JP2020065330A - Abnormality detection method, program, generation method for learnt model, and learnt model - Google Patents

Abstract

To provide an abnormality detection method, etc., capable of appropriately detecting an abnormality in a power distribution facility.SOLUTION: An abnormality detection method executes, at a computer, steps of: acquiring a captured image including power distribution facilities which are installed along a road; detecting abnormality in the power distribution facility from the acquired captured image by using a learnt model which is learned so as to output information indicating presence/absence of abnormality in the power distribution facility when the captured image is inputted; and outputting a detection result. Thus, abnormality in the power distribution facility can be appropriately detected.SELECTED DRAWING: Figure 1

Description

  The present invention relates to an abnormality detection method, a program, a learned model generation method, and a learned model.

  Various methods have been proposed for detecting abnormalities in power distribution equipment. For example, in Patent Document 1, a single infrared camera and a plurality of visible light cameras capture an image of a utility pole, an image area of an arm (arm) is extracted from an image captured by the visible light camera, and a thermal image captured by the infrared camera. Discloses an abnormality detection method for determining whether the pole equipment is normal or abnormal by calculating the temperature gradient of the pole equipment from the image area of the arm.

JP, 2002-366953, A

  However, the invention according to Patent Document 1 detects an abnormality by pattern matching of a temperature gradient, and it is necessary to manually create a rule for the computer to determine that the state is abnormal.

  In one aspect, it is an object to provide an abnormality detection method and the like that can appropriately detect an abnormality in power distribution equipment.

  In one aspect, the abnormality detection method is to learn to acquire a captured image of power distribution equipment installed along a road and output information indicating whether there is abnormality in the power distribution equipment when the captured image is input. It is characterized by causing the computer to execute a process of detecting an abnormality of the power distribution equipment from the acquired captured image using the learned model that has already been learned and outputting the detection result.

  In one aspect, it is possible to properly detect an abnormality in the distribution equipment.

It is a schematic diagram which shows the structural example of an abnormality detection system. It is a block diagram which shows the structural example of a server. It is explanatory drawing which shows an example of the record layout of telephone pole DB. It is explanatory drawing regarding the production | generation process of an abnormality detection model. It is explanatory drawing regarding an abnormality detection process. 7 is a flowchart illustrating an example of a processing procedure of generation processing of an abnormality detection model. It is a flow chart which shows an example of the processing procedure of abnormality detection processing. FIG. 8 is an explanatory diagram showing an outline of the second embodiment. 9 is a flowchart showing an example of a processing procedure of abnormality detection processing according to the second embodiment.

Hereinafter, the present invention will be described in detail with reference to the drawings showing an embodiment thereof.
(Embodiment 1)
FIG. 1 is a schematic diagram showing a configuration example of an abnormality detection system. In the present embodiment, as an example of power distribution equipment to be detected, a utility pole 3 installed on a road, a pole material installed on the utility pole 3, or a utility pole 3, 3, 3, ... An abnormality detection system that detects an abnormal portion from an image obtained by picking up an electric pole 3, a pole material, an electric wire, etc. will be described. The abnormality detection system includes the information processing device 1 and the vehicle 2. Each device is communicatively connected via a network N such as the Internet.

  The information processing device 1 is an information processing device capable of various information processing and information transmission / reception, and is, for example, a server device, a personal computer, or the like. In the present embodiment, the information processing device 1 is assumed to be a server device, and will be read as server 1 in the following description for simplicity. The server 1 acquires an image of the utility pole 3 or the like installed along the road, and performs a process of detecting an abnormal location of the utility pole 3 or the like from the image. In the present embodiment, the server 1 performs anomaly detection using an anomaly detection model (learned model) that has been learned so as to detect (identify) anomalous locations such as telephone poles 3 in an image by machine learning.

