JP7317264B1 - Risk value calculation device and risk value calculation system - Google Patents

Abstract

The risk value calculation device (4) includes a communication unit (410) that acquires imaging information including information on an image of the worker (6) from the imaging device (3), and an unsafe action performed by the worker. Stored in a behavior analysis unit (430) that identifies the content based on imaging information, and a risk value storage unit (471) that stores calculation elements for calculating risk values for assumed harm in association with assumed harm. a risk value correction unit (470) that calculates a new risk value by correcting the risk value by performing a correction process according to the content of the unsafe behavior performed by the worker on the calculated elements that have been calculated; It has

Description

This disclosure relates to a risk value calculation device and a risk value calculation system for calculating a risk value at a work site.

At a work site, a situation may occur in which a worker is injured, such as a hand being caught in a machine or a head being hit by a falling object.

Therefore, in order to ensure the safety of workers, it has been proposed to conduct risk assessments on machines installed at work sites. Risk assessment is to identify the danger in the work, estimate the risk by combining the severity of the danger and the probability of occurrence of the danger, and decide the priority of countermeasures based on the magnitude of the risk. It is a series of methods for considering measures to eliminate or reduce risks and recording the results. In addition, a technique has been proposed that utilizes information obtained at the work site to determine the safety level of the worker.

JP 2021-064257 A JP 2016-157357 A

Japanese Patent Laid-Open No. 2002-200001 detects the biological information of the worker such as the body temperature and the complexion of the worker, and the environmental information related to the working environment of the worker, and determines the safety level of the worker based on the biological information and the environmental information. Also, based on the determined safety level, the control of the machine operated by the operator is changed.

Furthermore, Patent Document 2 acquires flow line data such as the position of the center of gravity of the worker, the position of the head, and the joints on the premise that the same work is repeatedly performed, and determines whether the work situation of each worker is normal or abnormal. It automatically determines whether there is However, in any of the documents, the judgment of the safety level according to the characteristics of the worker such as habits and character is not performed.

Therefore, the present disclosure has been made in view of the above problem, and aims to calculate a risk value according to the characteristics of a worker.

This disclosure assumes a communication unit that acquires imaging information including an image of a worker from an imaging device, and a behavior analysis unit that identifies the content of unsafe behavior performed by the worker based on the imaging information. For the calculation elements stored in the risk value storage unit that stores the calculation elements for calculating the risk value for harm in association with the assumed harm, according to the content of the unsafe behavior performed by the worker and a risk value correction unit that calculates a new risk value by correcting the risk value by performing the correction process described above.

A risk value calculation system according to the present disclosure can calculate a risk value according to a worker's characteristics.

Configuration diagram of the risk value calculation system according to the first embodiment Functional block diagram of the risk value calculation system according to the first embodiment Functional block diagram of the risk value correction unit according to the first embodiment FIG. 4 is a diagram illustrating an unsafe behavior database according to the first embodiment; FIG. FIG. 2 is a diagram illustrating a factory information database according to the first embodiment; FIG. FIG. 2 is a diagram illustrating an example of a worker database according to the first embodiment; FIG. FIG. 2 is a diagram illustrating a risk assessment database according to the first embodiment; FIG. 4 is a flowchart showing processing of the worker equipment according to the first embodiment; 4 is a flowchart showing processing of the imaging device according to the first embodiment; 4 is a flowchart showing processing of the control device according to the first embodiment; 4 is a flowchart showing processing of the worker device input unit according to the first embodiment 4 is a flowchart showing processing of a behavior analysis unit according to the first embodiment; 4 is a flowchart showing processing of a worker identifying unit according to the first embodiment; 4 is a flowchart showing processing of an unsafe behavior analysis unit according to the first embodiment; Flowchart showing processing of the risk value correction unit according to the first embodiment Importance ratio correspondence table for severity and probability of occurrence according to the first embodiment Correspondence table of unsafe behavior risk level and severity correction amount according to the first embodiment Correspondence Table of Evaluation Information and Occurrence Possibility Correction Amount According to Embodiment 1 4 is a flowchart showing processing of a control unit according to the first embodiment; Hardware configuration diagram of worker equipment according to the first embodiment 1 is a hardware configuration diagram of a control device according to a first embodiment; FIG. Functional block diagram of the learning device according to the second embodiment Flowchart relating to learning processing of the learning device according to the second embodiment Functional block diagram of the inference device according to the second embodiment Flowchart relating to inference processing of the inference apparatus according to the second embodiment Functional block diagram of the risk value calculation system according to the modification of the first embodiment

Embodiments of the present disclosure will be described below with reference to the drawings. The same reference numerals in the figures represent the same or corresponding parts.

Embodiment 1.
FIG. 1 is a configuration diagram of a risk value calculation system 1 according to Embodiment 1 of the present disclosure. The risk value calculation system 1 is composed of a worker device 2 worn by a worker 6 , an imaging device 3 installed at a work site, a machine 5 and a control device 4 . The control device 4 corrects the risk value from the information acquired from the worker device 2 and the imaging device 3 and controls the machine 5 .

The worker device 2 is a device worn by the worker 6, and includes worker information that uniquely identifies the worker 6, position information indicating the current position of the worker, and time when the position information is acquired. Communicate information with the control device 4 . The time information that is the time when the position information is acquired is the first time information. An ID (identifier) for unique identification is used for the worker information. The position information uses latitude and longitude values measured by a GPS (Global Positioning System) that specifies the coordinates of the current position on the earth. For the time information, information delivered by an SNTP (Simple Network Time Protocol) server provided at the work site is synchronized with the time set inside the worker device 2, and the time after synchronization is used. The worker device 2 may be a device that realizes the above, and is realized by, for example, a transmitter.

An example of the machine 5 is a machine tool for cutting. Wireless communication is performed between the worker device 2 and the control device 4 . Communication is performed between the imaging device 3 and the control device 4 either by wire or by radio. Communication between the control device 4 and the machine 5 is either wired or wireless.

FIG. 2 is a functional block diagram of the risk value calculation system 1 according to Embodiment 1 of the present disclosure. The worker device 2 is composed of a processing unit 110 and a communication unit 100 . The processing unit 110 performs processing for specifying the current position of the worker 6 using GPS, and acquires the position information of the worker 6 . The processing unit 110 also acquires worker information and time information of the worker 6 . The processing unit 110 also creates a communication command for the communication unit 100 to transmit to the communication unit 420 of the control device 4 .

The communication unit 100 transmits the position information of the worker 6 , the worker information of the worker 6 and the time information to the communication unit 420 of the control device 4 .

