KR20230121229A - Occupational safety and health education system through artificial intelligence video control and method thereof - Google Patents

Abstract

The present invention relates to a system and method for occupational safety and health education, and more particularly, images obtained from an image acquisition device such as a CCTV installed at an arbitrary location in an industrial site and photographing a preset surveillance area in a dangerous situation for each occupational safety factor. It is applied to the dangerous situation detection artificial intelligence model that detects dangerous situations corresponding to industrial safety elements in the monitoring area, and selects workers corresponding to the monitoring areas in which dangerous situations are detected as training subjects, It relates to an occupational safety and health education system and method through artificial intelligence video control that provides and educates occupational safety and health education contents corresponding to the detected dangerous situation through a terminal.

Description

Occupational safety and health education system through artificial intelligence video control and method thereof}

The present invention relates to an occupational safety and health education system and method, and more particularly, to a closed circuit television (CCTV) that is installed at an arbitrary location in an industrial field and photographs a preset area (hereinafter referred to as "surveillance area"). ), etc., are applied to a dangerous situation detection artificial intelligence model that detects dangerous situations for each industrial safety factor to detect dangerous situations corresponding to industrial safety factors in the surveillance area, and detect dangerous situations Occupational safety and health education system and method through artificial intelligence video control that selects workers corresponding to the surveillance area as training subjects and provides occupational safety and health training contents corresponding to the detected dangerous situation through the mobile terminal of the training subjects and educates them. It is about.

With the development of industrial technology, industrial sites are becoming larger and more complicated due to numerous equipment, workers, and materials.

As industrial sites become larger and more complex, fire, building collapse, construction equipment (crane, excavator, etc.) fall, leakage (leakage of gas, water, etc.), worker fall, fall, material fall, collision, confined space suffocation, etc. Various dangerous situations and accidents occur frequently.

In order to keep industrial sites safe by preventing dangerous situations and accidents in these industrial sites, the installation of various safety management equipment is recommended and strengthened so that each site can identify and cope with the situation. The safety management equipment may be a gas detector, a fire alarm, an emergency treatment facility, and the like.

However, conventional safety management equipment uses sensors to monitor the occurrence of dangerous situations, and it is difficult to control safety conditions for various dangerous situations throughout the industrial site because it detects and notifies only the status of specific dangerous situations in a narrow area.

In addition, in order to solve these problems, Republic of Korea Patent Publication No. 10-2021-0050711, "Video-based industrial site safety condition prediction device and method" (2021.05.10. Publication: hereinafter referred to as "prior art") is Disclosed is a technique for monitoring whether a safe state is maintained in the image by detecting it with an image acquisition device such as CCTV and applying the captured image to an artificial intelligence model learned for a specific safe state.

However, the prior art only controls whether a specific safety state is maintained, and it is difficult to monitor various dangerous situations that may occur at industrial sites, and it is difficult to monitor a place where a fire may occur and media such as gas, water, and oil leak. It is possible to monitor various factors that require occupational safety (hereinafter referred to as "industrial safety factors") such as possible location, equipment in operation, materials in transportation and movement, worker's work location, worker's safety equipment, etc. There are no problems.

In addition, most dangerous situations and accidents in industrial sites are caused by negligence in industrial safety management by workers and managers who handle equipment of industrial safety elements and perform work in industrial safety elements.

Therefore, occupational safety and health education is compulsory for these workers and managers, and workers and managers who are subject to occupational safety and health education either directly attend the relevant occupational safety education content at the training center or online. should be taken.

Workers and managers do not use it well because it is cumbersome to go directly to the training center, and mainly take the corresponding training content online.

However, online education has problems in that it is difficult to check whether workers and managers are taking the training properly, and the efficiency of education is reduced because it does not provide educational contents closely related to industrial safety elements that workers and managers actually work and manage.

