Safety monitoring system and safety monitoring method

Abstract

To provide a safety monitoring system and safety monitoring method that accurately determine whether workers are wearing fall prevention equipment correctly.SOLUTION: A safety monitoring system 100 is composed of a computer having an arithmetic unit that executes a predetermined process and a storage device connected to the arithmetic unit, and includes a detection model that outputs a determination result of safety equipment of a worker from an image, and an application unit 110 that determines the status of the worker using the detection model to make an auxiliary determination for modifying the determination result using the detection model. The application unit 110 determines that the worker's condition is safe if either a master rope passes through a safety hook, the safety hook is away from the body, or the safety hook is on the back side, as auxiliary determination, and that the worker's condition is unsafe if the safety hook is higher than the master rope.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present invention relates to a safety monitoring system. [Background technology]

[0002] A conventional method for ensuring the safety of workers involves recording the work being done with a video camera and then having a person visually check the video footage to determine whether any dangerous work is being performed.

[0003] The following prior art exists as background art in this technical field: Patent Document 1 (JP 2006-285639 A) describes an on-site attendance support system for remotely supporting attendance work, which monitors the work status of workers working at a work site, and describes a dangerous work detection system that includes a network for communicating information necessary for attendance work, an information collection terminal that is installed at the work site and collects image information of work being done at the work site taken by a specified device and environmental information indicating the environmental conditions of the work site measured by the specified device, and transmits the collected image information and environmental information via the network, and a center server that receives the image information and environmental information transmitted from the information collection terminal via the network and monitors at least the safety of the work site and the appropriateness of the work process based on the image information and environmental information.

[0004] Patent document 2 (JP 2021-129953 A) describes a management system that includes an image acquisition means for acquiring on-site images of a work site where workers equipped with safety belts are working, an image processing means for generating, from the on-site images, color images of the colors attached to the hooks of the safety belts and color images of the colors attached to the handrails or life ropes at the work site, and a determination means for determining the usage status of the hooks based on the color images of the hooks and the color images of the handrails or life ropes. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Japanese Patent Application Laid-Open No. 2006-285639 [Patent Document 2] Patent Publication No. 2021-129953 Summary of the Invention [Problem to be solved by the invention]

[0006] At construction sites, workers may become overwhelmed with work or make careless mistakes, resulting in them forgetting to wear protective equipment or ignoring work rules. However, as with the background technology mentioned above, it is difficult for managers to monitor workers constantly, making thorough safety a key issue. Furthermore, using a computer to individually assess workers' equipment and behavior increases processing time, making it difficult to ensure real-time assessments. Furthermore, assessments using AI models may result in incorrect assessments depending on the progress of the AI ​​model's learning. For this reason, there is a demand for assessments that reduce error. [Means for solving the problem]

[0007] A representative example of the invention disclosed in the present application is as follows: That is, a safety monitoring system is configured by a computer having an arithmetic unit that executes predetermined processing and a storage device connected to the arithmetic unit, and is equipped with a detection model that outputs a judgment result of a worker's safety equipment from an image, and an application unit that judges the worker's condition using the detection model and executes auxiliary judgment to correct the judgment result using the detection model, and as the auxiliary judgment, the application unit judges that the worker's condition is safe if any of the following is satisfied: a lifeline is passing through a safety hook, the safety hook is away from the body, or the safety hook is on the back side; In the image of the safety hook If the safety hook is positioned higher than the lifeline, the system determines that the worker is in a dangerous state. [Effects of the Invention]

[0008] According to one aspect of the present invention, it is possible to accurately determine whether a worker is wearing fall protection equipment correctly. Objects, configurations, and effects other than those described above will become apparent from the following description of the embodiments. [Brief explanation of the drawings]

[0009] [Figure 1] 1 is a diagram showing the configuration of a safety monitoring system according to an embodiment of the present invention; [Figure 2] 1 is a block diagram showing the physical configuration of a safety monitoring system according to an embodiment of the present invention. [Figure 3] FIG. 10 is a diagram illustrating an example of the configuration of data stored in a result storage unit according to the present embodiment. [Figure 4] 3 is a flowchart of a process executed by the safety monitoring system of the present embodiment. [Figure 5] 10 is a detailed flowchart of the process of step S504 in this embodiment. [Figure 6] FIG. 10 is a diagram illustrating the processing of step S504 in this embodiment. [Figure 7] 10 is a detailed flowchart of the process of step S505 in this embodiment. [Figure 8] FIG. 10 is a diagram illustrating the processing of step S505 in this embodiment. [Figure 9] 10 is a detailed flowchart of the process of step S506 in this embodiment. [Figure 10] FIG. 10 is a diagram illustrating the processing of step S506 in this embodiment. [Figure 11] 10 is a detailed flowchart of the process of step S507 in this embodiment. [Figure 12] FIG. 10 is a diagram illustrating the processing of step S507 in this embodiment. [Figure 13] FIG. 10 is a diagram showing an example of a worker safety state determination result screen in the present embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0010] FIG. 1 is a diagram showing the configuration of a safety monitoring system 100 according to an embodiment of the present invention.