The imaging device 3 is composed of an imaging unit 200 and a communication unit 210 . The imaging unit 200 captures a moving image or still image of a work site including a worker, and creates imaging information in which the captured moving image or still image is associated with the date and time when the image was captured. A moving image or still image of the work site including the worker 6 is information of an image of the worker 6 . The communication section 210 transmits imaging information to the communication section 420 of the control device 4 .

The machine 5 is composed of a control section 300 and a communication section 310 . The communication unit 310 receives control values from the communication unit 420 of the control device 4 . The machine 5 has a movable part (not shown), and the controller 300 controls the movable part based on control values received from the control device 4 to physically process an object.

The control device 4 is composed of a risk value calculation section 400 , a control section 410 and a communication section 420 . The control device 4 is a risk value calculation device. The communication unit 420 acquires worker information, position information and time information from the communication unit 100 of the worker device 2 and acquires imaging information from the communication unit 210 of the imaging device 3 . The communication unit 420 transmits worker information, position information, time information, and imaging information to the risk value calculation unit 400 . Also, the communication unit 420 transmits the control value changed by the control unit 410 to the communication unit 310 of the machine 5 .

Although the processing in the risk value calculation unit 400 will be described later, the control unit 410 changes the control value based on the risk calculated by the risk value calculation unit 400. FIG. The means for controlling the machine 5 includes sequence control for operating the moving parts in order, motion control for controlling the number of rotations and rotation speed of the driving parts, and safety control for ensuring the safety of the operator 6. Examples include, but are not limited to. The control unit 410 changes the control values for controlling the machine 5 and transmits them to the communication unit 420 .

FIG. 3 is a functional block diagram of the risk value calculator 400 according to the first embodiment of the present disclosure. The worker information, position information and time information received from the communication unit 420 are sent to the worker device input unit 440 . The imaging information received from the communication unit 420 is sent to the behavior analysis unit 430 .

The behavior analysis unit 430 accesses the unsafe behavior database 431 using the imaging information as key information, details of which will be described later. Then, if there is a corresponding unsafe behavior, behavior information is acquired. The action information includes the content of the unsafe action corresponding to the imaging information and the corresponding unsafe action risk. That is, the behavior analysis unit 430 identifies the content of the unsafe behavior performed by the worker and the degree of risk corresponding to the content of the unsafe behavior performed by the worker based on the imaging information.

An unsafe action is an action that violates rules established to ensure the safety of a work site, or an action that leads to an accident that is intended to be prevented by the rule. In addition, unsafe behavior includes accidents such as collisions with machinery, injury due to being caught or caught in moving parts of machinery, falling from high places, and health hazards due to inhalation of hazardous substances. It is a behavior that becomes a factor that leads to Alternatively, an unsafe action is an action that is unsafe and unhealthy to be performed.

When the behavior analysis unit 430 acquires the behavior information from the unsafe behavior database 431 , the behavior analysis unit 430 associates the behavior information with the time information that is the time when the imaging device 3 captured the image, and transmits the behavior information to the worker identification unit 450 . The time information sent from the behavior analysis unit 430 to the worker identification unit 450 is second time information.

The unsafe behavior database 431 will now be described. FIG. 4 is a diagram illustrating an unsafe behavior database according to the first embodiment of the present disclosure; The unsafe behavior database 431 stores, for each unsafe behavior, image/moving image data in which images and moving images of the unsafe behavior are recorded, contents of the unsafe behavior for uniquely identifying the image/moving image data, and information about the unsafe behavior. This is a database in which unsafe behavior risk levels indicating risk levels are linked and stored. Contents of unsafe actions include unsafe actions that the worker performs.

The unsafe behavior database 431 is an unsafe behavior storage unit. That is, the unsafe behavior storage unit stores the content information indicating the content of the unsafe behavior and the image analysis data for identifying the unsafe behavior by image analysis from the imaging information in association with each other.

Here, as the content information, the specific content of the unsafe behavior is saved, but it is not limited to this. For example, it may be a number or a symbol that can specify the specific content of the unsafe behavior. In short, any information may be used as long as the content of the specific action of the unsafe action can be identified. Note that the content information may be different for each worker, because the unsafe actions that the worker performs differ for each worker depending on the worker's habits and character.

Images or videos documenting unsafe behavior are image analysis data. Here, the image analysis data may be other than the images and videos recording the unsafe behavior actually performed. For example, it may be an image or moving image created using an information processing device such as a computer. In short, any information may be used as long as it is information necessary for identifying unsafe behavior from images or moving images captured by the imaging device 3 . Common image analysis is used to identify unsafe behavior.

No. in FIG. column stores a number that uniquely identifies the content of the unsafe behavior. The contents of the unsafe behavior are not limited to those illustrated in FIG. 4 (not wearing a helmet, wearing gloves near the machine 5 in operation, walking in a section where walking is prohibited). Other examples include not wearing safety belts while working at heights and not wearing dust masks to prevent inhalation of particulate matter into the body.

The "image/video collection" column stores images and video data of unsafe behaviors. In the image/video collection, a plurality of images or videos may be provided for the same unsafe behavior in order to improve the determination accuracy. For example, No. If the helmet is not worn in 1, images/videos are taken from the front, left, right, and rear when the helmet is not worn, and the images/videos are stored. As a result, the accuracy of identifying whether the helmet is not worn is improved compared to the case where only the image/moving image from one direction when the helmet is not worn is stored.

The unsafe behavior risk level is information that quantifies the risk level of unsafe behavior. Note that the risk level is not a universal value, but a value evaluated and set by the user. In this embodiment, the higher the numerical value, the higher the risk, and the lower the value, the lower the risk. For example, in the case of a collision with a machine, a bruise or bruise will only result in a minor injury, so the degree of danger is set to be low. However, if the body is caught or entangled in the machine, resulting in the amputation of a hand or arm, or in the case of death due to falling from a high place, etc., the risk is set high.

Returning to FIG. 3 , worker identification unit 450 receives worker information and time information from worker device input unit 440 . The time information received from the worker device input unit 440 is the first time information. The worker identification unit 450 receives the behavior information and the second time information linked to the behavior information from the behavior analysis unit 430, and determines the behavior information and the work based on the first time information and the second time information. Person information is linked and transmitted to the frequency analysis unit 460 . That is, the worker identification unit 450 identifies the worker included in the image from the first time information and the second time information.

The worker device input unit 440 accesses the factory information database 441 from the received worker information, position information and time information, using the position information as key information. Although the details will be described later, using the factory information database 441 , the worker device input unit 440 identifies the corresponding factory information from the position information and transmits it to the frequency analysis unit 460 . The factory information is information indicating an installation in the factory and the position of the installation.