  The vehicle 2 is a vehicle on which a predetermined worker who carries out maintenance and inspection of the electric pole 3 and the like gets on, and is a patrol vehicle that patrols the road on which the electric pole 3 is installed. For example, the vehicle 2 is equipped with a camera 21 that captures an image of the surroundings of the vehicle and an in-vehicle device (not shown) that has a position information acquisition function and a communication function of the vehicle 2. The camera 21 is, for example, an omnidirectional camera, and captures an image of the scenery around the vehicle. The in-vehicle device is connected to the camera 21, transfers the captured image to the server 1, and also transfers the position information (for example, GPS information) of the vehicle 2 at the time of image capturing.

  The installation location of the camera 21 is not limited to the upper part of the vehicle body, and may be installed as an in-vehicle camera, such as a drive recorder system. Further, the server 1 may sequentially acquire the images captured by the camera 21 in real time, or may collectively acquire the images recorded (recorded) by the vehicle-mounted device for a certain period of time after the image capturing is completed.

  The server 1 inputs the captured image acquired from the camera 21 into the abnormality detection model, and acquires the identification result indicating the presence / absence of abnormality of the telephone pole 3 or the like as an output. Specifically, as will be described later, in addition to the presence / absence of abnormality in the pole equipment mounted on the pole 3, the presence / absence of nesting 4 of birds (animals) formed on the pole equipment or the pole 3 An identification result indicating whether or not there is contact between the electric wire erected in and the tree 5 planted in the vicinity is acquired. The server 1 notifies the operator of a predetermined notification urging the user to deal with an abnormal place such as the telephone pole 3 according to the identification result output from the abnormality detection model.

FIG. 2 is a block diagram showing a configuration example of the server 1. The server 1 includes a control unit 11, a main storage unit 12, a communication unit 13, and an auxiliary storage unit 14.
The control unit 11 has an arithmetic processing unit such as one or more CPUs (Central Processing Units), MPUs (Micro-Processing Units), and GPUs (Graphics Processing Units), and stores the programs P stored in the auxiliary storage unit 14. By reading and executing, various information processing, control processing, and the like related to the server 1 are performed. The main storage unit 12 is a temporary storage area such as an SRAM (Static Random Access Memory), a DRAM (Dynamic Random Access Memory), and a flash memory, and temporarily stores data necessary for the control unit 11 to execute arithmetic processing. Remember. The communication unit 13 is a communication module for performing processing relating to communication, and transmits / receives information to / from the outside.

  The auxiliary storage unit 14 is a large-capacity memory, a hard disk, or the like, and stores the program P and other data necessary for the control unit 11 to execute processing. The auxiliary storage unit 14 also stores a telephone pole DB 141 and an abnormality detection model 142. The utility pole DB 141 is a database that stores information on the utility poles 3 and the like that are the targets of abnormality detection. The abnormality detection model 142 is a classifier that identifies an abnormal portion such as the telephone pole 3 from the captured image, and is a learned model generated by machine learning.

  The auxiliary storage unit 14 may be an external storage device connected to the server 1. The server 1 may be a multi-computer including a plurality of computers, or may be a virtual machine virtually constructed by software.

  Further, in the present embodiment, server 1 is not limited to the above configuration, and may include, for example, a reading unit that reads information stored in a portable storage medium, an input unit that receives an operation input, a display unit that displays an image, and the like. .

  FIG. 3 is an explanatory diagram showing an example of a record layout of the telephone pole DB 141. The utility pole DB 141 includes a utility pole ID sequence, a position information sequence, and a utility pole information sequence. The telephone pole ID column stores telephone pole IDs for identifying each telephone pole 3. The position information sequence stores position information indicating the installation position of the electric pole 3 in association with the electric pole ID. The position information is, for example, coordinate values related to GPS (Global Positioning System) or information such as an address. The utility pole information stores information on each utility pole 3 in association with the utility pole ID. The utility pole information may include, for example, information on various devices and parts installed on the arm of the utility pole 3, information on a worker in charge of maintenance and inspection of the utility pole 3, and the like.

  FIG. 4 is an explanatory diagram regarding the generation processing of the abnormality detection model 142. FIG. 4 conceptually illustrates a process of performing the machine learning to generate the abnormality detection model 142. The process of generating the abnormality detection model 142 will be described with reference to FIG.

