B. display.

FIG. 5 is a diagram showing an example of a guide image. The user terminal device 100 superimposes a guide image on the image 110 captured by the camera device using the construction inspection application. The guide image is an image that assists the user terminal device 100 to photograph the construction member so as to include it within a predetermined range in the image. A guide image includes, for example, a guide frame image 120 and a mask image 122 . The guide frame image 120 indicates an area that preferably contains the construction member to be inspected. The mask image 122 indicates an area that the construction inspection server device 200 does not recognize as an area including construction members.

The area indicated by the guide frame image 120 and the area indicated by the mask image 122 are desirably determined based on the learning images. For example, if many of the positive example images (learning images) obtained by picking up a sign plate of a utility pole include the sign plate of the utility pole in the center of the image, the guide frame image 120 is a substantially central region in the image in the horizontal direction. It is desirable that the mask image 122 indicates an area other than the area indicated by the guide frame image 120 . Thereby, the construction inspection server device 200 can inspect with high accuracy. Note that the guide frame image 120 may be changeable by a setting file. As a result, the user terminal device 100 can flexibly change the region indicated by the guide frame image 120 based on the settings of the camera device and the display device, and can capture an image that appropriately includes the construction member.

FIG. 6 is a diagram showing another example of the guide image. The user terminal device 100 may acquire an image including a circular area (operation area) centered on the user's operation position from among the captured images, and transmit the image to the construction inspection server apparatus 200 . Also, the user terminal device 100 may acquire an image including an operation area where the finger is slid along the elongated branch line protection cover. The user terminal device 100 may generate coordinate information indicating that the operation area of the user is an image area including the construction member. The user terminal device 100 can transmit the captured image and coordinate information to the construction inspection server device 200 . For example, when an insulator and a branch line protection cover are included in a photographed image, the user terminal device 100, based on the user's operation position, outputs coordinate information of the image area including the insulator, coordinates of the image area including the branch line protection cover, and so on. Generate information.

In this way, the user terminal device 100 generates a captured image corresponding to the inspection item name based on the user's operation. As shown in FIG. 3 , in the segmentation work using segmentation hardware, the user terminal device 100 transmits an image of an area A1 including the segmentation hardware to the construction inspection server device 200 . In the utility pole burial work, the user terminal device 100 transmits to the construction inspection server device 200 an image of an area A2 including the utility pole and the support member. In installation work of the branch line, the user terminal device 100 transmits an image of the area A3 including the protractor to the construction inspection server device 200 . In the insulator installation work on the branch line, the user terminal device 100 transmits an image of an area A4 including the branch line and the insulator to the construction inspection server device 200 . In the protective cover installation work on the branch line, the user terminal device 100 transmits an image of the area A5 including the protective cover to the construction inspection server device 200 . In the installation work of the branch line, the user terminal device 100 transmits to the construction inspection server device 200 an image of an area A6 including the buried location. In the installation work of a utility pole signboard, the user terminal device 100 transmits an image of an area A7 including the utility pole signboard to the construction inspection server device 200 .

FIG. 7 is a flowchart showing an example of inspection processing for facility construction in the construction inspection server device. First, the cooperative API unit 210 acquires inspection item names in facility construction based on information received from the user terminal device 100 (step S100). The cooperative API unit 210 acquires the activation list of the inspection target recognition unit 220 at the time when the inspection item name is acquired. The cooperative API unit 210 compares the inspection item name acquired in step S100 with the activation list (step S104), and determines whether or not the image recognition engine 222 corresponding to the acquired inspection item name is activated (step S106). If the image recognition engine 222 corresponding to the inspection item name has not been activated (step S106: NO), the cooperative API unit 210 controls to activate the image recognition engine 222 corresponding to the inspection item name (step S108).

If the image recognition engine 222 corresponding to the inspection item name is activated (step S106: YES), or after the image recognition engine 222 corresponding to the inspection item name is activated in step S108, the cooperative API unit 210 extracts the image file And an inspection information file is obtained (step S110). Next, the cooperative API unit 210 performs predetermined image processing on the inspection image included in the acquired image file (step S112). The predetermined image processing includes, for example, edging (sharpening) processing and brightness correction processing. The user terminal device 100 transfers the inspection image that has undergone predetermined image processing to the image recognition engine 222 corresponding to the inspection item name.