Here, the factory information database 441 will be explained. FIG. 5 is a diagram illustrating a factory information database according to the first embodiment of the present disclosure; The factory information database 441 is a database in which factory installations and their positions are stored in association with each other. The installation column stores information identifying items such as machines, corridors, and other items installed in the factory or related to the layout of the factory.

The position information (longitude) and position information (latitude) columns store longitude and latitude information indicating the position of the installed object. If it is desired to set the longitude and latitude within a certain range, such as for a passageway or a large machine, the data are entered in two columns.

The worker device input unit 440 accesses the factory information database 441 using the position information as key information, and specifies the factory information corresponding to the position information. Identification of the factory information is performed by identifying an installed object existing within a threshold range from the coordinates included in the position information. In this case, the threshold range is set at any distance, such as from a few centimeters to several meters. Also, the factory information to be acquired may be two or more depending on the result of the above identification.

Note that if there is only one installation subject to risk assessment, it is possible to omit the factory information database. In that case, the worker device input unit 440 determines whether the position information of the worker and the position information of one preset installation object are close, and if they are close, the factory information of the installation object is sent to the frequency analysis unit 460. Just send it.

Returning to FIG. 3, the frequency analysis unit 460 accesses the worker database 461 using the action information, time information, and worker information received from the worker identification unit 450 as key information. The time information may be either the first time information or the second time information. Although details will be described later, the frequency analysis unit 460 acquires action frequency information from the worker database 461 . Based on the action frequency information acquired from the worker database 461, the frequency analysis unit 460 determines whether the worker intentionally or negligently performed the unsafe action. The frequency analysis unit 460 transmits evaluation information, which is the determined result, to the risk value correction unit 470 together with worker information, factory information, and action information.

Here, the worker database 461 will be described. FIG. 6 is a diagram illustrating a worker database according to the first embodiment of the present disclosure; The worker database 461 is a database that records, for each worker 6 working in the factory, the content of unsafe behavior and the behavior frequency, which is the frequency of unsafe behavior performed in the past. Action frequency information is frequency information. The worker ID column stores worker IDs, which are worker information. The name column stores the name of the worker corresponding to the worker ID. Unsafe behavior no. and contents of unsafe behavior, the No. stored in the unsafe behavior database 431 is displayed. and the contents of the unsafe behavior corresponding to it are stored.

Note that the worker database 461 is a worker storage unit. That is, the worker storage unit stores content information of unsafe actions and frequency information, which is the frequency at which the worker performs the action indicated by the content information, in association with each other.

In addition, although action frequency information is made into the frequency which performed unsafe action in the past here, it is not restricted to this. For example, it may be the frequency estimated by simulation or the like according to the characteristics of the worker, instead of the frequency at which the worker actually performed the unsafe behavior in the past. That is, frequency information (information in which time information and the number of times are associated with each other) according to the characteristics of the worker may be used.

As the action frequency, four types of values are stored according to the past period for which the number of unsafe actions is counted. The total number of unsafe actions performed after the start of operation of the system is stored in the column of the total number of times unsafe actions have been performed. The number of unsafe behaviors performed in the past week, the number of unsafe behaviors performed in the past month, and the number of unsafe behaviors performed in the past year store the number of times they have been performed in the past week, month, and year. It is

The frequency analysis unit 460 accesses the worker database 461 using the worker ID included in the worker information and the content of the unsafe behavior included in the action information as key information. Then, the frequency analysis unit 460 acquires how many times the person associated with the worker ID has performed the unsafe behavior included in the behavior information in the past year, month, and year. That is, the frequency analysis unit 460 acquires frequency information corresponding to the details of the unsafe actions performed by the worker from the worker storage unit. Specifically, the frequency analysis unit 460 acquires action frequency information corresponding to worker information, and links the worker information and the action frequency information.

Returning to FIG. 3, the risk value correction unit 470 accesses the risk assessment database 471 using the received factory information as key information. The risk value correction unit 470 acquires from the risk assessment database 471 the severity and probability of occurrence corresponding to the installation specified in the factory information. The risk value correction unit 470 performs a correction process for correcting the value indicating the severity or the value indicating the probability of occurrence based on the information indicating whether the severity or the probability of occurrence is emphasized, and corrects the risk value. The resulting new risk value is linked with the factory information and transmitted to the control unit 410 . Information indicating which of severity and probability of occurrence is emphasized is registered in advance by the user.

In general, the timing of risk assessment is when a work site is constructed or when a machine is introduced to the work site. In doing so, severity and likelihood are set as quantitative values. A risk value is generally calculated quantitatively by multiplying or adding these two elements (severity and likelihood). The risk assessment database 471 stores results of risk assessment as a database.

Here, the risk assessment database 471 will be explained. FIG. 7 is a diagram illustrating a risk assessment database according to the first embodiment of the present disclosure; The risk assessment database 471 contains all the results of risk assessments performed on each worksite installation. Note that the risk assessment database 471 is a risk value storage unit.

The column of risk assessment No stores a number that uniquely identifies which row of risk assessment is indicated. The factory information No. column stores numbers corresponding to the factory information numbers of the factory information database 441 for installations to be subjected to risk assessment. The risk assessment number is determined using the factory information number as a key. For example, the risk assessment number corresponding to factory information number 0001 is No. 00010, No. 00011.

The severity and probability of occurrence columns store numerical values of the degree of seriousness and probability of occurrence of the possible harm, respectively. The risk value column stores risk values calculated by combining severity and likelihood. That is, the risk assessment database 471 stores calculation elements for calculating risk values for assumed harm in association with assumed harm.

Note that the risk value correction unit 470 adjusts the content of the unsafe behavior performed by the worker to the risk value calculation element (the value indicating the severity) based on the degree of danger of the unsafe behavior of the worker 6. Correct the risk value by performing correction processing accordingly. Further, the risk value correction unit 470 calculates the content of the unsafe behavior performed by the worker for the calculation element of the risk value (the value indicating the possibility of occurrence) based on the implementation frequency of the unsafe behavior of the worker 6. Correct the risk value by performing correction processing according to That is, the risk value correction unit 470 determines the risk value calculation element (the value indicating the severity or the value indicating the possibility of occurrence) according to the content of the unsafe behavior performed by the worker identified by the behavior analysis unit 430. By correcting, a new risk value obtained by correcting the risk value is calculated.

The risk value correction unit 470 acquires a value indicating severity as a calculation element from the risk value storage unit, and performs a correction process of correcting the value indicating severity based on the degree of risk corresponding to the imaging information to create a new risk. Calculate the value. The risk value correction unit 470 obtains a value indicating the probability of occurrence as a calculation element from the risk value storage unit, and performs a correction process of correcting the value indicating the probability of occurrence based on the obtained frequency information to obtain a new risk value. Calculate The risk value correction unit 470 associates the corrected new risk value with the factory information and transmits it to the control unit 410 .

