Safety monitoring device, safety monitoring method, and safety monitoring program

Abstract

To provide a safety monitoring device, a safety monitoring method, and a safety monitoring program that can output unsafe conditions comprehensively and in natural language.SOLUTION: A safety monitoring device 1 is equipped with a determination unit 162 that determines whether a work site in the imaged state is safe from the captured image data input from an acquisition unit 161, and makes the determination using an AI model that outputs information related to unsafe factors and information related to accident prevention measures when the work site is not in a safe state, and an output control unit 163 that, when the determination unit 162 determines that the work site is unsafe, outputs to the outside a message that the work site is in an unsafe state and the information related to unsafe factors and accident prevention measures output by the AI model.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to a safety monitoring device, a safety monitoring method, and a safety monitoring program. [Background technology]

[0002] Traditionally, safety officers are assigned to factories and construction sites to ensure that workers work safely. These safety officers need to have the necessary experience and skill to determine and address unsafe conditions. Furthermore, the determination and response to unsafe conditions is often based on implicit knowledge or is not clearly documented, making it difficult to make appropriate decisions and pass on the skills. For this reason, Patent Document 1 discloses a technology for determining the safety status of workers from images using a detection model that outputs a determination result for the worker's safety equipment from images. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Application Publication No. 2023-137295 Summary of the Invention [Problem to be solved by the invention]

[0004] However, the above-mentioned Patent Document 1 merely determines whether or not the worker is wearing the safety device correctly, and there was a need for technology that could comprehensively output unsafe conditions that can be read from images of the worker and the environment in natural language.

[0005] The present invention has been made in consideration of the above, and aims to provide a safety monitoring device, a safety monitoring method, and a safety monitoring program that can output unsafe conditions at a work site comprehensively and in natural language. [Means for solving the problem]

[0006] In order to solve the above-mentioned problems and achieve the object, the safety monitoring device of the present invention comprises an acquisition unit that acquires captured image data from a camera, a judgment unit that determines whether the work site in the imaged state is safe from the captured image data input from the acquisition unit, and makes the judgment using an AI model that outputs information related to unsafe factors and information related to accident prevention measures if the work site is not safe, and an output control unit that, if the judgment unit determines that the work site is unsafe, outputs to the outside a fact that the work site is in an unsafe state and the information related to the unsafe factors and the accident prevention measures output by the AI ​​model, wherein the AI ​​model is obtained by machine learning using training data including information related to a plurality of images that represent each of a plurality of unsafe states, information related to the unsafe factors represented by each of the plurality of images, and information related to explanations of the accident prevention measures corresponding to each of the plurality of images.

[0007] In addition, the safety monitoring method of the present invention includes an acquisition step of acquiring captured image data from a camera; a determination step of using an AI model to determine whether the work site in the imaged state is safe from the captured image data input from the acquisition step, and to output information related to unsafe factors and information related to accident prevention measures if the work site is not safe; and an output control step of, if the work site is determined to be unsafe by the determination step, outputting to the outside a notice that the work site is in an unsafe state and the information related to the unsafe factors and the accident prevention measures output by the AI ​​model, wherein the AI ​​model is obtained by machine learning using training data including information related to a plurality of images representing each of a plurality of unsafe states, information related to the unsafe factors represented by each of the plurality of images, and information related to explanations of the accident prevention measures corresponding to each of the plurality of images.

[0008] In addition, the safety monitoring program and safety monitoring device of the present invention execute the following steps: an acquisition step of acquiring captured image data from a camera; a determination step of using an AI model to determine whether the work site in the imaged state is safe from the captured image data input from the acquisition step, and to output information related to unsafe factors and information related to accident prevention measures if the work site is not safe; and an output control step of, if the work site is determined to be unsafe by the determination step, outputting to the outside a notice that the work site is in an unsafe state and the information related to the unsafe factors and the accident prevention measures output by the AI ​​model, wherein the AI ​​model is obtained by machine learning using training data including information related to a plurality of images representing each of a plurality of unsafe states, information related to the unsafe factors represented by each of the plurality of images, and information related to explanations of the accident prevention measures corresponding to each of the plurality of images. [Effects of the Invention]