The image recognition engine 222 performs recognition processing for determining whether or not there is a construction member (inspection target) corresponding to the inspection item name based on the image acquired from the cooperative API unit 210 (step S114). At this time, the image recognition engine 222 calculates a certainty factor for each construction member. In addition, the image recognition engine 222 calculates the reliability of the construction member for each image processing. The image recognition engine 222 determines that the construction member exists when the reliability of the construction member is equal to or greater than a predetermined value. When the image recognition engine 222 recognizes that there is a construction member (step S116: YES), the process proceeds to step S120. When the image recognition engine 222 recognizes that there is no construction member (step S116: NO), the image recognition engine 222 acquires the inspection image again (step S118). When reacquiring the inspection image, the construction inspection server device 200 transmits a request to recapture the image to the user terminal device 100 .

The cooperative API unit 210 compares the reliability of the construction member calculated by the image recognition engine 222 (step S120). FIG. 8 is a diagram showing the relationship between the inspection image, the construction members a to e included in the inspection image, and the degree of certainty. When members a to e are included in the inspection image, the linked API unit 210 calculates the confidence levels Ca1 to Ce1 for each of the members a to e included in the inspection image, and calculates the total value C1 of the confidence levels Ca1 to Ce1. . The cooperative API unit 210 also calculates the confidence factors Ca2 to Ce2 for each of the members a to e, and calculates the total value C2 of the confidence factors Ca2 to Ce2 for the image obtained by subjecting the inspection image to edge processing. The cooperative API unit 210 also calculates the confidence factors Ca3 to Ce3 for each of the members a to e, and calculates the total value C3 of the confidence factors Ca3 to Ce3 for the images obtained by subjecting the inspection image to the brightness correction process.

The cooperative API unit 210 performs inspection item determination processing using the image corresponding to the highest total value among the plurality of total values (step S122). Note that the inspection item determination process will be described later. The cooperative API unit 210 transmits the inspection result to the user terminal device 100 as a result of inspection item determination processing (step S124).

The inspection process will be described below.
(Installation work of metal fittings for branch lines)
FIG. 9 is an example of an inspection image including metal fittings for segmentation, and FIG. 10 is a diagram showing construction members, tag names, and recognition conditions in installation work of metal fittings for segmentation.
The cooperative API unit 210 acquires an inspection image including a segmenting hardware (item number 1), a detachment stopper (item number 2), and two fasteners (item number 3) as construction members. The image recognition engine 222 corresponding to the segmenting metal fitting as the inspection item name determines whether or not the segmenting metal fitting, the detachment stopper, and the fastener exist in the inspection image as construction members. The image recognition engine 222 determines that the inspection image includes a parting metal fitting when the construction member recognized as the parting metal fitting is included in the inspection image. is provided with tag information corresponding to the metal fittings for the segment line. The image recognition engine 222 determines that a detachment stopper exists in the inspection image when the degree of certainty of the object recognized as the detachment stopper in the inspection image is equal to or greater than a predetermined value, and determines that the detachment stopper is present in the inspection image. Tag information corresponding to the detachment stopper is added to the containing area. The image recognition engine 222 determines that a fastener exists in the inspection image of the metal fitting for segmentation when the certainty factor of the object recognized as the fastener in the inspection image is equal to or higher than a predetermined value, and determines that the fastener exists in the inspection image. Tag information corresponding to the fastener is added to the area containing the .

The tag information includes tag names as recognition results of the image recognition engine 222 . A tag name is information indicating a construction member. The tag information may include coordinate information within the inspection image in addition to the tag name. The process of adding tag information includes, for example, the process of associating tags with regions in the inspection image. That is, the image recognition engine 222 may generate information representing the correspondence between the coordinates representing the regions in the inspection image and the tag information, and creates a table representing the correspondence between the coordinates representing the regions in the inspection image and the tag information. You may update.

Tag information includes at least a tag name. The tag information may include coordinate information indicating the position in the image of the recognized construction member. That is, the tag information may include in-image position information of the construction member recognized by the recognition process. The in-image position information may be information indicating the area (coordinate values, highest point, lowest point, shape, etc.) occupied by the construction member in the image.