In addition, the risk value correction unit 470 corrects the calculation elements of the risk value (the value indicating the severity and the value indicating the probability of occurrence) according to whether importance is placed on the severity or the probability of occurrence, Modify the risk value. The risk value correction unit 470 associates the corrected risk value with the factory information and transmits it to the control unit 410 .

The reason for modifying the severity value according to the degree of danger is that the severity may be greater than assumed in the risk assessment. For example, when conducting a risk assessment when introducing a machine, assume the risk value when the machine collides with the worker wearing a helmet. However, during actual work, unsafe behavior without wearing a helmet may occur due to the worker's characteristics such as habits and personality, and the severity is greater than assumed in the risk assessment.

Also, the reason for correcting the value indicating the probability of occurrence according to the implementation frequency is that the probability of occurrence may be higher than assumed in the risk assessment. For example, the frequency of unsafe behavior changes depending on the worker's characteristics such as habits and personality. For example, there are workers who think that there is no problem in violating some of the rules established at the work site, while there are other workers who think that any rules must not be violated.

In the former case, the worker may intentionally take unsafe actions, and the frequency of unsafe actions increases. In the latter case, the operator takes unsafe actions only when he or she is unintentionally negligent, so the implementation frequency is low. Therefore, if there are workers who frequently perform unsafe actions, the probability of occurrence is higher than when the risk assessment was assumed.

Moreover, even at the same work site, the risk value to be assumed at each site differs depending on which of the degree of danger and the frequency of implementation is emphasized. For example, at a work site where there are relatively many unsafe actions, it is necessary to set the risk value to a value that emphasizes the implementation frequency in order to reduce the number of unsafe actions. In a work site where unsafe behavior rarely occurs, the risk value is set so as to emphasize both the degree of danger and the frequency of implementation. In addition, at a work site where there are relatively many occurrences of harm, the risk value is calculated so as to emphasize the degree of danger.

In this way, the risk value correcting unit 470 can calculate a risk value according to the policy of the work site regarding which of the degree of risk and the frequency of implementation should be emphasized in addition to the worker's characteristics such as habits and personality.

Next, the flow of calculating the risk value between the worker device 2, the imaging device 3, the machine 5, and the control device 4 will be described.

FIG. 8 is a flowchart showing processing of the worker device 2 according to the first embodiment. The processing unit 110 of the worker's device 2 acquires current position information by GPS. Further, the processing unit 110 acquires worker information input by the worker 6 (step S11). Next, the processing unit 110 of the worker device 2 acquires the set time information, associates the worker information, the position information and the time information, and transmits them to the communication unit 100 (step S12).

The time when the worker 6 was at the location represented by the position information can be grasped from the time information linked to the position information. Next, the communication unit 100 transmits the associated worker information, position information, and time information to the control device 4 (step S13).

FIG. 9 is a flowchart showing processing of the imaging device 3 according to the first embodiment. The image capturing unit 200 of the image capturing device 3 captures moving images or still images of the work site. Further, the imaging unit 200 associates a moving image or a still image shot by referring to the time information set in the imaging device 3 with the time information when the moving image or the still image was shot (step S21). The time information at which the moving image or still image was shot is the second time information. In the case of moving images, time information is linked while shooting. Next, the communication unit 210 of the imaging device 3 transmits imaging information including the captured moving image or still image and associated time information to the control device 4 (step S22).

FIG. 10 is a flowchart showing processing of the control device 4 according to the first embodiment. The communication unit 420 of the control device 4 receives the linked worker information, position information, and time information from the worker device 2 and receives the imaging information from the imaging device 3 . The communication unit 420 of the control device 4 confirms the transmission source of the received information and determines whether the information is received from the worker device 2 . The communication unit 420, if received from the worker device 2, transmits the received information to the worker device input unit 440 (step S31). If the control unit 410 has received the information from the imaging device 3, the control unit 410 transmits the received information to the behavior analysis unit 430 (step S31).

Next, worker device input section 440 transmits worker information to worker identification section 450 . Also, the worker device input unit 440 accesses the factory information database 441 and transmits the factory information to the frequency analysis unit 460 (step S32). Next, the behavior analysis unit 430 performs processing such as accessing the unsafe behavior database 431 (step S33). The worker identification unit 450 identifies the worker 6 who was acting unsafely (step S34). The details of step S34 will be described later.

The frequency analysis unit 460 accesses the worker database 461, acquires and analyzes the past unsafe behavior frequency of the worker 6, and analyzes the evaluation information, which is the analysis result, as worker information, factory information, and behavior. Together with the information, it is transmitted to the risk value correction unit 470 (step S35). The details of step S35 will be described later. The risk value correction unit 470 corrects the risk value calculation elements (severity and occurrence probability values in the risk assessment database) based on the degree of risk and frequency of unsafe behavior of the worker 6, thereby reducing the risk Correct the value. Then, the modified risk value is transmitted to the control unit 410 (step S36).

The details of step S36 will be described later. The control unit 410 modifies the current control value according to the modified risk value and transmits the control value to the machine 5 (step S37).

FIG. 11 is a flowchart showing processing of the worker device input unit 440 according to the first embodiment. FIG. 11 shows details of the processing in step S32 of FIG. The worker device input unit 440 receives worker information, position information and time information from the worker device 2 (step S321). The worker device input unit 440 accesses the factory information database 441 using the position information as key information (step S322). The worker device input unit 440 acquires factory information corresponding to the position information from the factory information database 441 and transmits it to the frequency analysis unit 460 (step S323). The worker device input unit 440 transmits the worker information and the time information to the worker identification unit 450 (step S324).

FIG. 12 is a flowchart showing processing of the behavior analysis unit 430 according to the first embodiment. FIG. 12 shows details of the processing in step S33 of FIG. The behavior analysis unit 430 receives imaging information from the communication unit 420 (step S331). The behavior analysis unit 430 accesses the unsafe behavior database 431 using the imaging information as key information (step S332). The behavior analysis unit 430 compares the imaging information with information recorded in the unsafe behavior database 431, and determines whether unsafe behavior has occurred (step S333).

Specifically, the imaging information is analyzed, and if an image/moving image similar to the image/moving image collection in the unsafe behavior database 431 is included in the imaging information, an unsafe behavior corresponding to the image/moving image collection occurs. It is determined that The unsafe behavior risk level at that time is the value of the unsafe behavior risk level corresponding to the image/video collection. For example, No. of the unsafe behavior database 431 . If an image/moving image similar to the image/moving image collection indicated by 1 is included in the imaging information, it is determined that the helmet is not worn. In this case, the unsafe behavior risk level acquired from the unsafe behavior database 431 is "8".