[0009] According to the present invention, it is possible to comprehensively output unsafe conditions at a work site in natural language. [Brief explanation of the drawings]

[0010] [Figure 1] FIG. 1 is a block diagram showing the functional configuration of a safety monitoring device according to a first embodiment of the present invention. [Figure 2] FIG. 2 is a flowchart showing an outline of the processing executed by the safety monitoring device 1 according to the first embodiment of the present invention. [Figure 3] FIG. 3 is a schematic diagram illustrating an outline of the processing executed by the safety monitoring device 1 according to the first embodiment of the present invention. [Figure 4] FIG. 4 is a schematic diagram illustrating an outline of the processing executed by the safety monitoring device 1 according to the first embodiment of the present invention. [Figure 5] FIG. 5 is a block diagram showing a functional configuration of a safety monitoring device according to the second embodiment of the present invention. [Figure 6]FIG. 6 is a flowchart showing an outline of the processing executed by the safety monitoring device 1A according to the second embodiment of the present invention. [Figure 7] FIG. 7 is a diagram illustrating an outline of the processing executed by the safety monitoring device 1A according to the second embodiment of the present invention. [Figure 8] FIG. 8 is a block diagram showing a functional configuration of a safety monitoring device according to the third embodiment of the present invention. [Figure 9] FIG. 9 is a flowchart showing an outline of the processing executed by the safety monitoring device 1B according to the third embodiment of the present invention. [Figure 10] FIG. 10 is a diagram illustrating an outline of the processing executed by a safety monitoring device 1B according to the third embodiment of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0011] The safety monitoring device according to the present disclosure will be described in detail below with reference to the drawings. Note that the present disclosure is not limited to the following embodiments. Furthermore, the drawings referred to in the following description merely show the shape, size, and positional relationship in a schematic manner to enable understanding of the contents of the present disclosure. In other words, the present disclosure is not limited to only the shape, size, and positional relationship exemplified in each drawing.

[0012] (Embodiment 1) [Functional configuration of safety monitoring device] Fig. 1 is a block diagram showing the functional configuration of a safety monitoring device according to embodiment 1. The safety monitoring device 1 shown in Fig. 1 is configured using, for example, a desktop personal computer, a laptop personal computer, or a tablet personal computer. The safety monitoring device 1 includes an input unit 11, a communication unit 12, a display unit 13, an output unit 14, a recording unit 15, and a control unit 16.

[0013] The input unit 11 is configured using input interfaces such as a keyboard, a microphone, a mouse, and a touch panel. The input unit 11 accepts input of various operations by the user and outputs signals corresponding to the accepted operations to the control unit 16. The input unit 11 also accepts input of voice uttered by the user, converts the accepted voice into voice data, and outputs the voice data to the control unit 18. Note that the input unit 11 may convert the voice into voice data and accept input as text data by converting the voice data of the voice uttered by the user into text, instead of converting the voice into voice data.

[0014] Under the control of the control unit 16, the communication unit 12 communicates with external devices, such as an external server, cameras installed at the work site, environmental sensors installed at the work site, and mobile phones and vital sensors owned by the representative or individuals at the work site, via a network configured, for example, by the Internet or a mobile phone network, and outputs various information received from these external devices to the control unit 16. Here, the work site refers to a factory, a construction site, or the like. Multiple cameras are installed at the work site, capture images of a predetermined field of view, and generate captured image data. Here, the captured image data includes data in still image formats such as JPEG (Joint Photographic Experts Group), PNG (Portable Network Graphics), and TIFF (Tag Image File Format), as well as data in video formats such as MPEG (Moving Picture Experts Group), AVC (Advanced Video Coding: H.264), and HEVC (High Efficiency Video Coding: H.265). Multiple environmental sensors are installed at the work site and detect the temperature (°C), humidity (%), carbon dioxide concentration (CO2 concentration), air pressure (hPa), noise (dB), and other factors at the work site. The mobile phones are smartphones owned by representatives of each sector at the work site or by individuals. If the external device is a camera, the communication unit 12 receives image data of the work site captured by the camera. This image data includes location information indicating the camera's installation location and the shooting location, time, and identification information for identifying the camera. If the external device is an environmental sensor, the communication unit 12 receives detection data including the temperature (°C), humidity (%), carbon dioxide concentration (CO2 concentration), air pressure (hPa), and noise (dB) detected by the environmental sensor. The detection data includes the installation location of the environmental sensor and identification information for identifying the environmental sensor. Under the control of the control unit 16, the communication unit 12 transmits various information input from the control unit 16 to the external device. The communication unit 12 is configured using, for example, a communication module capable of Wi-Fi (registered trademark) or Bluetooth (registered trademark).