  In the present embodiment, the server 1 uses the captured image as an input by performing deep learning that learns the image feature amount of the abnormal portion in the captured image of the utility pole 3 or the like as the abnormality detection model 142, and detects the abnormality of the utility pole 3 or the like. A neural network that outputs the information indicating the presence or absence of is constructed (generated). For example, the neural network is a CNN (Convolution Neural Network), and an input layer that receives an input of a captured image, an output layer that outputs information (identification result) indicating the presence or absence of an abnormality, and an intermediate that extracts the image feature amount of the captured image. And layers.

  The input layer has a plurality of neurons that receive the input of the pixel value of each pixel included in the captured image, and transfers the input pixel value to the intermediate layer. The intermediate layer has a plurality of neurons that extract the image feature amount of the captured image, and passes the extracted image feature amount to the output layer. For example, when the abnormality detection model 142 is CNN, the intermediate layer alternates between the convolution layer that convolves the pixel value of each pixel input from the input layer and the pooling layer that maps the pixel value convolved by the convolution layer. The feature amount of the image is finally extracted while compressing the pixel information of the captured image. The output layer has one or a plurality of neurons that output an identification result that identifies an abnormal part such as the utility pole 3, and identifies the presence or absence of an abnormality of the utility pole 3 based on the image feature amount output from the intermediate layer.

  In the present embodiment, the abnormality detection model 142 is described as being CNN, but the abnormality detection model 142 is not limited to CNN, and a neural network other than CNN, SVM (Support Vector Machine), Bayesian network, regression tree It may be a trained model constructed by another learning algorithm such as.

  The server 1 provides teacher data in which a plurality of captured images of the utility pole 3, the mounting equipment provided on the utility pole 3, electric wires, and the like, and information indicating an abnormal portion of the utility pole 3 in each image are associated with each other. Use to learn. For example, as shown in FIG. 4, the teacher data is data in which the captured image of the telephone pole 3 or the like is labeled with the coordinate range of the image area corresponding to the abnormal portion and the content of the abnormality. In FIG. 4, the labeled image area is illustrated by a thick rectangular frame.

  The server 1 inputs the captured image, which is the teacher data, into the input layer, and through the arithmetic processing in the intermediate layer, acquires the identification result indicating the presence / absence of abnormality of the utility pole 3 or the like from the output layer. The identification result output from the output layer may be a value (for example, a value of “0” or “1”) discretely indicating the presence or absence of abnormality, and a continuous probability value (for example, “0” to “ Value in the range up to 1 ”). For example, the server 1 obtains, as the identification result output from the output layer, the identification result that identifies the image area corresponding to the abnormal location, the type of abnormality, the degree of abnormality, etc. in addition to the presence / absence of the abnormal location on the telephone pole 3 or the like. To do. Details of the type and degree of abnormality will be described later.

  The server 1 compares the identification result output from the output layer with the information labeled on the captured image in the teacher data, that is, the correct value, so that the output value from the output layer approaches the correct value. Optimize the parameters used in the calculation processing in. The parameter is, for example, a weight between neurons (coupling coefficient), a coefficient of an activation function used in each neuron, or the like. The method of optimizing the parameters is not particularly limited, but the server 1 optimizes various parameters using the error back propagation method, for example.

  The server 1 performs the above processing on each captured image included in the teacher data to generate the abnormality detection model 142. When the captured image is acquired from the camera 21 of the vehicle 2, the server 1 uses the abnormality detection model 142 to detect an abnormal portion such as the utility pole 3.

FIG. 5 is an explanatory diagram related to the abnormality detection processing. FIG. 5 conceptually illustrates how to detect an abnormal place such as a utility pole 3 from a captured image. An abnormality detection process for the utility pole 3 and the like will be described with reference to FIG.
The server 1 acquires the image captured by the camera 21 from the vehicle 2 and inputs it to the abnormality detection model 142. For example, the server 1 acquires the captured image from the vehicle 2 and simultaneously acquires the position information (GPS information or the like) indicating the imaging point.

  The server 1 performs arithmetic processing to extract the feature amount of the captured image in the intermediate layer of the abnormality detection model 142, inputs the extracted feature amount to the output layer of the abnormality detection model 142, and determines whether there is an abnormality in the utility pole 3 or the like. Is obtained as an output.