In addition, even in the case of visiting a training center and receiving training, there is a problem in that the efficiency of training is reduced because there are many contents that have nothing to do with the work directly processed by workers and managers.

Therefore, by detecting dangerous situations corresponding to each of the various industrial safety factors through video, and responding to the dangerous situations to workers who are or may be in dangerous situations and the management of industrial safety elements in which dangerous situations are about to occur. The development of an occupational safety and health education system capable of providing safety education contents is required.

Republic of Korea Patent Publication No. 10-2021-0050711 (published on May 10, 2021)

Therefore, an object of the present invention is to apply an image obtained from an image acquisition device such as a CCTV installed at an arbitrary location in an industrial site and photographing a preset surveillance area to a dangerous situation detection artificial intelligence model that detects a dangerous situation by industrial safety factor Industrial safety that detects dangerous situations corresponding to industrial safety elements in the monitoring area, selects workers corresponding to the monitoring area where dangerous situations are detected as training subjects, and responds to the detected dangerous situations with the training subject's mobile terminal It is to provide an occupational safety and health education system and method through artificial intelligence video control that provides and educates health education contents.

Industrial safety and health education system through artificial intelligence video control according to the present invention for achieving the above object is: video recording unit; a worker location detection unit installed at least one for each surveillance area and transmitting worker location information including worker identification information about a worker located in the surveillance region and location information about the location of the worker; and surveillance area location information about the location of the surveillance area in which the video capture unit is installed, tracks the location of the worker through the operator location detection unit, receives images from the plurality of image capture units, and receives each image. At least one dangerous situation detection artificial intelligence model corresponding to the industrial safety elements existing in the surveillance area is applied to monitor whether a dangerous situation is detected for the corresponding industrial safety element, and when a dangerous situation is detected, the dangerous situation is displayed on the detected image. After selecting a worker in the position of the corresponding monitoring area as a target for occupational safety and health education, industrial safety responding to industrial safety elements and detected risk situations in the monitoring area with a mobile terminal of the worker selected as a subject for occupational safety and health education It is characterized in that it includes an occupational safety and health education providing unit that provides health education content.

The occupational safety and health education providing unit provides learning data for each dangerous situation, including a plurality of images including the corresponding dangerous situation for each dangerous situation and dangerous situation information, which is labeling information included for each image, in response to the dangerous situation. A dangerous situation detection artificial intelligence model learning unit that applies to the detection artificial intelligence model and learns an image for each dangerous situation; After selecting and setting the dangerous situation detection artificial intelligence model to be applied to each surveillance area, including the learned dangerous situation detection artificial intelligence model for each danger situation, the dangerous situation detection artificial intelligence model applied for each surveillance area is applied to the corresponding surveillance area. a danger detection unit that monitors whether a dangerous situation occurs by applying a corresponding image, and outputs monitoring area location information and dangerous situation information about the location of the surveillance area when a dangerous situation is detected; The position of the worker is tracked through the worker position detection unit, and the position information of the surveillance region is input according to the detection of the dangerous situation, and the worker in the position of the surveillance region corresponding to the image in which the dangerous situation is detected is made a subject of occupational safety and health education. After selecting the education target recognition unit for outputting the worker identification information of the worker selected as the training target; And when the worker identification information of the worker selected as the training target is input from the training target recognition unit, the worker's mobile terminal corresponding to the worker identification information corresponds to the occupational safety element of the surveillance area and the dangerous situation of the dangerous situation information Characterized in that it includes an education providing unit that provides occupational safety and health education contents.