FIG. 11 is a diagram showing an example of a partial region of the metal fitting for the segment line. The cooperative API unit 210 rotates the inspection image so that the straight line portion of the segmenting hardware approaches the horizontal direction in the image in the inspection of the segmenting hardware. The cooperative API unit 210 extracts a region including metal fittings for segmentation as shown in FIG. 11 . The cooperative API unit 210 may rotate the inspection image after recognizing the presence of the segmentation hardware. The cooperative API section 210 divides the area including the metal fittings for segmentation into eight partial areas (1) to (8). The linking API unit 210 has the split line fittings in partial areas (1) to (8), the detachment stopper in partial area (3), and one of the two fasteners in the partial area. Recognize that one of the two fasteners is present over areas (4), (5) and in sub-area (7).

As described above, the construction inspection server device 200 can be used in the installation work of the metal fittings for the split line in the condition 1: that the metal fittings for the split line, the stopper, and the plurality of fasteners are present, and the condition 2: confidence of the stopper and condition 3: a specified number of areas where multiple fasteners exist in the specified part of the metal fitting for segmentation. Determine that there is.

The user terminal device 100 may display a screen for inputting the total number of fasteners installed, and may transmit information indicating the total number of installed fasteners to the installation inspection server device 200 based on the user's operation. The construction inspection server device 200 determines that the inspection result is acceptable on the condition that the total number of fasteners based on the information received from the user terminal device 100 and the total number of fasteners recognized by the image recognition engine 222 match. judge. Thereby, the accuracy of the inspection result can be further improved. The facility construction inspection system compares the number of attachments based on the user's operation with the number of attachments recognized by the image recognition engine 222, not only for the number of fasteners, but also for the number of attachments of other construction members. may

(Installation work of root hooks)
FIG. 12 is an example of an inspection image including a shackle, and FIG. 13 is a diagram showing construction members, tag names, and recognition conditions in the shackle installation work.
The cooperative API unit 210 acquires an inspection image. The image recognition engine 222 corresponding to the skein as the inspection item name, puts a measure, a vertical measure outer shell, a horizontal measure outer shell, a horizontal bar outer shell, an anchor, a utility pole section, a utility pole, and a tension in the inspection image as construction members. Determines whether a drawing board showing directions is included. The image recognition engine 222 determines whether one measure exists as a specific number of installation members with a measure representing each centimeter. The image recognition engine 222 calculates a certainty factor for each construction member. The image recognition engine 222 determines that the construction member exists in the inspection image when the reliability of the construction member exceeds a predetermined value, and adds tag information corresponding to the construction member to an area including the construction member in the inspection image. Give. Although FIG. 12 shows an example in which a horizontal bar exists instead of a horizontal measure, the image recognition engine 222 recognizes the existence of at least one of the horizontal measure and the horizontal bar as a depth reference member. and

FIG. 14 is a diagram showing processing for detecting the height of a horizontal bar or a horizontal measure in an inspection of an installation work of an embankment. The cooperative API unit 210 detects the height of the horizontal bar or horizontal measure to measure the buried depth of the utility pole. As prerequisites for detecting the height of the horizontal bar or horizontal measure, the image recognition engine 222 detects (condition 1) all numerical values displayed as tag information corresponding to the measure, and (condition 2) the outline of the horizontal measure. It is assumed that the tag information of either one of the outline of the horizontal bar and the horizontal bar has been detected.

As shown in FIG. 14, the vertical measure may be imaged tilted with respect to the horizontal direction (X direction) in the image. The cooperative API unit 210 calculates proportionally the distance a from the intersection coordinate position of the horizontal bar and the measure to the upper end of the vertical measure region and the distance b from the intersection coordinate position of the horizontal bar and the measure to the lower end of the vertical measure region. to calculate the height of the horizontal bar.

FIG. 15 is a diagram showing the positional relationship between the cross section of the utility pole, the shackle, and the drawing board indicating the tension direction. The cooperative API unit 210 inspects facility construction using the anchor (inspection unit) based on the positional relationship between the anchor (supporting member), the utility pole, and the drawing board indicating the direction of tension. Specifically, the linkage API unit 210 determines whether or not the center of gravity of the cross section of the utility pole and the center of gravity of the skein are arranged in the order indicated by the drawing board indicating the direction of tension. The cooperative API unit 210 judges that the inspection result is acceptable when the shackle is constructed correctly so as to resist the tension applied to the utility pole, and otherwise determines that the inspection result of the shackle installation work is unacceptable. judge.