  In the present embodiment, the server 1 performs processing for a plurality of detection items (diagnosis items) when detecting an abnormality in the telephone pole 3 or the like. As the first detection item, the server 1 detects an abnormality in the pole equipment mounted on the telephone pole 3. For example, the server 1 detects an abnormality of lightning protection equipment such as a discharge clamp. The discharge clamp is a part of a clamp insulator (insulator) that lays an electric wire, and is a device that protects the electric wire from being blown out from an abnormal current generated by a lightning strike or the like. When detecting the abnormality of the discharge clamp, the server 1 detects the darkness, cracks, and the like of the discharge clamp using the abnormality detection model 142 to evaluate the resistance to lightning strike and detect the abnormality.

  The electric resistant material to be detected is not limited to the discharge clamp, and may be, for example, a lightning arrester or an overhead ground wire. Further, the pillar equipment to be detected is not limited to the lightning protection equipment, and may include various equipment and parts such as a transformer, a switch, an insulator, a lead wire, and a nut for attaching each equipment. Alternatively, the presence or absence (for example, disconnection) of an electric wire installed between the plurality of utility poles 3, 3, 3 ... May be detected.

  For example, if the discharge clamp is cracked, a problem may occur in the discharge and an accident may occur, so the discharge clamp needs to be replaced. Therefore, the server 1 uses the abnormality detection model 142 to detect an abnormality in the pillar material. For example, the server 1 uses the abnormality detection model 142 to detect an abnormal location (a column equipment having an abnormality), and also detects the type and degree of the abnormality. For example, as shown in the upper right of FIG. 5, the server 1 detects that the abnormal portion is a discharge clamp, and that the type of abnormality is cracking and the degree of cracking (“rank” in FIG. 5).

  As described above, the server 1 uses the abnormality detection model 142 to detect an abnormal portion that requires replacement or repair of the pillar equipment. For example, the server 1 performs learning by using an image obtained by capturing a pillar material having an abnormal portion as teacher data, and detects an abnormality using the abnormality detection model 142 that has learned an abnormality of the pillar material. It should be noted that the server 1 may build a separate abnormality detection model 142 (neural network) for each of the pillar equipment (abnormal point), and the single abnormality detection model 142 learns the abnormality of all the pillar equipment. You may leave it.

  Further, the target of abnormality detection is not limited to the pillar equipment. For example, the server 1 detects, as the second detection item, an obstacle existing on the pillar material as an abnormal point. The obstacle is, for example, a nest 4 of birds (animals) formed on an arm mounted on a utility pole 3 (see FIG. 1).

  When the nesting 4 exists on the arm, for example, at the time of a lightning strike, a current may be conducted through the nesting 4 and an accident may occur. Especially when the nesting 4 is made of metal rather than wood, it is easy for the current to conduct, and immediate removal is required. Therefore, the server 1 detects the presence or absence of nesting 4 using the abnormality detection model 142, and also detects the type (eg, material), degree (eg, size), etc. of nesting 4.

  Although nesting 4 has been mentioned above as an obstacle existing on the arm, the present embodiment is not limited to this, and may be a flying object such as a balloon or a kite caught on the arm. . That is, the obstacle to be detected may be an object existing on the pillar material, and may be any object other than the object previously attached to the utility pole 3, and its specific content is not particularly limited.

  In addition, as a third detection item, the server 1 detects the presence or absence of a contact object, specifically a tree 5, that comes into contact with an electric wire laid on the utility pole 3. When the branches and leaves of the tree 5 are in contact with the electric wire, a current may flow from the electric wire to the tree 5 and a fire or the like may occur. In this case, it may be necessary to cut off a part of the branch or cover the electric wire with a protective cover. Therefore, the server 1 uses the abnormality detection model 142 to detect the presence or absence of contact between the electric wire and the tree 5. The contact state between the electric wire and the tree 5, which is detected as an abnormal state, may include not only the state where both are actually in contact but also the case where both are close to a predetermined separation distance.