[0011] The safety monitoring system 100 of this embodiment includes an application unit 110, an AI model 120, setting information 130, a result storage unit 140, and a data analysis unit 150.

[0012] The application unit 110 receives the video data captured by the monitoring camera 200, records the results of determining people appearing in the video data in the result storage unit 140, and outputs the results to the administrator's computer 300.

[0013] The AI ​​model 120 detects people from images, detects attributes of the people in the images, and outputs a confidence level for each attribute. The AI ​​model 120 may include multiple models (e.g., a person detection model, a hook detection model, and a lifeline detection model). The AI ​​model 120 receives an image of a worker and a person region, and learns a person detection model. The AI ​​model 120 also receives an image of a safety hook and a hook region, and learns a hook detection model. The AI ​​model 120 also receives an image of a support and a lifeline, and the coordinates of the support and lifeline, and learns a lifeline detection model. During inference, the AI ​​model 120 detects workers, safety hooks, supports, and lifelines from images of the work site, and outputs a person region where the worker is detected, a hook region where the safety hook is detected, and a lifeline region where the lifeline is detected. The supports are structures (e.g., temporary posts) to which lifelines are stretched, and scaffolding pipes or steel frames to which lifeline clamps are attached. Since the support is thicker than the lifeline and therefore easier to detect, it is advisable to detect the lifeline based on the position of the detected support. The AI ​​model 120 also determines the class of the detected safety hook (S503). The safety hook class is a classification of the safety hook, determined using AI or the like to determine whether the safety hook detected by the AI ​​model 120 is hooked on the lifeline or not.

[0014] The setting information 130 includes the name of the AI ​​model 120 used for processing, a threshold for determining the state from the confidence level output by the AI ​​model 120, and a predetermined number of times for determining the continuous detection state using a counter. The result storage unit 140 records the determination results of the person's attributes. Details of the result storage unit 140 will be described later with reference to FIG. 3. The result storage unit 140 is configured to be able to output the recorded attribute determination results as a log. The data analysis unit 150 analyzes the data recorded in the result storage unit 140 and outputs the analysis results.

[0015] The surveillance camera 200 is a camera installed at a work site and captures images of workers working. It is advisable to install multiple surveillance cameras 200 at the work site so that the entire work site is captured.

[0016] The administrator computer 300 is a computer having an arithmetic unit and a storage device, and receives the judgment results and monitoring images from the safety monitoring system 100. An example of a screen displayed by the administrator computer 300 will be described later with reference to Fig. 7. The administrator computer 300 operates an alarm device (for example, a patrol lamp or an alarm sound generator) according to the judgment results notified from the safety monitoring system 100.

[0017] For example, a surveillance camera 200 is installed at a work site, and it takes pictures of workers at work. If it detects that the worker is not wearing the required safety equipment, it will turn on a patrol lamp and issue an audio alert that the situation is dangerous.

[0018] FIG. 2 is a block diagram showing the physical configuration of the safety monitoring system 100 of this embodiment.

[0019] The safety monitoring system 100 of this embodiment is configured by a computer having a processor (CPU) 1, a memory 2, an auxiliary storage device 3, and a communication interface 4. The safety monitoring system 100 may also have an input interface 5 and an output interface 8.

[0020] The processor 1 is a computing device that executes programs stored in the memory 2. The processor 1 executes various programs to realize the functions of each functional unit (e.g., application unit 110, AI model 120, data analysis unit 150, etc.) of the safety monitoring system 100. Note that some of the processing performed by the processor 1 by executing the programs may be executed by another computing device (e.g., hardware such as ASIC or FPGA).

[0021] The memory 2 includes a ROM, which is a non-volatile storage element, and a RAM, which is a volatile storage element. The ROM stores unchanging programs (e.g., BIOS), etc. The RAM is a high-speed, volatile storage element such as a DRAM (Dynamic Random Access Memory), and temporarily stores programs executed by the processor 1 and data used when the programs are executed.