[0015] The display unit 13 is configured using a display such as a liquid crystal display or an organic electroluminescent display (EL display), etc. Under the control of the control unit 16, the display unit 13 displays various information.

[0016] The output unit 14 is configured using, for example, a speaker, etc. Under the control of the control unit 16, the output unit 14 outputs various types of information.

[0017] The recording unit 15 is configured using a hard disk drive (HDD), a solid state drive (SSD), a flash memory, a volatile memory, a non-volatile memory, etc. The recording unit 15 has a program recording unit 151, a generated AI recording unit 152, and a notifier information recording unit 153.

[0018] The program recording unit 151 records various programs executed by the safety monitoring device 1 and various data that the safety monitoring device 1 is executing.

[0019] The generative AI recording unit 152 records a generative artificial intelligence (AI) model. The generative AI model is a trained model that is obtained by performing machine learning in advance using training data that includes at least one or more images that are said to represent an unsafe state, unsafe factors corresponding to the unsafe state, and explanations of accident prevention measures for the unsafe state, and that detects and outputs unsafe states from captured image data that is input to the model, and also outputs explanations of the unsafe factors and accident prevention measures.

[0020] Here, "at least one image representing an unsafe condition" refers to an image containing an incident that immediately led to a serious disaster or accident, or an image containing an incident in which the causes of an accident or mistake identified during an internal accident investigation and the underlying causes were analyzed and countermeasures were implemented. Specifically, the image is a still image corresponding to captured image data of a worker working without a helmet. Furthermore, the unsafe factor corresponding to the unsafe condition is text, graphics, etc., included in the image that describes the unsafe condition. Specifically, if the text is "Working without a helmet in a work area within a factory," the text may be "Please wear a helmet correctly." Furthermore, the explanation of the accident prevention measures is text, images, etc., that explain the accident prevention measures for the unsafe condition. Specifically, if the text is "Please wear a helmet correctly," the text may be "Please wear a helmet correctly." If the image is an image, the image may be an image of the worker wearing a helmet correctly. Furthermore, machine learning using training data can use well-known techniques, including neural networks. This generation AI may be configured to work in conjunction with an external generation AI such as ChatGPT (registered trademark) via the communication unit 12, detect and output unsafe conditions, as well as output an explanation of the unsafe factors and accident prevention measures.

[0021] In addition, instead of a generation AI model, the generation AI recording unit 152 may record an inference model (trained model) learned by machine learning such as deep learning, which infers and outputs unsafe conditions from captured image data, as well as infers and outputs explanations of unsafe factors and accident prevention measures.

[0022] The notifier information recording unit 153 records, for each worker, notifier information that corresponds to the facial features, name, age, affiliation, position (supervisor, manager or leader), industry of the worker at the work site, and contact information (email address or telephone number) of an external device such as a mobile phone that is linked to the contact information of the notifier, who is the worker who notifies of unsafe conditions.

[0023] The control unit 16 is realized using a processor having hardware such as an FPGA (Field-Programmable Gate Array), a GPU (Graphics Processing Unit), or a CPU (Central Processing Unit), and a memory that is a temporary storage area used by the processor. The control unit 16 controls each unit that constitutes the safety monitoring device 1. The control unit 16 has an acquisition unit 161, a determination unit 162, and an output control unit 163.

[0024] The acquisition unit 161 acquires, via the communication unit 12, image data captured by a camera installed at the work site.

[0025] Next, the judgment unit 162 uses the generation AI model recorded by the generation AI recording unit 152 to determine whether the work site in the captured state is safe or not from the still image corresponding to the image data acquired by the acquisition unit 161, and outputs information related to unsafe factors and information related to accident prevention measures if the work site is not in a safe state.