Furthermore, the cooperation API unit 210 determines the direction of the arrow. The cooperative API unit 210 rotates the inspection image so that the longitudinal direction of the shackle is parallel to the horizontal direction in the image when the shackle and the drawing board indicating the tension direction are detected. The cooperative API unit 210 calculates the angle between the longitudinal direction of the skein and the direction of the arrow indicated by the drawing board indicating the tension direction. If the calculated angle is within a predetermined range (for example, 80 degrees to 100 degrees), the cooperative API unit 210 determines that the inspection result of the installation work of the shackle is acceptable. The inspection result of the installation work is determined to be unsatisfactory.

(Installation work for the lower part of the branch line)
FIG. 16 is an example of an inspection image including the lower part of the branch line, and FIG. 17 is a diagram showing construction members, tag names, and recognition conditions in installation work of the lower part of the branch line.
The cooperative API unit 210 acquires an inspection image. The image recognition engine 222, which corresponds to the lower part of the branch line as the inspection item name, adds a vertical measure, a horizontal measure, a centimeter measure indicating each height, and an anchor color (for example, black, red, brown, and orange) to the inspection image as construction members. color) is included. FIG. 16 shows a case where an anchor color corresponding to any item number among item numbers 3 to 6 corresponding to construction members is recognized. The image recognition engine 222 calculates a certainty factor for each construction member. The image recognition engine 222 determines that the construction member exists in the inspection image when the reliability of the construction member exceeds a predetermined value, and adds tag information corresponding to the construction member to an area including the construction member in the inspection image. Give.

FIG. 18 is a diagram showing processing for calculating the buried depth of the branch line in the inspection of the lower portion of the branch line. The cooperative API unit 210 calculates an intersection point between a region tagged with a vertical measure tag and a region tagged with a horizontal measure tag. The cooperative API unit 210 calculates the position of the calculated intersection point in the Y direction as the embedding depth of the lower part of the branch line. The cooperative API unit 210 calculates the position of the intersection point in the Y direction in the image based on the tag information of the area including the intersection point. The cooperative API unit 210 determines that the inspection result of the lower part of the branch line is acceptable if the buried depth of the lower part of the branch line is appropriate, and that the inspection result of the lower part of the branch line is unacceptable if not.

FIG. 19 is a diagram for explaining an example of calculating the embedding depth of the lower part of the branch line. When calculating the burying depth of the lower part of the branch line, the cooperative API unit 210 first obtains an area where the vertical major area and the horizontal major area intersect, as shown in FIG. 19(A). Next, the cooperative API unit 210 acquires the position where the scale is 0 in the region of the vertical measure as the minimum value of the embedding depth in the Y direction in the image. Next, the cooperative API unit 210 acquires the minimum value of the intersecting area in the Y direction in the image as the maximum value of the embedding depth in the Y direction. Next, the linked API unit 210 calculates the intra-image distance between the minimum and maximum embedding depths. Next, as shown in FIG. 19B, the cooperative API unit 210 determines the distance (a pixel) in the image from the position indicating 10 cm to the position indicating 20 cm in the vertical measure area and the minimum embedding depth. From the ratio of the in-image distance (b pixels) between the value and the maximum value, the actual length indicated by the in-image distance between the minimum and maximum values of the embedding depth is calculated. That is, the cooperative API unit 210 calculates the embedding depth of the branch line by obtaining the value of x in 10 (cm):x (cm)=a (pixel):b (pixel).

In addition, when the upper branch line or the soil classification is selected as the inspection item name, the linked API unit 210 selects black, red, brown, and orange with the highest degree of certainty when multiple anchor color tags are detected. Detect colors. The cooperative API unit 210 may compare the detected color with the anchor color of the lower branch line based on the ground conditions of the facility location, economic efficiency, etc., and determine whether or not the anchor color is appropriate.

(Construction at the bottom of utility poles)
FIG. 20 is an example of an inspection image including a lower portion of a utility pole, and FIG. 21 is a diagram showing construction members, tag names, and recognition conditions in construction of the lower portion of a utility pole.
The cooperative API unit 210 acquires an inspection image. The image recognition engine 222 corresponding to the lower part of the utility pole as the inspection item name determines whether or not the inspection image includes a scaffolding, a utility pole sign plate, and a measure as construction members. Among the members, the confidence of the sign board of the utility pole and the confidence of the measure are calculated. When each reliability exceeds a predetermined value, the image recognition engine 222 determines that the construction member exists in the inspection image, and adds tag information corresponding to the construction member to an area including the construction member in the inspection image. . The image recognition engine 222 adopts the scaffold having the lowest installation position among the construction members recognized as scaffolds.