[0022] The auxiliary storage device 3 is a large-capacity, non-volatile storage device such as a magnetic storage device (HDD) or a flash memory (SSD). The auxiliary storage device 3 also stores data (e.g., setting information 130, result storage unit 140, etc.) used by the processor 1 when executing a program, and the program executed by the processor 1. That is, the program is read from the auxiliary storage device 3, loaded into the memory 2, and executed by the processor 1 to realize each function of the safety monitoring system 100.

[0023] The communication interface 4 is a network interface device that controls communication with other devices (for example, the surveillance camera 200, the administrator's computer 300) in accordance with a predetermined protocol.

[0024] The input interface 5 is an interface to which input devices such as a keyboard 6 and a mouse 7 are connected and which receives input from an operator. The output interface 8 is an interface to which output devices such as a display device 9 and a printer (not shown) are connected and which outputs the results of program execution in a format that can be viewed by the user. Note that a user terminal connected to the safety monitoring system 100 via a network may provide the input and output devices. In this case, the safety monitoring system 100 may have a web server function, and the user terminal may access the safety monitoring system 100 using a predetermined protocol (for example, http).

[0025] The program executed by the processor 1 is provided to the safety monitoring system 100 via removable media (CD-ROM, flash memory, etc.) or a network, and is stored in a non-volatile auxiliary storage device 3, which is a non-transitory storage medium. For this reason, the safety monitoring system 100 preferably has an interface for reading data from removable media.

[0026] The safety monitoring system 100 is a computer system configured on a single physical computer or on multiple logically or physically configured computers, and may operate on a virtual computer constructed on multiple physical computer resources. For example, the application unit 110, the AI ​​model 120, and the data analysis unit 150 may each operate on separate physical or logical computers, or multiple units may be combined to operate on a single physical or logical computer.

[0027] FIG. 3 is a diagram showing an example of the structure of data stored in the result storage unit 140 of this embodiment.

[0028] The result storage unit 140 records a person region where a person is detected from the camera image and a hook region where the safety hook worn by the person is detected. The person region is represented by the coordinates of the four corners of a rectangle in the image where the person is detected, and the hook region is represented by the coordinates of the four corners of a rectangle in the image where the safety hook is detected.

[0029] FIG. 4 is a flowchart of the process executed by the safety monitoring system 100 of this embodiment.

[0030] The application unit 110 acquires video data in a stream format captured by the surveillance camera 200 (S500).

[0031] Next, the AI ​​model 120 detects people from frame images of the video data using a person detection model, and detects the support and lifeline using a lifeline detection model (S501). Then, it detects safety hooks using a hook detection model (S502) and determines the class of the detected safety hook (S503). If it is determined that the safety hook is engaged with the lifeline (Yes in S503), the worker's status is set to safe (S510). Note that, depending on the learning progress of the AI ​​model 120, the AI ​​model 120 may erroneously determine that a safety hook that is not engaged with the lifeline is safe, or that a safety hook that is engaged with the lifeline is dangerous. Therefore, a supplementary determination is made after the AI ​​class determination to reduce errors.

[0032] Therefore, if it is determined that the safety hook is not attached to the main rope (No in S503), the auxiliary determination in steps S504 to S508 detects an erroneous determination by the AI ​​model 120 and classifies the worker's condition as safe or dangerous.

[0033] In the auxiliary determination, first, the application unit 110 determines whether the life rope is passing through the safety hook (S504). If the life rope is passing through the safety hook, the status of the worker is set to safe (S510). Details of the processing in step S504 will be described with reference to FIGS. 5 and 6.

[0034] On the other hand, if the lifeline is not passing through the safety hook, the application unit 110 determines whether the safety hook detection position is higher than the lifeline (S505). If the safety hook detection position is higher than the lifeline, it determines that the safety hook is attached to the harness worn by the worker, and sets the worker's status to danger (S513). Details of the processing in step S505 will be described with reference to Figures 7 and 8.

[0035] On the other hand, if the safety hook detection position is lower than the master rope, the application unit 110 determines whether the safety hook is away from the body (S506). If the safety hook is away from the body, it determines that the safety hook is not attached to the harness worn by the worker, and sets the worker's status to safe (S510). Details of the processing in step S506 will be described with reference to Figures 9 and 10.

[0036] On the other hand, if the safety hook is not separated from the body, the application unit 110 determines whether the safety hook is on the back side (S507). If the safety hook is on the back side, it determines that the worker is walking while dragging the safety hook attached to the life rope, and sets the worker's status to safe (S510). Details of the processing in step S507 will be described with reference to Figures 11 and 12.