[0026] The output control unit 163 externally outputs the determination result of the determination unit 162. Specifically, when the determination unit 162 determines that the work site is unsafe, the output control unit 163 displays on the display unit 13 the information output by the determination unit 162 that the work site is in an unsafe state, the information related to the unsafe factors output by the generation AI model, the information related to accident prevention measures, the text of these, and a still image showing that the work site has been determined to be unsafe.

[0027] [Safety monitoring device processing] Next, there will be explained the processing executed by the safety monitoring device 1. Fig. 2 is a flowchart showing an outline of the processing executed by the safety monitoring device 1. Fig. 3 is a schematic diagram showing an outline of the processing executed by the safety monitoring device 1.

[0028] As shown in FIG. 2, first, the acquisition unit 161 acquires image data captured by a camera installed at the work site via the communication unit 12 (step S101). Specifically, as shown in FIG. 3, the acquisition unit 161 acquires a still image P1 (hereinafter simply referred to as "still image P1") corresponding to the image data captured by the camera installed at the work site via the communication unit 12. Note that the acquisition unit 161 may acquire multiple pieces of image data or video data continuously captured by the camera at predetermined intervals, for example, one second intervals, via the communication unit 12. Of course, the interval at which the acquisition unit 161 acquires image data can be set appropriately, and may be, for example, 10 ms intervals or 10 seconds intervals.

[0029] Next, the determination unit 162 uses the generative AI model recorded by the generative AI recording unit 152 to input a still image corresponding to the image data acquired by the acquisition unit 161 into the generative AI model to determine whether the work site is unsafe or not (step S102). Specifically, as shown in Fig. 3, the determination unit 162 inputs the still image P1 acquired by the acquisition unit 161 into the generative AI model recorded by the generative AI recording unit 152 to determine whether the work site is unsafe or not.

[0030] Thereafter, the determination unit 162 determines whether the work site is unsafe using the generative AI model (step S103). Specifically, the determination unit 162 determines whether the work site is unsafe based on text or a flag indicating that the work site is unsafe, which is included in the output result output from the generative AI model. In this case, the determination unit 162 determines that the work site is unsafe when the output result output from the generative AI model includes a flag indicating that the work site is unsafe, whereas the determination unit 162 determines that the work site is safe when the output result output from the generative AI does not include a flag indicating that the work site is unsafe. If the determination unit 162 determines that the work site is unsafe (step S103: Yes), the safety monitoring device 1 proceeds to step S104. On the other hand, if the determination unit 162 determines that the work site is not unsafe based on the output result output from the generative AI (step S103: Yes), the safety monitoring device 1 terminates this process.

[0031] In step S104, the output control unit 163 outputs the determination result of the determination unit 162 to the outside. Specifically, as shown in Fig. 3, when the determination unit 162 determines that the work site is unsafe, the output control unit 163 displays text T1 and a still image P1 on the display unit 13 to the effect that the work site is in an unsafe state, including information related to the unsafe factors and information related to accident prevention measures output by the determination unit 162, thereby presenting this to a user U1 or the like who is responsible for the work site. This allows the user U1 to understand that the work site is in an unsafe state and also to intuitively recognize the unsafe state of the work site based on the still image P1. Of course, the output control unit 163 may refer to the notifier information recording unit 153 and output, via the communication unit 12, the fact that the work site is in an unsafe state, the explanation of the unsafe factors, and the accident prevention measures output by the generation AI model to a smartphone, PC, etc. of a manager or supervisor associated with the work site recorded by the notifier information recording unit 153, without displaying the fact that the work site is in an unsafe state and the explanation of the unsafe factors and the accident prevention measures output by the determination unit 162 on the display unit 13. This allows the user U1 to grasp the unsafe state of the work site from a remote location even if he or she is not near the safety monitoring device 1. After step S104, the safety monitoring device 1 terminates this process.

[0032] According to the first embodiment described above, when the determination unit 162 determines that the work site is unsafe, the output control unit 163 presents the fact that the work site is in an unsafe state and the information relating to the unsafe factors and information relating to accident prevention measures output by the determination unit 162 to the user U1 or the like who is in charge of the work site by displaying text T1 and still image P1 on the display unit 13. This allows the user U1 to understand that the work site is in an unsafe state and also to intuitively recognize the unsafe state of the work site based on the still image P1.