The danger detection unit may include: an image acquiring unit that receives images and capturing unit identification information from image capturing units of each monitoring area and outputs corresponding images and capturing unit identification information of the image capturing unit that has transmitted the images for each monitoring area; A monitoring region location table having mapping unit identification information for each of the image capturing units and surveillance region location information of a surveillance region in which the image capturing unit is installed is provided, and when the capturing unit identification information is input from the image acquisition unit, an image acquisition location identification unit for identifying the position of the surveillance region by looking for surveillance region location information corresponding to the photographing unit identification information with reference to the surveillance region location table, and then outputting the surveillance region location information to the training target recognition unit; an artificial intelligence model selection unit that selects and sets an artificial intelligence model for detecting a dangerous situation to be applied to each surveillance area; and the learned dangerous situation detection artificial intelligence model for each dangerous situation, wherein an image corresponding to the surveillance area is applied to the dangerous situation detection artificial intelligence model selected by the artificial intelligence model selection unit and applied to each monitoring area to detect a dangerous situation. It is characterized in that it includes an educational target risk detection unit that monitors whether a dangerous situation occurs and outputs dangerous situation information on the detected dangerous situation to the education providing unit when a dangerous situation is detected.

The risk detection unit for training is applied with an artificial intelligence model for detecting a risk of collision, and monitors whether a collision with an industrial safety element occurs in the surveillance area by applying the image to the artificial intelligence model for detecting a risk of collision. a collision detection unit that outputs collision detection information that is dangerous situation information when a collision is detected; An artificial intelligence model for detecting a risky leakage situation is applied, and the image is applied to the artificial intelligence model for detecting a risky leakage situation to monitor whether leakage of media for industrial safety elements occurs in the surveillance area, and to detect media leakage for industrial safety elements. a leakage detection unit outputting leakage detection information, which is information about a dangerous situation; A fire risk situation detection artificial intelligence model is applied, and the image is applied to the fire risk situation detection artificial intelligence model to monitor whether a fire occurs for industrial safety elements in the surveillance area, and a dangerous situation when a fire is detected for industrial safety elements. a fire detection unit outputting fire detection information, which is information; and a dangerous situation detection artificial intelligence model for unwearing safety equipment is applied, and the image is applied to the artificial intelligence model for detecting a dangerous situation without safety equipment to monitor whether non-wearing of safety equipment for industrial safety elements is detected in the surveillance area, and safety equipment It is characterized in that it includes a safety equipment wearing detection unit that outputs safety equipment non-wearing detection information, which is dangerous situation information on industrial safety elements when non-wearing is detected.

The risk detection unit for education is applied with a dangerous object detection artificial intelligence model, applies the image to the dangerous object detection artificial intelligence model, detects a dangerous object that may cause a dangerous situation in the image, and detects a dangerous object Further comprising a dangerous object detecting unit for providing information on a type and location to the collision detecting unit, wherein the educational target danger detecting unit monitors whether a dangerous situation occurs in the object detected by the dangerous object detecting unit. .

The artificial intelligence model selection unit includes a risk situation table for each industrial safety factor defining one or more dangerous situations that may occur for each industrial safety factor, and has an artificial intelligence model selection artificial intelligence model, and input from the image acquisition unit The image is applied to the artificial intelligence model selection artificial intelligence model to classify industrial safety factors existing in the monitoring area of the image, and one or more risk situations corresponding to the classified industrial safety factors are classified from the risk situation table for each industrial safety factor. It is characterized by recognizing, selecting a dangerous situation detection artificial intelligence model corresponding to the recognized dangerous situation, and providing the information of the selected dangerous situation detection artificial intelligence model to the risk detection unit to be trained.

The artificial intelligence model selection unit includes a risk situation table for each industrial safety factor defining one or more dangerous situations that may occur for each industrial safety factor, and refers to the dangerous situation table for each industrial safety factor, It is characterized in that a dangerous situation detection artificial intelligence model corresponding to each of at least one or more risk situations defined for industrial safety elements is selected, and information of the selected dangerous situation detection artificial intelligence model is provided to a risk detection unit to be trained.

The medium is any one of smoke, gas, water, and petroleum, the safety equipment is a safety helmet, the leakage detecting unit detects leakage of any one or more of the medium, and the safety equipment non-wearing detection unit detects the safety helmet. It is characterized in that the non-wearing detection information is output when a worker not wearing the safety equipment is detected.