  In the above description, the tree 5 is mentioned as a contact object that comes into contact with the electric wire, but the present embodiment is not limited to this. For example, a flying object such as a balloon caught on the electric wire, a kite, or the tree 5 may be used. It may be an object (for example, a building) located near the electric wire. That is, the contact object to be detected may be any object that comes into contact with the electric wire, and its specific content is not particularly limited.

  When the server 1 detects the above-mentioned various abnormal parts from the captured image, the server 1 notifies (outputs) the detection result indicating that the abnormality of the telephone pole 3 or the like has been detected. For example, the server 1 notifies the vehicle-mounted device of the vehicle 2 traveling on the road to notify the worker. Alternatively, the server 1 may notify a predetermined mobile terminal carried by the worker.

  For example, the server 1 also transmits, as the information to be notified, the detection result indicating the abnormal location of the detected utility pole 3 or the like, the type of abnormality, and the degree thereof, as well as the position information of the utility pole 3 that detected the abnormal location. The position information is information such as GPS information and a utility pole ID, and is information indicating the installation position of the corresponding utility pole 3. The server 1 notifies the worker that the abnormality of the telephone pole 3 or the like is detected, and also transmits the position information of the telephone pole 3 together.

  An example of the notification screen showing the content of the notification from the server 1 is illustrated on the lower side of FIG. For example, the in-vehicle device displays each notification screen shown in the lower side of FIG. 5 on a navigation display (not shown) mounted on the vehicle 2. For example, as shown in the lower left of FIG. 5, the in-vehicle device displays a notification screen showing the position of the corresponding utility pole 3 on the map with an icon or the like, and guides the worker to the work site (power pole 3). Further, for example, the in-vehicle device displays a notification screen that allows the detected abnormal place to be identified by a rectangular frame or the like in the image captured by the camera 21, as shown in the lower right part of FIG. To notify the worker in advance. The in-vehicle device displays information such as the detected abnormal place, type of abnormality, and degree on the screen in text (not shown). In the case of the example in the upper right of FIG. 5, the in-vehicle device displays “discharge clamp” as the name of the abnormal place, “crack” as the type of abnormality, and rank “B” as the degree of abnormality. The worker is guided to the telephone pole 3 by receiving a guide by a map or a captured image, and performs necessary work.

  Note that, for example, when the server 1 notifies the detection result of an abnormal place such as the telephone pole 3, the server 1 may also notify the reliability indicating the certainty of the detection result. The reliability is, for example, a probability value output from the output layer of the abnormality detection model 142, and the degree of reliability with which the detected abnormal portion is detected is determined from “0” to “1”. It is a numerical value indicated by a continuous value. By notifying the worker of the reliability of the detection result together with the detection result, the worker can know how accurate the detection result is, and the convenience of the worker can be improved.

  In addition, for example, the server 1 may receive correct / incorrect input of the detection result from the worker via the in-vehicle device after the above notification. Then, the server 1 may re-learn the abnormality detection model 142 using the captured image of the utility pole 3 or the like in which the abnormal portion is detected and the correctness of the detection result input by the operator as teacher data. As a result, as the system is continued, the abnormality detection model 142 is updated, and more accurate abnormality detection can be performed.

  The server 1 may change the notification content or the timing according to the abnormality detection result of the utility pole 3 or the like. For example, the server 1 may determine whether or not urgent action is required, that is, the degree of need for action, according to the type and degree (rank) of the detected abnormality, and notify the operator of the determination result. Further, for example, the server 1 may perform the notification only when it is determined that the necessity degree is high. In this way, the server 1 may change the notification mode according to the detection result output from the abnormality detection model 142.

  As described above, the server 1 uses the abnormality detection model 142 (learned model) in which the characteristics of the abnormal place have been learned from the captured image of the electric pole 3 to detect the abnormal place of the electric pole 3. As a result, the operator does not need to perform visual inspection, and the efficiency of maintenance and inspection work can be improved.