[0037] If the auxiliary determination fails to classify the worker's status as safe or dangerous, the application unit 110 determines the status of the previous and next frames (S508). In the previous and next frame status determination process, if the auxiliary determination in steps S504 to S507 fails to classify the worker's status as safe or dangerous, a counter corresponding to the worker is incremented. If the number of consecutive times that a safety hook is not attached is a predetermined number (n times) or less, the status of the worker is set to alert (S511). On the other hand, if the number of consecutive times that a safety hook is not attached is a predetermined number or more, the status of the worker is set to danger (S513). If the worker's status is safe, the counter is initialized and the status of the worker is set to safe.

[0038] If the worker's condition is alert, the application unit 110 notifies the administrator computer 300 (S512). The notification to the administrator computer 300 may be displayed on the worker safety condition determination result screen 700, which will be described later with reference to FIG. 13, or may be displayed as a pop-up outside the worker safety condition determination result screen 700. Furthermore, if the worker's condition is dangerous, the application unit 110 activates an alarm device installed at the work site to notify the worker of the dangerous condition (S513). At this time, the notification may be sent to the administrator computer 300. In this way, when unsafe behavior is detected a predetermined number of times in succession, the condition is set to dangerous and an alert is issued, so that the condition can be accurately determined while preventing erroneous detections that capture momentary changes.

[0039] As described above, in steps S504 to S507, the worker is judged to be safe or unsafe based on four criteria, but the order of judgment based on these criteria is not limited to the order shown in the figure. Also, judgment may be made by selectively using only some of the four criteria. Selection of which criteria to use can be set on the worker safety status judgment result screen 700, which will be described later with reference to FIG. 13. Furthermore, while the worker is judged to be safe or unsafe based on each of the four criteria, the worker may be judged to be safe or unsafe by calculating a judgment value as a weighted sum of the judgment results of each criterion, and then the worker may be judged to be safe or unsafe based on whether the calculated judgment value exceeds a predetermined threshold. In this way, by making it possible to adjust the adoption and weighting of multiple criteria, the accuracy of the auxiliary judgment can be adjusted.

[0040] FIG. 5 is a flowchart showing the details of the process in step S504, and FIG. 6 is a diagram for explaining the process in step S504.

[0041] First, the application unit 110 cuts out image 5046A of the hook area of ​​the safety hook detected by the hook detection model (S5041). Next, it detects the lifeline 5047 from the image in which the safety hook has been detected by the lifeline detection model (S5042). For example, the lifeline detection model uses location information of the installed lifeline, the installation position of the monitoring camera 200, and the shooting direction of the monitoring camera 200 to identify an area in the captured image where the lifeline may be captured, detects the lifeline in that area, and converts the detected lifeline area into a binary image (e.g., white). Next, it compares the coordinates of the hook area with those of the lifeline (S5043). If the coordinates of the hook area and the lifeline overlap, it is determined that the lifeline is passing through the safety hook. On the other hand, if the coordinates of the hook area and the lifeline do not overlap, it is determined that the lifeline is not passing through the safety hook.

[0042] FIG. 7 is a flowchart showing the details of the process in step S505, and FIG. 8 is a diagram for explaining the process in step S505.

[0043] First, the application unit 110 acquires the coordinates of the hook area (S5051). If the image of the hook area has not been cut out, it cuts out the image of the hook area of ​​the detected safety hook. For example, it acquires the coordinates of the four vertices of the hook area of ​​the safety hook detected by the hook detection model, and calculates the coordinate of the center point of the bottom side of the hook area. Then, it detects the life rope within the person area 5057 (S5052). If multiple life rope areas have been cut out, it selects the topmost life rope area (S5053). Then, it compares the coordinates of the hook area with the positional relationship of the life rope (S5054). Then, it determines whether the coordinates of the hook area are higher than the life rope (S5055).

[0044] For example, since the center point of the lower side of hook area 5046B shown in Fig. 8 is above the closest point of main rope 5047 (5056B), it is determined that the safety hook is attached to the harness worn by the worker. On the other hand, since the center point of the lower side of hook area 5046C is below the closest point of main rope 5047 (5056C), it is not determined that the safety hook is attached to the harness worn by the worker.

[0045] FIG. 9 is a flowchart showing the details of the process in step S506, and FIG. 10 is a diagram for explaining the process in step S506.