The industrial safety and health education contents are characterized in that they are short-time microlearning video contents.

The location detector obtains location information by calculating a location based on location data received from GPS satellites outdoors, and at least one is installed in a monitoring area indoors where location data is not received from GPS satellites to detect its own location information. A worker's mobile terminal that acquires the location information from a beacon sent periodically and transmits the location information and worker location information including worker identification information for a worker working in the surveillance area to the training subject recognition unit, The subject recognition unit includes: a danger occurrence location acquisition unit that receives and outputs surveillance area location information about a surveillance area in which a dangerous situation occurs from the danger detection unit; A worker identification information registration unit for receiving and registering worker identification information including worker information for each worker performing work at the industrial site and terminal identification information of a mobile terminal used by each worker; Receives operator location information from the location detection unit, tracks the location the operator is working by referring to the registered operator identification information, and outputs operator identification information located in the monitoring area of the monitoring area location information upon receiving surveillance area location information. Worker location tracking unit to; And when the monitoring area location information is input from the danger occurrence location acquisition unit, the monitoring area location information is transmitted to the work location tracking unit to receive worker identification information for the operator in the monitoring area of the monitoring area location information and to train the subject. and an education subject identification unit for selecting and providing education subject selection information including the selected worker identification information to the education providing unit.

Industrial safety and health education method through artificial intelligence video control according to the present invention for achieving the above object is: a plurality of video recording units installed in different surveillance areas of an industrial site, corresponding surveillance including at least one industrial safety element An image providing process of transmitting an image taken of an area to an occupational safety and health education providing unit; A worker location providing process in which at least one worker location detection unit installed for each surveillance area transmits worker location information including worker identification information about a worker located in the surveillance region and location information about the location of the worker; and an occupational safety and health education providing unit having monitoring area location information about the location of the monitoring area in which the image capturing unit is installed tracks the location of the worker through the worker position detection unit, and receives images from the plurality of image capturing units. , At least one dangerous situation detection artificial intelligence model corresponding to the industrial safety element existing in the surveillance area where each video is received is applied to monitor whether a dangerous situation for the corresponding industrial safety element is detected, and when a dangerous situation is detected, a dangerous situation After selecting a worker at the location of the monitoring area corresponding to the detected image as a subject for occupational safety and health education, the occupational safety elements of the monitoring area and the detected risk situation are transmitted to the mobile terminal of the worker selected as a subject for occupational safety and health education. It is characterized by including an occupational safety and health education provision process that provides occupational safety and health education contents corresponding to

The process of providing occupational safety and health education includes dangerous situation information, which is labeling information included for each image, and a plurality of images including the corresponding dangerous situation for each dangerous situation in the dangerous situation detection artificial intelligence model learning unit of the occupational safety and health education providing unit. A dangerous situation detection artificial intelligence model learning step of learning an image for each dangerous situation by applying the learning data for each dangerous situation to a dangerous situation detection artificial intelligence model corresponding to the corresponding dangerous situation; The risk detection unit of the occupational safety and health education providing unit includes the learned risk situation detection AI model for each risk situation, selects and sets the risk situation detection AI model to be applied for each monitoring area, and then sets the risk applied for each monitoring area. A risk detection step of monitoring whether a dangerous situation occurs by applying an image corresponding to the surveillance area to the situation detection artificial intelligence model, and outputting surveillance area location information and dangerous situation information about the location of the surveillance area when a dangerous situation is detected; The education subject recognition unit of the industrial safety and health education providing unit tracks the position of the worker through the worker position detection unit, receives the location information of the surveillance area according to the detection of the dangerous situation, and monitors the surveillance area corresponding to the image in which the dangerous situation is detected. A training target recognition step of selecting a worker at the location as an occupational safety and health training target and then outputting worker identification information of the worker selected as the training target; And when the education providing unit of the occupational safety and health education providing unit inputs worker identification information of a worker selected as a training target from the education target recognition unit, the occupational safety element of the monitoring area and It is characterized in that it comprises an education provision step of providing occupational safety and health educational contents corresponding to the dangerous situation of the dangerous situation information.