FIG. 6 is a flowchart showing an example of a processing procedure of generation processing of the abnormality detection model 142. Processing contents of the generation processing of the abnormality detection model 142 will be described based on FIG. 6.
The control unit 11 of the server 1 handles a plurality of captured images of the utility pole 3, the pole equipment provided on the utility pole 3, electric wires, and the information indicating an abnormal location of the utility pole 3 in each captured image. The attached teacher data is acquired (step S11). The teacher data is, for example, data obtained by labeling an imaged image of a telephone pole 3 or the like with an image region corresponding to an abnormal place and the type and degree of the abnormality.

  The control unit 11 uses the teacher data to generate an abnormality detection model 142 (learned model) that outputs information indicating whether or not there is an abnormality in the telephone pole 3 or the like when a captured image of the telephone pole 3 or the like is input (step S12). ). Specifically, the control unit 11 inputs the captured image, which is the teacher data, to the input layer of the neural network, and displays the identification result obtained by identifying the image region corresponding to the abnormal portion in the captured image, the type of abnormality, the degree, and the like. Get from output layer. The control unit 11 compares the acquired identification result with the correct value of the teacher data (information labeled on the captured image), so that the identification result output from the output layer approaches the correct value. The parameters (weight etc.) used for the arithmetic processing are optimized. The control unit 11 stores the generated abnormality detection model 142 in the auxiliary storage unit 14, and ends the series of processes.

FIG. 7 is a flowchart showing an example of the processing procedure of the abnormality detection processing. Based on FIG. 7, the processing content of the abnormality detection processing of the telephone pole 3 and the like will be described.
The control unit 11 of the server 1 acquires a picked-up image of the electric pole 3 and the pole equipment, electric wires, and the like installed on the electric pole 3 from the vehicle 2 traveling on the road on which the electric pole 3 is installed (step S31). For example, the control unit 11 acquires a captured image of the utility pole 3 and the like and position information indicating an imaging point.

  The control unit 11 uses the abnormality detection model 142 to detect an abnormality of the utility pole 3 or the like from the acquired captured image (step S32). For example, the control unit 11 detects whether there is an abnormality in the utility pole 3 or the like, and when there is an abnormality, detects the corresponding abnormal location, type of abnormality, degree, and the like.

  The control unit 11 determines whether or not there is an abnormality in the utility pole 3 or the like based on the detection result in step S32 (step S33). Note that in step S33, the control unit 11 may make a determination including not only the presence / absence of an abnormality, but also the degree of need for the operator to respond, depending on the abnormal location, the type of abnormality, the degree, and the like. When it is determined that there is no abnormality (S33: NO), the control unit 11 ends the series of processes.

  When it is determined that there is an abnormality (S33: YES), the control unit 11 notifies the operator of the detection result indicating that the abnormality of the telephone pole 3 or the like has been detected (step S34). For example, the control unit 11 specifies the corresponding utility pole 3 from the position information acquired together with the captured image in step S31, and includes position information (for example, GPS information, utility pole ID, etc.) that can identify the installation position of the specified utility pole 3. Notify me. For example, the control unit 11 notifies the operator of the detection result indicating the abnormal place detected in step S32, the type of abnormality, the degree, and the like. The control unit 11 may notify the in-vehicle device mounted on the vehicle 2 or the terminal device of the worker who is in the vehicle 2. The control unit 11 ends the series of processes.

  In the above description, it is assumed that the worker is in the vehicle 2, but it is not necessary for the worker to be in the vehicle 2, and the vehicle 2 may be configured to only image the utility pole 3. . For example, when the vehicle 2 travels around the road to capture an image of the utility pole 3 and the like, and the server 1 detects an abnormal portion from the captured image, the above-mentioned notification is sent to the terminal device of the worker who stands by outside so that he / she can take action. You can

  Further, in the above description, the electric pole 3 and the like are imaged from the vehicle 2, but in the present embodiment, it is possible to image the electric pole 3 and the like from a moving body that moves along the road on which the electric poles 3, 3, 3 ... Are installed. For example, instead of the vehicle 2, the image of the utility pole 3 or the like may be taken from a drone.

  Further, in the above description, the server 1 is described as performing the abnormality detection remotely, but the present embodiment is not limited to this, and a series of local terminal devices such as an in-vehicle device and a portable terminal of a worker can be used. Processing may be performed.