FIG. 22 is a diagram for explaining the process of inspecting the height of scaffolding. The cooperative API unit 210 recognizes the measure as the distance from the lowest point of the measure recognized by the image recognition engine 222 to the highest point of the measure, and detects tag information indicating the lowest scaffolding. Based on the tag information of the detected scaffolding, the cooperative API unit 210 determines that the scaffolding height inspection result is acceptable if the lowest scaffolding is higher than a certain height from the measure, or otherwise. If so, the inspection result is judged to be unsatisfactory.

FIG. 23 is a diagram for explaining the process of inspecting the height of the sign board of the utility pole. The cooperative API unit 210 recognizes the height of the measure as the distance from the lowest point of the measure recognized by the image recognition engine 222 to the highest point of the measure, and detects tag information indicating the sign plate of the utility pole. Based on the tag information of the sign board of the utility pole, the cooperative API unit 210 determines that the inspection result of the height of the sign board of the utility pole is acceptable if the height of the top end of the sign board of the utility pole is lower than a certain height. otherwise, the test result is determined to be unsatisfactory. For example, the cooperative API unit 210 may determine that the inspection result of the height of the signboard of the utility pole is acceptable when the heights of the top end and the bottom end of the signboard of the utility pole meet a certain height.

(Insulator installation work)
FIG. 24 is an example of an inspection image including insulators, and FIG. 25 is a diagram showing construction members, tag names, and recognition conditions in insulator installation work.
The cooperative API unit 210 acquires an inspection image. The image recognition engine 222 corresponding to the insulator as the inspection item name determines whether or not the inspection image includes the insulator and the measure as construction members. The image recognition engine 222 calculates confidences for insulators and measures. The image recognition engine 222 determines that an insulator and a measure are present in the inspection image when each certainty factor exceeds a predetermined value, and adds tag information corresponding to the construction member to an area including the construction member in the inspection image. do. The image recognition engine 222 adopts insulators contained in a predetermined area including the position specified by the user. The image recognition engine 222 adopts the insulator when there is a construction member recognized as an insulator in a predetermined area including the detection position of the user operation on the display section of the user terminal device 100 .

FIG. 26 is a diagram for explaining the process of inspecting the height of insulators. The cooperative API unit 210 recognizes the height of the measure as the distance from the lowest point of the measure recognized by the image recognition engine 222 to the highest point of the measure, and detects tag information indicating an insulator. Based on the tag information of the detected insulator, the cooperative API unit 210 determines that the inspection result of the height of the insulator is acceptable if the insulator is higher than a certain height. Judged as unacceptable.

(Installation work of branch lines)
FIG. 27 is an example of a photographed image including a branch line and a protractor for measuring branch line angles, and FIG. 28 is a diagram showing member names, tag names, and recognition conditions of the protractor.
The cooperative API unit 210 acquires an inspection image. The image recognition engine 222 corresponding to the branch line angle as the inspection item name displays a needle as a member of the protractor and angle indications (0, 10, 20, 30, 40, 50, 60 degrees) on the inspection image. , 70 degrees, 80 degrees, and 90 degrees) are included. The image recognition engine 222 determines whether or not the inspection image includes an image indicating the probability value for each angle value of the angle display. The image recognition engine 222 computes needle confidence and angle representation (for each angle value) confidence. The image recognition engine 222 determines that a needle and an angle indication (for each angle value) are present in the inspection image when each certainty exceeds a predetermined value, and places a protractor in the area including the needle and the angle indication in the inspection image. Add tag information corresponding to the needle and angle display (for each angle value).