If it is determined that unsafe behavior has occurred (step S333 Yes), behavior information is received from the unsafe behavior database, and time information included in the imaging information is associated with the behavior information (step S334). The behavior information is information composed of the details of the unsafe behavior performed by the worker and the degree of risk of the unsafe behavior. Next, the behavior analysis unit 430 transmits the linked behavior information and time information to the worker identification unit 450 . (Step S335)

On the other hand, if it is not determined that an unsafe action has occurred (step S333 No), the process ends at step S333. After that, the process proceeds to step S34 in FIG. 10, but since the action information and the time information are not output from the action analysis unit 430, nothing is actually performed.

FIG. 13 is a flowchart showing processing of the worker identifying unit 450 according to the first embodiment. FIG. 13 shows details of the processing in step S34 of FIG. The worker identification unit 450 receives worker information and time information from the worker equipment input unit 440 (step S341). The worker identification unit 450 receives the linked behavior information, unsafe behavior risk level, and time information from the behavior analysis unit 430 (step S342).

Based on the time information received from the worker device input unit 440 and the time information received from the behavior analysis unit 430, the worker identification unit 450 links the time information, the worker information, and the behavior information (step S343). This makes it possible to grasp which worker 6 took what unsafe action when. The worker identification unit 450 transmits the linked worker information, action information, and time information to the frequency analysis unit 460 (step S344).

FIG. 14 is a flowchart showing processing of the frequency analysis unit 460 according to the first embodiment. FIG. 14 shows the details of the processing in step S35 of FIG. The frequency analysis unit 460 receives worker information, action information, and time information from the worker identification unit 450 (step S351). The frequency analysis unit 460 receives factory information from the factory information database 441 (step S352). The frequency analysis unit 460 accesses the worker database 461 using the received worker information and behavior information as key information (step S353). If time information is also used as key information, it is possible to update the total number of unsafe actions in the worker database 461 and the number of times they have been performed in the past week, month, or year.

The frequency analysis unit 460 acquires the action frequency of the worker 6 from the worker database 461 (step S354). The frequency analysis unit 460 compares the acquired action frequency with a threshold of action frequency set for each unsafe action, and if the frequency of the unsafe action exceeds the threshold, the worker 6 performs the unsafe action. shall be deemed to have been carried out with the intention of taking Note that the threshold value provided for each unsafe action is set in advance.

If the frequency of the unsafe behavior does not exceed the threshold, it is considered that the worker 6 did not intend to take the unsafe behavior, and that he/she did it by mistake. In this way, the frequency analysis unit 460 calculates evaluation information indicating whether the worker 6 intentionally committed unsafe behavior or was negligent (step S355).

The frequency analysis unit 460 transmits the worker information, factory information, behavior information and evaluation information to the risk value correction unit 470 (step S356).

FIG. 15 is a flowchart showing processing of the risk value correction unit 470 according to the first embodiment. FIG. 15 shows details of the processing in step S36 of FIG. The risk value correction unit 470 receives worker information, factory information, behavior information and evaluation information from the frequency analysis unit 460 (step S361). The risk value correction unit 470 determines whether the risk value calculation elements (values indicating severity and probability of occurrence) corresponding to the received factory information have been acquired from the risk assessment database 471 (step S362).

For the received factory information, if calculation elements for all possible harms have been received, a new corrected risk value is transmitted to the control unit 410 . (Step S362 Yes)

If the severity and probability of occurrence of all possible harms have not been received for the received factory information (step S362 No), the risk assessment database 471 is accessed (step S363). The risk value correction unit 470 acquires the severity and probability of occurrence for the installed object corresponding to the factory information (step S364). Steps S365 to S368 are performed for each factory information.

The risk value correction unit 470 sets which of seriousness and probability of occurrence is to be emphasized (step S365). This processing is performed by the user selecting an arbitrary combination from combinations of values of severity and probability of occurrence stored in advance.

FIG. 16 is an emphasis ratio correspondence table of severity and probability of occurrence according to the first embodiment. The values of this correspondence table are registered in advance by the user of the risk value calculation system 1 . The respective ratios are determined so that the sum of severity and likelihood is 1. Here, if you want to calculate a risk value that emphasizes the degree of risk of unsafe behavior, the severity is larger than the probability of occurrence. 4 or No. Choose one of 5. If you want to calculate a risk value that emphasizes the probability of occurrence, the probability of occurrence is greater than the severity. Choose one from 1 to 3. No. Which one of 1 to 3 should be selected may be determined depending on how much importance is placed on the probability of occurrence.

If you want to calculate a risk value that emphasizes both, the No. where the severity and the probability of occurrence have the same value. Select 3. These selections are made by the user of the risk value calculation system 1 . In this embodiment, the combination of severity and occurrence probability is selected from pre-stored combinations, but the present invention is not limited to this. For example, the user may enter arbitrary values.

Next, the risk value correction unit 470 corrects the value indicating the severity based on the unsafe behavior risk included in the behavior information (step S366). The reason for correcting the value indicating the severity based on the risk of unsafe behavior is that the severity may be greater than the risk assessment assumption (assumed when creating the risk assessment database 471) as described above. be. As the value of the degree of risk of unsafe behavior increases, it is considered that the risk of serious damage to the worker 6 is high, and correction is made to increase the value of severity.

FIG. 17 is a correspondence table of unsafe behavior risk levels and severity correction amounts according to the first embodiment. This correspondence table is registered in advance by the user of the risk value calculation system 1 . The risk value correction unit 470 acquires the unsafe behavior risk level from the behavior information received from the frequency analysis unit 460, and acquires the severity correction amount based on the correspondence table of the correction amount of risk and severity. The risk value correction unit 470 multiplies or adds the severity acquired from the risk assessment database 471 and the severity correction amount.

Next, the risk value correction unit 470 corrects the value indicating the probability of occurrence (step S367). FIG. 18 is a correspondence table of evaluation information and occurrence probability correction amounts according to the first embodiment. This correspondence table is registered in advance by the user of the risk value calculation system 1 . The risk value correction unit 470 acquires the correction amount of occurrence probability corresponding to the evaluation information received from the frequency analysis unit 460 and the correspondence table of the evaluation information and the occurrence possibility correction amount. The risk value correction unit 470 corrects the value of the probability of occurrence so that when the evaluation information is "intentional" the value of the probability of occurrence is greater than when the evaluation information is "accidental". Specifically, the occurrence probability acquired from the risk assessment database 471 is added or multiplied to the occurrence probability correction amount acquired in FIG.