[0046] First, the application unit 110 acquires the coordinates of the hook area (S5061). If the image of the hook area has not been cut out, the application unit 110 cuts out the image of the hook area of ​​the detected safety hook. For example, the application unit 110 acquires the coordinates of the four vertices of the hook area of ​​the safety hook detected by the hook detection model. Then, the application unit 110 measures the distance between the hook area and the person area associated with the hook area (S5062). Then, the application unit 110 compares the hook area with the person area (S5063) and determines whether the hook area and the person area are separated by a predetermined distance or more (S5064).

[0047] 10 is separated from person area 5057 by a predetermined distance or more (5067), it is determined that the safety hook is away from the body. On the other hand, hook area 5046E is included in person area 5057 and is not separated from person area 5057 by a predetermined distance or more, so it is determined that the safety hook is not away from the body.

[0048] FIG. 11 is a detailed flowchart of the process in step S507, and FIG. 12 is a diagram for explaining the process in step S507.

[0049] First, the application unit 110 acquires the coordinates of the hook area (S5071). If the image of the hook area has not been cut out, it cuts out the image of the hook area of ​​the detected safety hook. For example, the application unit 110 acquires the coordinates of the four vertices of the hook area of ​​the safety hook detected by the hook detection model, and calculates the coordinate of the center point of the hook area. Then, it acquires multiple frames before and after the video captured by the surveillance camera 200, tracks the person, and calculates the person's movement vector (S5072). It then compares the direction of the person's movement vector with the position of the hook area (S5073), and determines whether there is a safety hook in the direction opposite to the direction of the person's movement vector (S5074). The positional relationship between the hook area and the person area can be determined by a vector from the center point of the person area to the center point of the hook area.

[0050] As shown in FIG. 12 , a person movement vector 5076 is calculated from changes in the person area over multiple frames, and hook position vectors 5077 and 5078 are calculated from the center point of the person area 5057 to the center points of hook areas 5046F and 5046G. Then, the angle between the person movement vector 5076 and the hook position vectors 5077 and 5078 is calculated, and it is determined whether the calculated angle is greater than a predetermined threshold. For example, the angle between the person movement vector 5076 and the hook position vector 5077 is 170 degrees, which is greater than a predetermined threshold (e.g., 150 degrees), so it is determined that the worker is walking while dragging the safety hook attached to the lifeline, and that the safety hook is behind the worker. On the other hand, the angle between the person movement vector 5076 and the hook position vector 5078 is 40 degrees, which is smaller than the predetermined threshold, so it is determined that the safety hook is not behind the worker.

[0051] FIG. 13 is a diagram showing an example of a worker safety state determination result screen 700 output by the safety monitoring system 100 of this embodiment.

[0052] The manager of the work site operates the manager computer 300 to select a location to be monitored (i.e., the monitoring camera 200), and displays the image of the location and the monitoring results on the screen. For example, when an alert is issued from the manager computer 300, the manager displays the image of the location and the monitoring results on the screen of the manager computer 300. An example of the worker safety state judgment result screen 700 includes an image display area 710, operation buttons 751 to 753, and a judgment condition setting area 760.

[0053] The image display area 710 displays a frame image captured by the surveillance camera 200. Detected person areas 721 to 723 are displayed in the frame image, and the worker judgment result (safe = SAFE, caution = WRNING, danger = DANGER!!) of each person area is displayed.

[0054] The worker safety state determination result screen 700 includes operation buttons 751 to 753. The on-site notification button 751 is operated when activating an alarm device installed at the site. The notification stop button 752 is operated when halting the operation of an alarm device installed at the site. The camera switching button 753 is operated when viewing an image captured by another monitoring camera 200 installed at the site.

[0055] In the judgment condition setting area 760, it is set for each item whether or not to use multiple judgment conditions for auxiliary judgment.

[0056] As described above, it is possible to accurately determine in real time whether a worker is wearing fall protection equipment correctly.

[0057] The present invention is not limited to the above-described embodiments, but includes various modifications and equivalent configurations within the spirit and scope of the appended claims. For example, the above-described embodiments have been described in detail to clearly explain the present invention, and the present invention is not necessarily limited to configurations including all of the described configurations. Furthermore, part of the configuration of one embodiment may be replaced with the configuration of another embodiment. Furthermore, the configuration of another embodiment may be added to the configuration of one embodiment. Furthermore, part of the configuration of each embodiment may be added, deleted, or replaced with other configurations.