FIG. 29 is a diagram for explaining the process of inspecting branch line angles. The cooperative API unit 210 sets the intermediate position between the half position and the tip position of the area from the half position to the tip position of the needle recognized by the image recognition engine 222 as the reference coordinates of the needle. The cooperative API unit 210 acquires the tag information of the top two angle indications (50 degrees and 30 degrees) closest to the reference coordinates of the needle. The cooperative API unit 210 calculates the distance (the number of pixels) between the images of the two acquired angle representations, and divides the calculated number of pixels by the angle difference (20 degrees) between the two angle representations, thereby obtaining pixels around 1 degree. Calculate the number. Although the tag information for each angle value includes area information including the angle value, the linked API unit 210 calculates the distance (in-image horizontal direction) between the centroids of the areas including the angle value. The cooperative API unit 210 calculates the angle (branch line angle) indicated by the needle based on the number of pixels between the needle and the angle display (30 degrees) in the horizontal direction in the image and the angle (30 degrees) indicated by the angle display. do. If the branch line angle is within a predetermined range, the linked API unit 210 determines that the inspection result of the branch line angle is acceptable, and otherwise determines that the inspection result is unacceptable.

(Installation work for winding grips)
FIG. 30 is an example of a photographed image including a wrapped grip, and FIG. 31 is a diagram showing member names, tag names, and recognition conditions in installation work of the wrapped grip.
The cooperative API unit 210 acquires an inspection image. The image recognition engine 222 corresponding to the wrapped grip as the inspection item name determines whether or not the captured image includes red tape, purple tape, brown tape, orange tape, and thimble as construction members. The image recognition engine 222 calculates the reliability of each construction member. The image recognition engine 222 determines that a construction member exists in the inspection image when each reliability exceeds a predetermined value, and adds tag information corresponding to the construction member to a region including the construction member in the inspection image.

The cooperative API unit 210 adopts the color with the highest degree of certainty among the colors of the tape. The cooperative API unit 210 compares the adopted tape color with the tape color corresponding to the geographical conditions, economic efficiency, etc. of the facility location. The cooperative API unit 210 determines that the inspection result of the tape color is acceptable if the colors match each other, and determines that the inspection result is unacceptable if not. Note that the cooperative API unit 210 may perform the tape color inspection when the user's operation is received and the upper part of the branch line in the inspection image is selected.

The cooperative API unit 210 determines whether or not tag information corresponding to the thimble is attached to the inspection image. Otherwise, the test result is determined to be unsatisfactory. If all the inspection results for the construction member pass, the cooperative API unit 210 determines that the inspection result for the construction of the wrapped grip has passed.

(Clogs installation work)
FIG. 32 is an example of an inspection image including clogs, and FIG. 33 is a diagram showing member names, tag names, and recognition conditions in the installation work of clogs.
The cooperative API unit 210 acquires an inspection image. The image recognition engine 222 corresponding to clogs as an inspection item name determines whether clogs are included in the inspection image as construction members. When the clogs are included, the image recognition engine 222 adds tag information corresponding to the clogs to a region including the clogs in the inspection image. The cooperative API unit 210 determines that the inspection result is acceptable if the region in the inspection image selected by the user includes the geta tag, and otherwise determines that the inspection result is unsatisfactory.

(Protection of branch lines and installation of insulators)
FIG. 34 is an example of an inspection image including branch line protection and insulators, and FIG. 35 is a diagram showing member names, tag names, and recognition conditions in installation work of branch line protection and insulators.
The cooperative API unit 210 acquires an inspection image. The image recognition engine 222 corresponding to branch line protection and insulator as inspection item names determines whether branch line protection and insulator are included as construction members in the inspection image. When a construction member is included, the image recognition engine 222 adds tag information corresponding to the construction member to a region including the construction member in the inspection image. The cooperative API unit 210 determines that the inspection result is acceptable if the region in the inspection image selected by the user includes a branch line protection tag, and otherwise determines that the inspection result is unsatisfactory. . The cooperative API unit 210 determines that the inspection result is acceptable if the region in the inspection image selected by the user includes an insulator tag, and otherwise determines that the inspection result is unacceptable. .

(Insulator installation work)
FIG. 36 is an example of an inspection image including insulators, and FIG. 37 is a diagram showing member names, tag names, and recognition conditions in insulator installation work. An insulator is an insulator attached to a line (power line, communication line, etc.) connected to a utility pole.
The cooperative API unit 210 acquires an inspection image. The image recognition engine 222 corresponding to the insulator as the inspection item name determines whether or not the inspection image includes the insulator as the construction member. When an insulator is included, the image recognition engine 222 adds tag information corresponding to the insulator to a region including the insulator in the inspection image. The cooperative API unit 210 determines that the inspection result is acceptable if the region in the inspection image selected by the user includes an insulator tag, and otherwise determines that the inspection result is unacceptable. .