  As described above, according to the first embodiment, it is possible to appropriately detect the abnormality of the power distribution equipment by performing the abnormality detection using the abnormality detection model 142 that has learned the characteristics of the abnormal portion of the power distribution equipment.

  Further, according to the first embodiment, it is possible to improve the efficiency of maintenance / inspection work of the utility pole 3 and the like by detecting an abnormality in the pole material installed on the utility pole 3.

  Further, according to the first embodiment, it is possible to more appropriately support the maintenance and inspection of the utility pole 3 and the like by detecting the obstacle (the nesting 4) existing on the pillar material.

  Further, according to the first embodiment, by detecting a contact object (tree 5) that comes into contact with the electric wire laid on the utility pole 3, maintenance and inspection of the utility pole 3 and the like can be more appropriately supported.

  Further, according to the first embodiment, at least the abnormal place of the power distribution equipment and the type of the abnormality are notified to the worker, so that the maintenance inspection can be more appropriately supported.

  Further, according to the first embodiment, the vehicle 2 makes a patrol to capture an image of the utility pole 3, and the abnormality is detected from the image captured by the vehicle 2, so that the efficiency of the maintenance / inspection work can be more suitably achieved. be able to.

(Embodiment 2)
In the present embodiment, a mode will be described in which an abnormal portion is detected from an image provided by a predetermined imager, not an image captured from a dedicated vehicle 2 that is patrol for maintenance and inspection. In addition, the same reference numerals will be given to the same contents as those in the first embodiment, and the description thereof will be omitted.
FIG. 8 is an explanatory diagram showing an outline of the second embodiment. Since the contents of FIG. 8 are almost the same as those of FIG. 1, description of common items will be omitted. In the present embodiment, the camera 21 is mounted on the operation vehicle 203 that regularly operates on the road on which the utility pole 3 is installed, instead of the dedicated vehicle 2 that travels on the road to image the utility pole 3 and the like. The electric pole 3 and the like are imaged from the running vehicle 203.

  The operating vehicle 203 is, for example, a bus, a garbage truck, etc., and is a vehicle operated by a business operator other than a business operator (for example, an electric power company) that performs maintenance and inspection of the utility pole 3 and the like for purposes other than maintenance and inspection of the utility pole 3. is there. Since these operating vehicles 203 pass by the utility pole 3 regularly, by mounting the camera 21 on the operating vehicle 203, it is not necessary to make the dedicated vehicle 2 circulate. The server 1 acquires an image captured by the camera 21 mounted on the running vehicle 203 and detects an abnormality.

  When an abnormality of the telephone pole 3 or the like is detected from the captured image, the server 1 gives a predetermined reward to a business operator (imager) who operates the operation vehicle 203 in return for cooperation in maintenance and inspection. The reward may be paid, for example, in cash, or may be paid in other ways. When paying money, for example, the server 1 transmits a remittance request to a financial institution and causes the business operator to perform remittance. As described above, by giving a reward when an abnormality of the telephone pole 3 or the like is detected, it is possible to give an incentive to cooperate with the present system to the business operator.

  In the above description, the camera 21 is installed in the operating vehicle 203 to receive the captured image from the operator who operates the operating vehicle 203, but the present embodiment is not limited to this. For example, the server 1 may acquire an image in which a pedestrian walking on a sidewalk images a utility pole 3 or the like with a mobile terminal such as a smartphone. That is, the server 1 receives the captured image from the pedestrian. For example, an application program for realizing this system is installed in a pedestrian's mobile terminal, and the pedestrian provides the captured image to the server 1 via the application. When an abnormal part is detected from the provided captured image, the server 1 gives a predetermined reward to the pedestrian through the application. In this way, the server 1 is only required to be able to receive a captured image from a predetermined imager and give a reward, and the image capturing subject is not limited to the traveling vehicle 203.

FIG. 9 is a flowchart showing an example of the processing procedure of the abnormality detection processing according to the second embodiment.
The control unit 11 of the server 1 acquires a captured image provided by a predetermined imager (step S201). For example, the control unit 11 acquires a captured image from the operation vehicle 203 that regularly operates on the road on which the utility pole 3 is installed, or the mobile terminal of a pedestrian who walks on a sidewalk. The control unit 11 shifts the processing to step S32.