That is, the risk value is greater when the evaluation information is "intentional" than when it is "accidental". Next, the reason is described.

If the evaluation information is "intentional", unsafe behavior is considered to be occurring regularly. If this is done on a regular basis, the worker 6 will not recognize that unsafe behavior is dangerous, and it will be difficult for him or her to take action to avoid a dangerous situation when it actually occurs. . On the other hand, if the evaluation information is "negligence," it means that the work was done by mistake, and the worker 6 can be considered to be able to take avoidance action when a dangerous situation occurs. Therefore, the risk value is greater when the evaluation information is "intentional" than when it is "accidental".

Next, the risk value correction unit 470 calculates the risk value by multiplying the corrected severity and corrected probability of occurrence by the values selected in step S365. (Step S368) Here, the corrected severity, the corrected probability of occurrence, and the importance ratio of severity and probability of occurrence are multiplied. good. In short, based on the importance ratio (importance) of the severity and the probability of occurrence, the amount of correction of the severity and the probability of occurrence is weighted to correct the risk value.

That is, the risk value correction unit calculates a new risk value by correcting the risk value in the risk assessment database 471 based on the calculation elements for calculating the risk value for the assumed harm in the risk assessment database 471 .

After calculating the risk value in S368, the process returns to S362 to determine whether all the risk assessment results for the factory information have been acquired. If all risk assessment results have been obtained, the risk value is transmitted to the control unit (step S369).

FIG. 19 is a flow chart showing processing of the control unit 410 . FIG. 19 shows details of the processing in step S369 of FIG. Control unit 410 receives the factory information and the risk value from risk value correction unit 470 (step S381). The control unit 410 identifies the machine corresponding to the factory information, and determines whether or not the correction of the control value of the machine has been completed. (Step S382 Yes). If the correction has been completed, the process relating to the correction of the risk value is terminated, and if the correction has not been completed, the process proceeds to step S384.

In step S384, the modified risk value is compared with a preset threshold, and if the modified risk value is greater than the threshold, the control value of machine 5 is changed. For example, changes such as stopping control of the machine 5 or decelerating motion control are assumed, but are not limited to these.

Control unit 410 repeats steps S382 and S384 until all comparisons between the preset threshold value and the corrected risk value are completed.

FIG. 20 is a hardware configuration diagram of worker equipment 2 according to the first embodiment of the present disclosure. The worker device 2 has a processor 11 for specifying the coordinates of the current position and processing wireless communication using GPS, a memory 12 for storing various programs and various information, and a communication interface 13 for processing wireless communication with the control device 4. . Various programs in the memory 12 include programs including the processing of FIG. Memory 12 and communication interface 13 are connected to processor 11 via bus 14 and communicate with processor 11 respectively.

Memory 12 includes non-volatile memory and volatile memory. The memory 12 also stores worker information, position information, and time information of the worker 6 to whom the worker device 2 is attached. Further, the memory 12 temporarily stores results being processed by the processor 11 .

The communication interface 13 is used for wireless communication with the control device 4 . The communication interface 13 wirelessly transmits the processing result stored in the memory 12 to the control device 4 under the control of the processor 11 .

The communication unit 100 is implemented by the processor 11 and the memory 12 of the worker's device 2 . The processing unit 110 is implemented by the processor 11 and memory 12 of the worker device 2 and the communication interface 13 .

FIG. 21 is a hardware configuration diagram of the control device 4 according to the first embodiment. The control device 4 includes a memory 41 for storing a user program for controlling the machine 5 created by the user, a communication interface 42 for wireless or wired communication with the worker device 2 and the imaging device 3, and the entire control device 4. and a processor 43 that controls the

In addition to the user program, the memory 41 stores information acquired from the worker device 2 and the imaging device 3 and information necessary for the processor 43 to execute the user program. The memory 41 and the communication interface 42 are each connected to the processor 43 via the bus 44 and communicate with each other.

Memory 41 includes non-volatile memory and volatile memory. The memory 41 stores firmware, which is a processing program for functions provided by the control device 4 . Furthermore, the memory 41 temporarily stores results during processing of the user program by the processor 43 . The memory 41 also stores an unsafe behavior database 431 , a factory information database 441 , a worker database 461 and a risk assessment database 471 . However, these databases may be provided outside the control device 4 without being stored in the memory 41 , or may be provided outside the control device 4 only partially.

The communication interface 42 is a communication interface for wireless or wired communication with the worker device 2 and the imaging device 3 . The interface 42 stores received information in the memory 41 and transmits information stored in the memory 41 under the control of the processor 43 .

The processor 43 executes firmware and user programs stored in the memory 41 . Control unit 410 is implemented by processor 43 and memory 41 . Communication unit 420 is implemented by processor 43 , memory 41 , and communication interface 42 . The risk value calculator 400 is implemented by the memory 41 and processor 43 of the control device 4 .

According to the risk value calculation system in Embodiment 1 of the present disclosure, the behavior analysis unit acquires the degree of risk of unsafe behavior according to the unsafe behavior, and the risk assessment results performed on the premise that there is no unsafe behavior is corrected based on the degree of risk acquired by the behavior analysis department, so even if unsafe behavior occurs according to the worker's characteristics, the risk value is calculated according to the characteristics of the worker. can do.

Further, according to the risk value calculation system according to the first embodiment of the present disclosure, the frequency analysis unit acquires the frequency information, which is the frequency at which the worker performs unsafe behavior, and performs it on the premise that there is no unsafe behavior. The probability of occurrence, which is the result of risk assessment, is corrected based on the frequency information obtained by the frequency analysis department. Although the frequency of occurrence of unsafe behavior changes according to the characteristics of the worker, according to the configuration of the present disclosure, it is possible to calculate the risk value according to the characteristics of the worker.

Further, according to the risk value calculation system according to the first embodiment of the present disclosure, the worker information, the position information indicating the position of the worker, and the imaging device acquired from the worker device 2 worn by the worker are acquired. The operator is identified from the captured image information, the frequency information of unsafe behavior is acquired for each identified worker, and the value indicating the probability of occurrence is corrected for each identified worker, so even if there are multiple workers, The risk value can be modified according to the characteristics of each worker.

In addition, in Embodiment 1, it is assumed that there are a plurality of workers 6, but there may be only one worker. In that case, S32 and S34 of FIG. 10 are not necessary for the control device 4 . The process of calculating the risk value may be performed by the risk value calculation unit 400 . Thereby, it is possible to calculate a risk value according to the characteristics of the worker.

In addition, in the present embodiment, the processing unit 110 acquires the current location information by GPS, but the present invention is not limited to this. For example, the position information of the worker 6 may be obtained by analyzing the moving image captured by the imaging device 3 . In this case, the behavior analysis unit 430 transmits the imaging information to the worker identification unit 450, and the worker identification unit 450 identifies the position of the worker by image analysis from the received imaging information.