(Installation work of fasteners)
FIG. 38 is an example of an inspection image including fasteners, and FIG. 39 is a diagram showing member names, tag names, and recognition conditions in fastener installation work. A fastener is a construction member for drawing a line (power line, communication line, etc.) connected to a utility pole into a house.
The cooperative API unit 210 acquires an inspection image. The image recognition engine 222 corresponding to the fastener as the inspection item name determines whether or not the inspection image includes the fastener as the construction member. If a fastener is included, the image recognition engine 222 adds tag information corresponding to the fastener to a region including the fastener in the inspection image. The cooperative API unit 210 determines that the inspection result is acceptable when the region in the inspection image selected by the user includes the tag of the fastener, and otherwise determines that the inspection result is unsatisfactory. .

According to the facility construction inspection system of the embodiment described above, an inspection image obtained by imaging the metal fittings for segmentation and inspection item information indicating that the inspection item is the metal fittings for segmentation are acquired from the user terminal device 100, Calculating a degree of certainty that indicates the certainty of the object to be inspected based on the inspection image, recognizing whether or not the object to be inspected is included in the inspection image based on the degree of certainty, and if the object to be inspected is included in the inspection image Then, based on the positional relationship of the object to be inspected, the equipment construction using the metal fittings for branching is inspected, and the result of the inspection performed by the inspecting unit is notified to the terminal device. According to this facility construction inspection system, inspection is performed based on the positional relationship of the object to be inspected in the inspection of the hardware for segmentation, so the construction state using a plurality of members can be inspected with high accuracy.

In addition, according to the equipment construction inspection system of the embodiment, an inspection image obtained by imaging a support member of a utility pole and inspection item information indicating that the inspection item is the support member of the utility pole are acquired from the terminal device, and from the inspection image, Calculating a certainty factor indicating the certainty of the supporting member and the inspection object around the supporting member, determining whether or not the supporting member and the inspection object are included in the inspection image based on the certainty factor, and supporting the inspection image When a member and an object to be inspected are included, inspect equipment construction using the supporting member based on the positional relationship between the supporting member and the object to be inspected, and notify the terminal device of the results of the inspection performed by the inspection department. do. According to this facility construction inspection system, the inspection of the support member of the utility pole is performed based on the positional relationship of the inspection object, so the construction state using a plurality of members can be inspected with high accuracy.

Furthermore, according to the facility construction inspection system of the embodiment, the terminal device presents a plurality of icons corresponding to a plurality of inspection items included in the facility construction, and an icon is selected from among the plurality of icons based on the user's operation. In response to the selection, the camera device of the terminal device is activated, and the guide frame of the inspection object in the search item corresponding to the selected icon and the mask indicating the area not recognized as the inspection object are displayed on the terminal device. , an inspection image captured by a camera device and inspection item information corresponding to the selected icon are transmitted from the terminal device to the information processing device, and the recognition unit recognizes the inspection included in the inspection image transmitted from the terminal device. calculating a degree of certainty indicating the certainty of an object, determining whether or not the inspection object is included based on the degree of certainty, and if the recognition unit determines that the inspection object is included, the information processing device The inspection unit performs an inspection corresponding to the inspection item based on the positional relationship of the inspection object. According to this equipment construction inspection system, it is possible to acquire an appropriate inspection image by displaying a guide frame and a mask when imaging an object to be inspected. can do.

Although each embodiment and modifications have been described, these are only examples and are not limited to these. A section may be combined with one or more other embodiments or one or more other modifications to realize one aspect of the present invention.

A program for executing each process of the user terminal device 100 and the construction inspection server device 200 in this embodiment is recorded on a computer-readable recording medium, and the program recorded on the recording medium is transferred to the computer system. The above-described various processes related to the user terminal device 100 and the construction inspection server device 200 may be performed by reading and executing.

Note that the “computer system” referred to here may include hardware such as an OS and peripheral devices. The "computer system" also includes the home page providing environment (or display environment) if the WWW system is used. In addition, "computer-readable recording medium" means writable non-volatile memory such as flexible disk, magneto-optical disk, ROM, flash memory, portable medium such as CD-ROM, hard disk built in computer system, etc. storage device.