  After determining that the utility pole 3 or the like is abnormal (S33: YES) and notifying the worker (step S34), the control unit 11 gives a reward to the image capturing person who is the acquisition source of the captured image. Yes (step S205). For example, the control unit 11 gives a reward to a business operator who operates the operating vehicle 203 or a pedestrian who images the utility pole 3 or the like with a mobile terminal. The control unit 11 ends the series of processes.

  As described above, according to the second embodiment, by providing the captured image from the predetermined imager, it is not necessary to prepare a dedicated patrol vehicle, and the present system can be more suitably realized. it can.

  The embodiments disclosed this time are to be considered as illustrative in all points and not restrictive. The scope of the present invention is shown not by the above meaning but by the scope of the claims, and is intended to include meanings equivalent to the scope of the claims and all modifications within the scope.

1 server (information processing device)
11 control unit 12 main storage unit 13 communication unit 14 auxiliary storage unit P program 141 telephone pole DB
142 Anomaly detection model 2 Vehicle 21 Camera 3 Power pole 4 Nesting 5 Tree

Claims (11)

Acquire a captured image of the power distribution equipment installed along the road,
Using a learned model that has been learned to output information indicating whether there is an abnormality in the distribution facility when the captured image is input, detects an abnormality in the distribution facility from the acquired captured image,
An anomaly detection method characterized by causing a computer to execute a process of outputting a detection result. The power distribution equipment is a pole material installed on a pole.
The abnormality detection method according to claim 1, wherein an abnormality of the pillar material is detected using the learned model. The power distribution equipment is a pole material installed on a pole.
The obstacle detection method according to claim 1, wherein an obstacle existing on the pillar material is detected using the learned model. The power distribution equipment is an electric wire erected on a power pole,
The contact detection object which contacts the said electric wire is detected using the said learned model. The abnormality detection method of Claim 1 characterized by the above-mentioned. The abnormality detection method according to any one of claims 1 to 4, wherein the captured image obtained by capturing the distribution facility is acquired from a moving body that moves along the road. The moving body is a patrol vehicle that patrols the road,
The detection result and the position information of the power distribution equipment are output to an in-vehicle device mounted on the patrol vehicle or a terminal device of an operator who performs maintenance work on the power distribution equipment. Abnormality detection method described. Obtaining the captured image provided by a predetermined imager,
The abnormality detection method according to any one of claims 1 to 6, wherein when the abnormality of the power distribution equipment is detected from the captured image, a reward is given to the imager. The abnormality detection method according to any one of claims 1 to 7, wherein the detection result indicating an abnormal location of the power distribution facility and a type of abnormality is output. Acquire a captured image of the power distribution equipment installed along the road,
Using a learned model that has been learned to output information indicating whether there is an abnormality in the distribution facility when the captured image is input, detects an abnormality in the distribution facility from the acquired captured image,
A program that causes a computer to execute a process of outputting a detection result. Acquired teacher data in which a captured image of a power distribution facility installed along a road and information indicating an abnormal location of the power distribution facility are associated with each other,
Generate a learned model that outputs information indicating whether there is an abnormality in the power distribution equipment when the captured image is input based on the acquired teacher data. Generate a learned model that causes a computer to execute processing. Method. An input layer that receives an input of a captured image obtained by capturing an image of power distribution equipment installed along a road,
Outputs from the input layer are combined so as to be input, and based on teacher data in which the captured image and information indicating an abnormal location of the power distribution facility are associated with each other, an image feature amount of the captured image A middle layer that has already learned the parameters used in the arithmetic processing to extract,
An output layer that is coupled so that an output from the middle layer is input, and that outputs information indicating whether or not there is an abnormality in the power distribution equipment based on a calculation result in the middle layer,
A learned model for causing a computer to input the captured image to the input layer, perform arithmetic processing in the intermediate layer, and output information indicating the presence or absence of the abnormality from the output layer.

Images