[0058] Furthermore, the aforementioned configurations, functions, processing units, processing means, etc. may be realized in part or in whole in hardware, for example by designing them as integrated circuits, or may be realized in software by having a processor interpret and execute a program that realizes each function.

[0059] Information such as programs, tables, and files that realize each function can be stored in a storage device such as a memory, a hard disk, or an SSD (Solid State Drive), or in a recording medium such as an IC card, an SD card, or a DVD.

[0060] In addition, the control lines and information lines shown are those that are considered necessary for explanation, and do not necessarily represent all the control lines and information lines that are necessary for implementation. In reality, it can be assumed that almost all components are interconnected. [Explanation of symbols]

[0061] 1 processor 2. Memory 3 Auxiliary storage 4. Communication Interface 5 Input Interface 6 Keyboard 7. Mouse 8 Output Interfaces 9 Display Devices 100 Safety Monitoring System 110 Application Section 120 AI models 130 Setting Information 140 Result storage section 150 Data Analysis Department 200 surveillance cameras 300 Administrator calculator

Claims

1. A safety monitoring system, comprising: The computer is configured by an arithmetic unit that executes predetermined processing and a storage device connected to the arithmetic unit, A detection model that outputs a judgment result of the safety equipment of the worker from the image; an application unit that determines the state of the worker using the detection model and executes auxiliary determination to correct a determination result using the detection model; The application unit performs the auxiliary determination by: If any of the following conditions are met, the worker's condition is determined to be safe: the lifeline is passing through the safety hook, the safety hook is away from the body, or the safety hook is on the back side. A safety monitoring system characterized in that if the safety hook is positioned higher than the lifeline in the image of the safety hook, it is determined that the worker's condition is dangerous.

2. 2. The safety monitoring system according to claim 1, It is connected to the manager's computer used by the manager and the alarm device installed at the work site, The application unit If the determination using the detection model does not determine that the worker is safe, and the auxiliary determination does not determine whether the worker is safe or dangerous, determine that the worker is in an alert state; If the worker's state is determined to be alert a predetermined number of times or more in succession, the worker's state is determined to be dangerous; If the worker's status is alert, sending a notification to the manager's computer; A safety monitoring system characterized in that the alarm device is activated when the condition of the worker is dangerous.

3. 3. The safety monitoring system according to claim 1 or 2, The application unit Detecting a lifeline from an image of a hook region of the safety hook detected by the detection model; comparing the hook area with the coordinates of the detected lifeline; A safety monitoring system characterized in that if the coordinates of the hook area and the lifeline overlap, it is determined that the lifeline is passing through the safety hook.

4. 3. The safety monitoring system according to claim 1 or 2, The application unit Obtaining the coordinates of the hook region of the safety hook detected by the detection model; Detecting a lifeline within a person area of ​​the person detected by the detection model; By comparing the coordinates of the hook area with the positional relationship of the detected lifeline, A safety monitoring system characterized by determining whether the safety hook is at a higher position than the lifeline.

5. 3. The safety monitoring system according to claim 1 or 2, The application unit Obtaining the coordinates of the hook region of the safety hook detected by the detection model; measuring a distance between the hook region and a person region associated with the hook region; A safety monitoring system characterized in that it is determined whether the safety hook is away from the worker's body depending on whether the hook area and the person area are separated by a predetermined distance or more.

6. 3. The safety monitoring system according to claim 1 or 2, The application unit Obtaining the coordinates of the hook region of the safety hook detected by the detection model; Tracking a person from previous and next frames of the video to calculate the person's movement vector, A safety monitoring system characterized by comparing the calculated movement vector direction with the position of the hook area to determine whether the safety hook is on the back side depending on whether the safety hook is in the opposite direction to the person's movement vector direction.

7. A safety monitoring method executed by a safety monitoring system, comprising: the safety monitoring system is configured by a computer having an arithmetic unit that executes predetermined processing and a storage device connected to the arithmetic unit, a detection model that outputs a judgment result of the safety equipment of a worker from an image; and an application unit that judges the state of the worker using the detection model and executes auxiliary judgment to correct the judgment result using the detection model, The safety monitoring method includes: The computing device determines the state of the worker using the detection model; the computing device performs auxiliary determination to correct a determination result using the detection model; A safety monitoring method characterized in that the calculation device determines that the worker's condition is safe if the auxiliary judgment satisfies any of the following conditions: the main rope is passing through the safety hook, the safety hook is away from the body, or the safety hook is on the back side, and determines that the worker's condition is dangerous if the safety hook is positioned higher than the main rope in the image of the safety hook.

Images