In the risk detection step, the image acquisition unit of the danger detection unit receives images and identification information of the capturing unit from the image capturing units of each monitoring area, and outputs the corresponding image and identification information of the capturing unit of the image capturing unit that transmitted the image for each monitoring area. image acquisition step; The image acquisition location identification unit of the danger detection unit is provided with a surveillance area location table that each of the image capture units has by mapping capture unit identification information and surveillance area location information of a surveillance area in which the image capture unit is installed, and the image acquisition unit When the identification information of the recording unit is input from the monitoring area location table, the location of the monitoring area is identified by referring to the monitoring area location table, and then the location information of the monitoring area, which is the location information of the monitoring area, is generated and the image acquisition position is output to the training target recognition unit. identification step; an artificial intelligence model selection step in which the artificial intelligence model selection unit of the risk detection unit selects and sets an artificial intelligence model for detecting a dangerous situation to be applied for each surveillance area; And the risk detection unit to be trained in the danger detection unit includes a dangerous situation detection artificial intelligence model for each of the learned dangerous situations, and the surveillance area corresponding to the dangerous situation detection artificial intelligence model selected by the artificial intelligence model selection unit and applied to each surveillance area and monitoring whether a dangerous situation occurs by applying an image corresponding to , and outputting dangerous situation information on the detected dangerous situation to the education providing unit when the dangerous situation is detected.

In the risk detection step for education, the collision detection unit of the education target risk detection unit to which the collision risk situation detection artificial intelligence model is applied applies the image to the collision risk situation detection artificial intelligence model to collide with industrial safety elements in the surveillance area. A collision detection step of monitoring whether a collision occurs and outputting collision detection information, which is dangerous situation information, when a collision with an industrial safety element is detected; The leak detection unit of the education target risk detection unit to which the leakage dangerous situation detection artificial intelligence model is applied applies the image to the leakage dangerous situation detection artificial intelligence model to monitor whether leakage of media for industrial safety elements occurs in the surveillance area, A leakage detection step of outputting leakage detection information, which is dangerous situation information, when media leakage is detected for industrial safety elements; The fire detection unit to which the fire hazard detection artificial intelligence model is applied applies the image to the fire hazard detection artificial intelligence model to monitor whether a fire occurs for industrial safety elements in the surveillance area, and when a fire is detected for industrial safety elements A fire detection step of outputting fire detection information that is dangerous situation information; and the safety equipment detection unit to which the artificial intelligence model is applied to detect dangerous situations in which safety equipment is not worn applies the image to the artificial intelligence model to detect dangerous situations in which safety equipment is not worn, and monitors whether or not non-wearing of safety equipment for industrial safety elements is detected in the surveillance area. And, in the case of detecting non-wearing of safety equipment, during the safety equipment wearing detection step of outputting safety equipment non-wearing detection information, which is dangerous situation information for industrial safety elements, the step of applying the artificial intelligence model selected in the artificial intelligence model selection step is selectively performed. characterized by

In the step of detecting danger for education, a dangerous object detection unit to which a dangerous object detection artificial intelligence model is applied applies the image to the dangerous object detection artificial intelligence model to detect and detect a dangerous object in which a dangerous situation may occur in the image. Further comprising a dangerous object detection step of providing information on the type and location of the dangerous object to the collision detection unit, wherein in the education target risk detection step, the collision detection unit detects a dangerous situation in the object detected by the dangerous object detection unit. It is characterized by monitoring whether this occurs.