[0033] In the first embodiment, the determination unit 162 inputs the still image P1 acquired by the acquisition unit 161 into the generative AI model and outputs information indicating that the work site is in an unsafe state, information related to unsafe factors, and information related to accident prevention measures. However, for example, after a predetermined time has elapsed, the user U1 may take accident prevention measures based on the accident prevention measures output from the safety monitoring device 1, and then input a still image into the generative AI model again to determine whether the work site is safe. Specifically, as shown in FIG. 4 , the user U1 inputs a still image P2 corresponding to the progress image data after the accident prevention measures into the safety monitoring device 1 via the input unit 11 after a predetermined time has elapsed. In this case, the determination unit 162 inputs the still image P2 corresponding to the image data after the accident prevention measures into the generative AI model, and determines whether the work site is unsafe based on the output result output by the generative AI model. The determination unit 162 determines that the work site is not unsafe, i.e., that the work site is safe. At this time, when the determination unit 162 determines that the work site is safe, the output control unit 163 presents a text T2 indicating that the unsafe state of the work site has been resolved to the user U1 or the like who is responsible for the work site by displaying the text T2 on the display unit 13. This allows the user U1 to understand that the work site has gone from an unsafe state to a safe state.

[0034] (Embodiment 2) Next, a second embodiment will be described. In the first embodiment, the safety monitoring device 1 acquires a still image P1 corresponding to image data of the work site, but in the second embodiment, in addition to the still image P1, work environment data of the work site and worker vital data of the workers are acquired to determine whether the work site is unsafe. Below, the functional configuration of the safety monitoring device according to the second embodiment will be described, and then the processing executed by the safety monitoring device according to the second embodiment will be described. Note that the same components as those in the safety monitoring device 1 according to the first embodiment described above will be assigned the same reference numerals, and detailed description thereof will be omitted.

[0035] [Functional configuration of safety monitoring device] Fig. 5 is a block diagram showing the functional configuration of a safety monitoring device according to embodiment 2. The safety monitoring device 1A shown in Fig. 5 includes a recording unit 15A instead of the recording unit 15 of the safety monitoring device 1 according to embodiment 1 described above. The recording unit 15A includes a generated AI recording unit 152A instead of the generated AI recording unit 152 according to embodiment 1 described above.

[0036] The generation AI recording unit 152A records a generation AI model. The generation AI model is a trained model that is generated by machine learning using training data that combines at least one image that is said to represent an unsafe state, unsafe factors corresponding to the unsafe state, explanations of accident prevention measures for the unsafe state, work environment data for each of the multiple unsafe states, vital sign data of workers for each of the multiple unsafe states, and explanations of accident prevention measures for each of the multiple unsafe states. The generation AI model is a trained model that detects and outputs unsafe states from the captured image data input to the model, and also outputs explanations of the unsafe factors and accident prevention measures. Machine learning using training data can be performed using known techniques, including neural networks.

[0037] In addition, instead of a generation AI model, the generation AI recording unit 152A may record an inference model (trained model) learned by machine learning such as deep learning, which infers and outputs unsafe conditions from captured image data, work environment data, and worker vital data, as well as infers and outputs explanations of unsafe factors and accident prevention measures.

[0038] [Safety monitoring device processing] Next, the processing executed by the safety monitoring device 1A will be described. Fig. 6 is a flowchart showing an outline of the processing executed by the safety monitoring device 1A. Fig. 7 is a diagram explaining an outline of the processing executed by the safety monitoring device 1A. In Fig. 6, the safety monitoring device 1A executes step S101A instead of step S101 in Fig. 2 described above, and the processing is otherwise the same, so detailed description will be omitted. For this reason, step S101A will be described below.