Embodiment 2.
In Embodiment 1, the worker equipment input unit 440 accesses the factory information database 441 using the worker's position information as key information, and uniquely identifies an installation object that is close to the worker based on the worker's position information. Then, the factory information of the equipment was acquired. However, the installed objects included in the factory information database 441 include passages and other items unrelated to the control of the control unit 410 . Therefore, in the present embodiment, whether or not the installed object is unrelated to control is learned by unsupervised learning of AI (artificial intelligence), and if the installed object is unrelated to control, worker device input unit 440 Avoid getting information.

FIG. 22 is a functional block diagram of learning device 500 according to the second embodiment of the present disclosure. The learning device 500 includes a data acquisition section 510 and a model generation section 520 . The data acquisition unit 510 acquires the distance between the worker and the installed object and the control value of the installed object as learning data. Note that the control value is treated as 0 when the control object is not a passage or the like. Also, the learning device 500 exists within the risk value calculation unit 400 and is connected to the worker equipment input unit 440 and the risk value correction unit 470 .
<Learning phase>

The model generating unit 520 generates the installed object unrelated to the control based on the learning data created based on the combination of the distance between the worker and the installed object and the control value of the installed object output from the data acquisition unit 510 . learn whether or not Here, the information as to whether or not the installed object is irrelevant to the control is referred to as factory information acquisition necessity information. That is, from the distance between the worker and the installed object in the risk value calculation system 1 and the control value of the installed object, a learned model is generated for inferring the necessity information of factory information acquisition. Here, the learning data is data in which the distance between the worker and the installed object and the control value of the installed object are associated with each other. The distance between the worker and the installation object is input from the worker device input unit 440 to the data acquisition unit 510 , and the control value of the installation object is input from the risk value correction unit 470 to the data acquisition unit 510 .

A known algorithm for unsupervised learning can be used as the learning algorithm used by the model generation unit 520 . As an example, a case where the K-means method (clustering), which is unsupervised learning, is applied will be described. Unsupervised learning refers to a method of learning the features of the learning data by providing the learning device 500 with learning data that does not contain results (labels).

The model generation unit 520 learns factory information acquisition necessity information by so-called unsupervised learning according to a grouping method based on the K-means method, for example.

The K-means method is a non-hierarchical clustering algorithm, and is a method of classifying a given number of clusters into K using the average of clusters.

Specifically, the K-means method is processed in the following flow. First, clusters are randomly assigned to each data xi. Next, the center Vj of each cluster is calculated based on the allocated data. Then, find the distance between each xi and each Vj, and reassign xi to the closest central cluster. If the assignment of all xi clusters does not change in the above process, or if the amount of change falls below a predetermined threshold, it is determined that convergence has occurred, and the process ends.

In the present application, the necessity of factory information acquisition is determined by so-called unsupervised learning according to the learning data created based on the combination of the distance between the worker and the installation object acquired by the data acquisition unit 510 and the control value of the installation object. learn information.

The model generation unit 520 generates and outputs a learned model by executing the learning as described above.

The learned model storage unit 530 stores the learned model output from the model generation unit 520 .

Next, with reference to FIG. 23, processing for learning by the learning device will be described. FIG. 23 is a flowchart relating to learning processing of the learning device.

In step S401, the data acquisition unit 510 acquires the distance between the worker and the installed object and the control value of the installed object. Although the distance between the worker and the installed object and the control value of the installed object are acquired at the same time, it is sufficient to input the distance between the worker and the installed object and the control value of the installed object in association with each other. and the installed object, and the data of the control value of the installed object may be obtained at different timings.

In step S402, the model generation unit 520 performs so-called unsupervised learning according to learning data created based on the combination of the distance between the worker and the installation object acquired by the data acquisition unit 510 and the control value of the installation object. learns the factory information acquisition necessity information and generates a learned model.

In step S<b>403 , the learned model storage unit 530 stores the learned model generated by the model generation unit 520 .
<Utilization phase>

FIG. 24 is a functional block diagram of an inference device 600 relating to the risk value calculation system 1. As shown in FIG. The inference device 600 includes a data acquisition unit 510 and an inference unit 610 . The inference device 600 exists within the risk value calculation unit 400 and is connected to the worker equipment input unit 440 .

The data acquisition unit 510 acquires the distance between the worker and the installed object from the worker equipment input unit 440 .

The inference unit 610 infers factory information acquisition necessity information obtained by using the learned model stored in the learned model storage unit. That is, by inputting the distance between the worker and the installation object acquired by the data acquisition unit 510 into this trained model, it is inferred which cluster the distance between the worker and the installation object belongs to, and based on the inference result Factory information acquisition necessity information can be output.

In the present embodiment, the learned model learned by the model generation unit 520 of the risk value calculation system 1 is used to output the factory information acquisition necessity information. A learned model may be acquired from the outside, and the factory information acquisition necessity information may be output based on this learned model.

In this way, the inference unit 610 transmits the factory information acquisition necessity information obtained based on the distance between the worker and the installed object to the worker equipment input unit 440 .

Next, a process for obtaining factory information acquisition necessity information using the inference device 600 will be described with reference to FIG. 25 .

In step S411, the data acquisition unit 510 acquires the distance between the worker and the installed object.

In step S412, the inference unit 610 inputs the distance between the worker and the installed object into the learned model stored in the learned model storage unit 530, and obtains factory information acquisition necessity information.

In step S<b>413 , the inference unit 610 outputs factory information acquisition necessity information obtained from the learned model to the worker equipment input unit 440 .

In step S414, the worker device input unit 440 accesses the factory information database 441 based on the factory information acquisition necessity information. That is, if the factory information acquisition necessity information is "required", the factory information database 441 is accessed to acquire the factory information. If the result of the factory information acquisition necessity information is “no”, the worker device input unit 440 does not access the factory information database 441 . When factory information is acquired, the worker equipment input unit 440 transmits the factory information to the frequency analysis unit 460 . This makes it possible to calculate a risk value limited to factory information related to control.

In this embodiment, the learning device 500 and the inference device 600 are assumed to exist within the risk value calculation unit 400, but the present invention is not limited to this. For example, the learning device 500 and the reasoning device 600 may be separate devices from the control device 4 in the risk value calculation system 1 . Alternatively, it may be a device connected to the risk value calculation system 1 via a network and separate from the risk value calculation system 1 . Furthermore, the learning device 500 and the reasoning device 600 may reside on a cloud server.