Furthermore, "computer-readable recording medium" means a volatile memory (e.g., DRAM (Dynamic
Random Access Memory)), which holds a program for a certain period of time. Also, the program may be transmitted from a computer system storing the program in a storage device or the like to another computer system via a transmission medium or by a transmission wave in the transmission medium.

Here, the "transmission medium" for transmitting the program refers to a medium having a function of transmitting information, such as a network (communication network) such as the Internet or a communication line (communication line) such as a telephone line. Further, the program may be for realizing part of the functions described above. Further, it may be a so-called difference file (difference program) that can realize the above-described functions in combination with a program already recorded in the computer system.

As described above, the embodiments of the present invention have been described in detail with reference to the drawings, but the specific configuration is not limited to these embodiments, and includes designs within the scope of the gist of the present invention. For example, as an embodiment of the present invention, the user terminal device 100 may be used alone.

100 User terminal device 120 Guide frame image 122 Mask image 200 Construction inspection server device 210 Cooperation API unit 212 API unit 214 Image processing unit 216 Inspection unit 218 Control unit 220 Inspection object recognition unit 222 Image recognition engine

Claims (8)

an acquisition unit that acquires, from a terminal device, an inspection image obtained by imaging a support member of a utility pole and inspection item information indicating that the inspection item is the support member of the utility pole;
calculating from the inspection image a certainty factor indicating the likelihood of the supporting member and the inspection object around the supporting member, and based on the certainty factor, whether the supporting member and the inspection object are included in the inspection image; a recognition unit that determines whether
an inspection unit that, when the inspection image includes the support member and the inspection object, inspects equipment construction using the support member based on the positional relationship between the support member and the inspection object;
a notification unit that notifies the terminal device of the result of the inspection performed by the inspection unit;
Equipment construction inspection device. The recognition unit is a recognition model trained using a learning image including a normal inspection object, and recognizes that the inspection object exists when an inspection image including a normal inspection object is input. Using a recognition model whose processing parameters have been learned as
The equipment construction inspection device according to claim 1 . The recognition unit recognizes, as the inspection object, a virtual cross section of a utility pole supported by the support member,
The inspection unit performs inspection based on the positional relationship between the position of the support member and a virtual point based on the area occupied by the virtual cross section.
3. The facility construction inspection device according to claim 1 or 2. The recognition unit recognizes an arrow indicating a pulling direction of the utility pole from the inspection image as the inspection object,
The inspection unit performs inspection based on the relationship between the position of the support member, the virtual point based on the area occupied by the virtual cross section, and the direction indicated by the arrow.
The equipment construction inspection device according to claim 3 . The inspection unit performs an inspection based on the relationship between the center-of-gravity position of the support member, the center-of-gravity position of the utility pole based on the area occupied by the virtual cross section, and the direction indicated by the arrow.
The equipment construction inspection device according to claim 4 . When the utility pole is of the first type, the inspection unit determines whether an angle between the longitudinal direction of the support member and the direction of the arrow is within a first angle. type, determining whether the angle between the longitudinal direction of the support member and the direction of the arrow is within a second angle;
6. The facility construction inspection device according to claim 4 or 5. acquiring from the terminal device an inspection image obtained by imaging a support member of a utility pole and inspection item information indicating that the inspection item is the support member of the utility pole;
calculating a certainty factor indicating the likelihood of the support member and the inspection object around the support member from the inspection image;
determining whether or not the support member and the inspection object are included in the inspection image based on the certainty;
a step of inspecting equipment construction using the support member based on the positional relationship between the support member and the inspection object when the inspection image includes the support member and the inspection object;
a step of notifying the terminal device of the result of the inspection;
A method for inspecting facility construction. to the computer,
acquiring from the terminal device an inspection image obtained by imaging a support member of a utility pole and inspection item information indicating that the inspection item is the support member of the utility pole;
calculating a certainty factor indicating the likelihood of the support member and the inspection object around the support member from the inspection image;
determining whether or not the support member and the inspection object are included in the inspection image based on the certainty;
a step of inspecting equipment construction using the support member based on the positional relationship between the support member and the inspection object when the inspection image includes the support member and the inspection object;
a step of notifying the terminal device of the result of the inspection;
The program that causes the to run.

Images