[0039] In step S101A, the acquisition unit 161 acquires, via the communication unit 12, image data captured by a camera installed at the work site, work environment data detected by environmental sensors installed at the work site, and worker data detected by vital sensors worn by workers at the work site. Specifically, as shown in FIG. 7, the acquisition unit 161 acquires a still image P1, work environment data, and worker data. Here, the work environment data includes noise, temperature, humidity, and the like at the work site. The worker vital data includes the temperature, heart rate, body temperature, and the like around the worker. When acquiring the work environment data, if multiple environmental sensors are installed at the work site, the acquisition unit 161 may acquire the environmental data from an environmental sensor specified by the user U1, or may acquire the environmental data by selecting the highest detection result from among multiple detection results detected by the multiple environmental sensors, or may acquire the average value of the multiple detection results as the environmental data. Furthermore, when acquiring worker data, if multiple workers are working at a work site, the acquiring unit 161 may acquire worker vital data from a vital sensor worn by a worker designated by the user U1, or may acquire worker vital data of each worker from a vital sensor worn by each worker. After step S101A, the safety monitoring device 1A proceeds to step S102.

[0040] According to the second embodiment described above, the determination unit 162 determines whether the work site is unsafe by inputting the work environment data detected by the environmental sensors installed at the work site and the worker data detected by the vital sensors worn by the workers at the work site, which are acquired by the acquisition unit 161, into the generation AI model recorded by the generation AI recording unit 152A. If the determination unit 162 determines that the work site is unsafe, the output control unit 163 displays on the display unit 13 text T1 and a still image P1 that include information on the unsafe state of the work site determined by the determination unit 162, as well as information on unsafe factors and accident prevention measures, thereby presenting this to a user U1 or the like who is responsible for the work site. This allows the user U1 to understand that the work site is unsafe and to intuitively recognize the unsafe state of the work site based on the still image P1.

[0041] (Embodiment 3) Next, a third embodiment will be described. In the first embodiment, when a work site is determined to be unsafe, the safety monitoring device 1 outputs, via the communication unit 12, to the smartphone or PC of the manager or supervisor associated with the work site recorded by the notifier information recording unit 153, a message that the work site is in an unsafe state, an explanation of the unsafe factors output by the generation AI model, and accident prevention measures. However, in the third embodiment, when a work site is determined to be unsafe, the message that the work site is in an unsafe state, information related to the unsafe factors output by the determination unit, and information related to the accident prevention measures are output to the worker shown in the still image P1. In the following, the functional configuration of the safety monitoring device according to the third embodiment will be described, and then the processing executed by the safety monitoring device according to the third embodiment will be described. Note that the same components as those in the safety monitoring device 1 according to the first embodiment described above will be assigned the same reference numerals, and detailed description thereof will be omitted.

[0042] [Functional configuration of safety monitoring device] Fig. 8 is a block diagram showing the functional configuration of a safety monitoring device according to embodiment 3. The safety monitoring device 1B shown in Fig. 8 includes a recording unit 15B and a control unit 16B instead of the recording unit 15 and the control unit 16 of the safety monitoring device 1 according to embodiment 1 described above.

[0043] The recording unit 15B further includes a facial recognition AI recording unit 154 in addition to the functional configuration of the recording unit 15 of the safety monitoring device 1 according to the first embodiment described above. The facial recognition AI recording unit 154 records a facial recognition AI model. The facial recognition AI model is a trained model that has been trained through machine learning using training data that includes information on a plurality of images in which people appear and information on feature points of the people appearing in each of the plurality of images, and that identifies and outputs the face of a worker (person) appearing in a still image corresponding to input image data.

[0044] The control unit 16B further includes an identification unit 164 in addition to the functional configuration of the control unit 16 of the safety monitoring device 1 according to the above-described embodiment 1. The identification unit 164 identifies the worker appearing in the still image based on the face recognition AI model recorded by the face recognition AI recording unit 154, the still image corresponding to the image data acquired by the acquisition unit 161, and the face (feature amount) of the worker recorded by the notifier information recording unit 153.

[0045] [Safety monitoring device processing] Next, the processing executed by the safety monitoring device 1B will be described. Fig. 9 is a flowchart showing an outline of the processing executed by the safety monitoring device 1B. Fig. 10 is a diagram explaining an outline of the processing executed by the safety monitoring device 1B. In Fig. 9, steps S201 to S203 correspond to steps S101 to S103 in Fig. 2 described above, respectively, but steps S204 and S205 are different. Steps S204 and S205 will be described below.