In addition, the model generation unit 520 may learn the factory information acquisition necessity information according to learning data created for a plurality of risk value calculation systems 1 . Note that the model generation unit 520 may acquire learning data from a plurality of risk value calculation systems 1 used in the same area, or may acquire learning data from a plurality of risk value calculation systems 1 that operate independently in different areas. The information on whether or not to obtain factory information may be learned by using the collected learning data. It is also possible to add or remove the risk value calculation system 1 that collects the learning data from the target on the way. Furthermore, the learning device 500 that has learned the factory information acquisition necessity information for a certain risk value calculation system 1 is applied to a different risk value calculation system 1, and the factory information acquisition requirement for the different risk value calculation system 1 is applied. The negative information may be re-learned and updated.

(Modification)
In Embodiment 1, as shown in FIG. 3, the risk value calculation section 400 of the control device 4 calculates the risk value, and the control section 410 controls the machine 5 . However, it is not necessary to be limited to these configurations, and an information processing device 7 such as a personal computer that can perform information processing may be configured to include the risk value calculation unit 400 .

26 is a functional block diagram of a system of a modification of Embodiment 1. FIG. The information processing device 7 communicates with the worker device 2, the imaging device 3, and the machine 5 via the communication unit 710, and the risk value calculation unit 400 calculates the risk value. The information processing device 7 then transmits the calculated risk value to the control device 4 . The control device 4 transmits the risk value received by the communication unit 420 to the control unit 410 and controls the machine 5 with the corrected risk value.

Even in this modified example, the same effect as in the first embodiment can be obtained. Note that learning device 500 and reasoning device 600 of Embodiment 2 are applicable to this modification. Also in this case, the same effects as in the second embodiment are obtained.

The configuration shown in the above embodiment shows an example of the content of the present disclosure, and can be combined with another known technique, and the combination of the above embodiment and the gist of the present disclosure It is also possible to omit or change part of the configuration without departing from

1 Risk value calculation system 2 Worker device 3 Imaging device 4 Control device 5 Machine 6 Worker 400 Risk value calculation unit 410 Control unit 420 Communication unit 430 Behavior analysis unit 431 Unsafe behavior database , 440 worker device input unit, 450 worker identification unit, 460 frequency analysis unit, 461 worker database, 470 risk value correction unit 470, 471 risk assessment database, 500 learning device, 510 data acquisition unit, 520 model generation unit, 530 learned model storage unit, 600 reasoning device, 610 reasoning unit.

Claims (11)

a communication unit that acquires imaging information including information of an image of a worker from an imaging device;
a behavior analysis unit that identifies the content of the unsafe behavior performed by the worker based on the imaging information;
Unsafety performed by the worker with respect to the calculation elements stored in the risk value storage unit that stores the calculation elements for calculating the risk value for the assumed harm in association with the assumed harm. A risk value calculation device comprising: a risk value correction unit that calculates a new risk value by correcting the risk value by performing correction processing according to the content of the action. The behavior analysis unit stores the risk of unsafe behavior and image analysis data for identifying unsafe behavior by image analysis from the imaging information in association with content information indicating details of the unsafe behavior. Accessing the behavior storage unit, acquiring the degree of risk corresponding to the content of the unsafe behavior performed by the worker,
The risk value storage unit stores a value indicating the severity of the assumed harm in association with the assumed harm,
The calculation element includes a value indicating the severity, and the correction processing includes processing for correcting the value indicating the severity based on the degree of risk corresponding to the content of the unsafe behavior performed by the worker. The risk value calculation device according to claim 1. Corresponding to the content of the unsafe behavior performed by the worker from the worker storage unit that stores the content information and frequency information that is the frequency of the action indicated by the content information in association with each other. further comprising a frequency analysis unit for acquiring frequency information,
The risk value storage unit stores a value indicating the probability of occurrence of the assumed harm in association with the assumed harm,
The calculated element further includes a value indicating the likelihood of occurrence;
3. The risk value calculation device according to claim 2, wherein the correction process further includes a process of correcting the value indicating the probability of occurrence based on the frequency information. The correction process includes a process of correcting the value indicating the severity or the probability of occurrence based on information registered in advance by the user and indicating which of the severity and the probability of occurrence is emphasized. The risk value calculation device according to claim 3 . Content information indicating the content of unsafe behavior and frequency information indicating the frequency at which the worker performs the behavior indicated by the content information are stored in association with each other from the worker storage unit. further comprising a frequency analysis unit that acquires frequency information corresponding to the content of the action,
The risk value storage unit stores a value indicating the probability of occurrence of the assumed harm in association with the assumed harm,
the calculated element includes a value indicative of the probability of occurrence;
2. The risk value calculation device according to claim 1, wherein the correction processing includes processing for correcting the value indicating the probability of occurrence based on the frequency information. From a worker device that is worn by the worker and acquires worker information that identifies the worker, position information that indicates the position of the worker, and first time information that is the time when the position information is acquired Further comprising a worker identification unit that acquires worker information, position information indicating the position of the worker, and the first time information,
The imaging information includes second time information that is the time when the imaging device captured the image,
The risk value calculation device according to claim 1 or 5 , wherein the worker identification unit identifies the worker included in the image from the first time information and the second time information. The unsafe action is an action that is undesirable in terms of safety and health, an action that violates rules established to ensure safety, or an action that leads to an accident that is intended to be prevented by rules. 6. The risk value calculation device according to any one of 1 to 5 . 2. The risk value calculation device according to claim 1, further comprising a control unit that modifies a control value for controlling an installation at the work site according to the new risk value. a data acquisition unit that acquires the distance between the worker and the installed object and the control value as learning data;
A model generation unit that uses the learning data to generate a learned model for inferring information as to whether or not the installed object is unrelated to the control of the control unit from the distance between the worker and the installed object. and,
The risk value calculation device according to claim 8, further comprising : a data acquisition unit that acquires the distance between the worker and the installed object;
The worker acquired from the data acquisition unit using a trained model for inferring information about whether or not the installed object is unrelated to the control of the control unit from the distance between the worker and the installed object an inference unit that outputs information as to whether or not the installation object is unrelated to the control of the control unit from the distance between the installation object and the installation object;
The risk value calculation device according to claim 8, further comprising : a communication unit that acquires imaging information including information of an image of a worker from an imaging device;
a behavior analysis unit that identifies the content of the unsafe behavior performed by the worker based on the imaging information;
Unsafety performed by the worker with respect to the calculation elements stored in the risk value storage unit that stores the calculation elements for calculating the risk value for the assumed harm in association with the assumed harm. a risk value calculation device having a risk value correction unit that calculates a new risk value by correcting the risk value by performing correction processing according to the content of the action;
the imaging device;
A risk value calculation system with

Images