[0046] In step S204, the identification unit 164 identifies an individual worker who appears in the still image based on the face recognition AI model recorded by the face recognition AI recording unit 154, the still image corresponding to the image data acquired by the acquisition unit 161, and information related to the facial features of the worker recorded by the notifier information recording unit 153. Specifically, as shown in Fig. 10, the identification unit 164 inputs the still image P1 into the face recognition AI model, and identifies the individual worker based on the information related to the facial features of the worker output by the face recognition AI model and the information related to the facial features of each worker recorded by the notifier information recording unit 153.

[0047] Next, the output control unit 163 references the notifier information recording unit 153 and outputs the determination result of the determination unit 162 to the external device of the contact person associated with the worker identified by the identification unit 164 (step S205). Specifically, as shown in FIG. 10, the output control unit 163 outputs text T1 and a still image P1 to the external device of the contact person associated with the face identified by the identification unit 164, the text T1 including information on the unsafe state of the work site and information on the unsafe factors and accident prevention measures output by the determination unit 162. This allows the user U1 working at the work site to understand that the state is unsafe and to intuitively recognize the unsafe state of the work site based on the still image P1. After step S205, this process ends.

[0048] According to the third embodiment described above, the output control unit 163 outputs the judgment result of the judgment unit 162 to the external device of the contact person linked to the face of the worker identified by the identification unit 164, so that the worker at the work site can know that the work site or the worker himself / herself is in an unsafe state.

[0049] (Other embodiments) In the safety monitoring devices according to the above-mentioned embodiments 1 to 3, the generated AI model is recorded (stored) in the recording unit, but this is not limited to this. Image data of the work site captured by a camera can be sent to an external generated AI model via a network, and the output results of detecting unsafe conditions at the work site, an explanation of the unsafe factors, and accident prevention measures can be received and displayed from the external generated AI model, or can be output to an external device carried by the user.

[0050] Various inventions can be formed by appropriately combining multiple components disclosed in the safety monitoring devices according to the above-mentioned embodiments 1 to 3. For example, some components may be omitted from all the components described in the safety monitoring devices according to the above-mentioned embodiments of the present disclosure. Furthermore, the components described in the safety monitoring devices according to the above-mentioned embodiments of the present disclosure may be appropriately combined.

[0051] Furthermore, in the safety monitoring devices according to the first to third embodiments, the "unit" described above can be read as "means" or "circuit," etc. For example, the control unit can be read as control means or control circuit.

[0052] In addition, the programs to be executed by the safety monitoring devices according to embodiments 1 to 3 are provided as file data in an installable or executable format recorded on a computer-readable recording medium such as a CD-ROM, flexible disk (FD), CD-R, DVD (Digital Versatile Disk), USB medium, or flash memory.

[0053] Furthermore, the programs executed by the safety monitoring devices according to the first to third embodiments may be configured to be stored on a computer connected to a network such as the Internet, and provided by being downloaded via the network.

[0054] In the explanation of the flowcharts in this specification, the order of processing between steps is clearly indicated using expressions such as "first," "then," and "continue," but the order of processing required to implement the present invention is not uniquely determined by these expressions. In other words, the order of processing in the flowcharts described in this specification can be changed within a consistent range.

[0055] Although some of the embodiments of the present application have been described in detail above with reference to the drawings, these are merely examples, and the present invention can be implemented in other forms that have undergone various modifications and improvements based on the knowledge of those skilled in the art, including the aspects described in the disclosure of the present invention. [Explanation of symbols]

[0056] 1,1A,1B Safety monitoring device 11 Input section 12 Communications Department 13 Display section 14 Output section 15, 15A, 15B Recording section 16,16B Control section 151 Program Recording Section 152,152A Generation AI Recording Unit 153 Notifier Information Recording Section 154 Facial Recognition AI Recording Department 161 Acquisition Department 162 Judgment section 163 Output control section 164 Specific part

Claims

1. an acquisition unit that acquires captured image data from the camera; a determination unit that determines whether the work site in the imaged state is safe or not based on the captured image data input from the acquisition unit, and that determines using an AI model that outputs information related to unsafe factors and information related to accident prevention measures when the work site is not in a safe state; an output control unit that, when the determination unit determines that the work site is unsafe, outputs to the outside a message that the work site is in an unsafe state, information related to the unsafe factors output by the AI ​​model, and information related to the accident prevention measures; Equipped with The AI ​​model is The information is obtained by performing machine learning using training data including information on a plurality of images each representing a plurality of unsafe situations, information on unsafe factors represented by each of the plurality of images, and information on explanations of accident prevention measures corresponding to each of the plurality of images. Safety monitoring device.

2. 2. The safety monitoring device according to claim 1, The AI ​​model is a generation AI model that determines whether a work site of the captured image data is in a safe state based on the captured image data input from the acquisition unit, and generates and outputs new information related to unsafe factors and new information related to the accident prevention measures when the work site is not in a safe state, The output control unit new information related to unsafe factors generated by the generation AI model when the work site is not in a safe state as information related to the unsafe factors, and information related to the accident prevention measures generated by the generation AI model, and outputting the new information related to the accident prevention measures to the outside; Safety monitoring device.

3. 2. The safety monitoring device according to claim 1, The output control unit When the determination unit determines that the work site is unsafe, the determination unit outputs to a display unit a still image corresponding to the captured image data input to the AI ​​model. Safety monitoring device.

4. 2. The safety monitoring device according to claim 1, The teacher data is Further including work environment information relating to an environment at the work site and worker vital information relating to a worker, The determination unit The image data, the work environment data, and the worker vital data are input to the AI ​​model, and a determination is made as to whether the work site is unsafe. Safety monitoring device.

5. 2. The safety monitoring device according to claim 1, a notifier information recording unit that records notifier information that associates facial features of a worker with an external device to which contact information for the worker is linked; Machine learning is performed using training data including information on a plurality of images in which people appear and information on feature points of the people appearing in each of the plurality of images, and the captured image data is input into a face recognition AI model that identifies people appearing in the images, and an identification unit that identifies the worker appearing in a still image corresponding to the captured image data based on an output result output by the face recognition AI model and the notifier information; Furthermore, The output control unit When the determination unit determines that the work site is unsafe, the determination unit outputs to the external device linked to the contact information of the worker identified by the identification unit a message that the work site is in an unsafe state, information related to the unsafe factors output by the AI ​​model, and information related to the accident prevention measures. Safety monitoring device.

6. 2. The safety monitoring device according to claim 1, The acquisition unit After the output control unit outputs a message indicating that the work site is in an unsafe state, the output control unit acquires progress image data captured by the camera after a predetermined time has elapsed, The determination unit The progress image data is input to the AI ​​model to determine whether the work site is unsafe, and the output control unit When the determination unit determines that the work site is not unsafe, an output is made to indicate that the unsafe state at the work site has been resolved. Safety monitoring device.

7. A safety monitoring method executed by a safety monitoring device, an acquisition step of acquiring captured image data from the camera; a determination step of determining whether the work site in the imaged state is in a safe state from the captured image data input from the acquisition step, and using an AI model to output information on unsafe factors and information on accident prevention measures when the work site is not in a safe state; an output control step of outputting to the outside, when the work site is determined to be unsafe by the determination step, a message that the work site is in an unsafe state, information related to the unsafe factors output by the AI ​​model, and information related to the accident prevention measures; Including, The AI ​​model is The information is obtained by performing machine learning using training data including information on a plurality of images each representing a plurality of unsafe situations, information on unsafe factors represented by each of the plurality of images, and information on explanations of accident prevention measures corresponding to each of the plurality of images. Safety monitoring methods.

8. Safety monitoring devices, an acquisition step of acquiring captured image data from the camera; a determination step of determining whether the work site in the imaged state is in a safe state from the captured image data input from the acquisition step, and using an AI model to output information on unsafe factors and information on accident prevention measures when the work site is not in a safe state; an output control step of outputting to the outside, when the work site is determined to be unsafe by the determination step, a message that the work site is in an unsafe state, information related to the unsafe factors output by the AI ​​model, and information related to the accident prevention measures; Execute The AI ​​model is The information is obtained by performing machine learning using training data including information on a plurality of images each representing a plurality of unsafe situations, information on unsafe factors represented by each of the plurality of images, and information on explanations of accident prevention measures corresponding to each of the plurality of images. Safety monitoring program.

Images