The location detector obtains location information by calculating a location based on location data received from GPS satellites outdoors, and at least one is installed in a monitoring area indoors where location data is not received from GPS satellites to detect its own location information. A worker's mobile terminal that acquires the location information from a beacon sent periodically and transmits the location information and worker location information including worker identification information for a worker working in the surveillance area to the training subject recognition unit, In the training target recognition step, the worker identification information register of the training target recognition unit registers worker identification information including worker information for each worker performing work at the industrial site and terminal identification information of a mobile terminal used by each worker. Operator identification information registration step of receiving input and registering; a danger occurrence location acquisition step of receiving and outputting surveillance area location information about a surveillance area in which a dangerous situation occurs from the danger detection unit by the location acquisition unit of the education subject recognition unit; The operator location tracking unit of the training target recognition unit receives the location information of the operator from the location detection unit and tracks the location where the operator is working by referring to the registered operator identification information. Worker location tracking step of outputting worker identification information located in the monitoring area; And when the education target identification unit of the education target recognition unit transmits the monitoring area location information to the work location tracking unit when the monitoring area location information is input from the danger occurrence location acquisition unit, It is characterized in that it includes an education target identification step of receiving worker identification information, selecting a training target, and providing training target selection information including the selected worker identification information to an education providing unit.

The present invention applies a plurality of risk situation artificial intelligence models learned for each of a plurality of risk situations that may occur in industrial sites, and selectively applies the risk situation artificial intelligence model corresponding to industrial safety elements monitored by acquired images. Therefore, it has the effect of effectively detecting various risk situations that may occur in industrial sites.

In addition, the present invention tracks the location of workers and managers through mobile terminals of workers and managers and indoor positioning devices such as beacons and access points (APs), and for each occupational safety factor of the pre-registered surveillance area. Identifies the manager of industrial safety elements through manager information and defines the location of the surveillance area where industrial safety elements exist in advance, so the effect of identifying workers and managers existing in the surveillance area with industrial safety elements in a dangerous situation there is

In addition, since the present invention can identify a worker existing in a monitoring area in which a dangerous situation occurs and a manager who manages industrial safety elements in the monitoring area, the operator and manager can receive the optimum response to the dangerous situation occurring in the industrial safety element. Occupational safety and health education contents are provided, so the efficiency of education can be maximized.

In addition, since the present invention provides occupational safety and health education contents as microlearning contents in a short time, workers and managers can immediately watch them at industrial sites, so there is an effect of improving the efficiency of education.

1 is a diagram showing the configuration and installation area concept of an occupational safety and health education system through artificial intelligence video control according to the present invention.
2 is a diagram showing a block configuration diagram of an occupational safety and health education system through artificial intelligence video control according to the present invention.
3 is a diagram showing the detailed configuration of the risk detection unit of the occupational safety and health education system according to the present invention.
4 is a diagram showing the configuration of an education subject recognition unit of an occupational safety and health education system according to the present invention.
5 is a flowchart illustrating an occupational safety and health education method through artificial intelligence image control according to an embodiment of the present invention.
6 is a flowchart illustrating a risk detection method through video among industrial safety and health education methods through artificial intelligence video control according to an embodiment of the present invention.

Hereinafter, with reference to the accompanying drawings, the configuration and operation of the occupational safety and health education system through artificial intelligence video control according to the present invention will be described in detail, and the occupational safety and health education method in the system will be described.

1 is a diagram showing the configuration and installation area concept of an occupational safety and health education system through artificial intelligence video control according to the present invention, and FIG. 2 is a block configuration of an occupational safety and health education system through artificial intelligence video control according to the present invention. It is a drawing showing the figure. It will be described with reference to FIGS. 1 and 2 below.

Reference numeral 1 in the drawings indicates an industrial site, 2 indicates a monitoring area monitored by the image capture unit 20, and 3 is defined by the position of the operator position detection unit 30 present in the monitoring area. It shows the location recognition area. The surveillance area 2 will contain at least one occupational safety element to be monitored.

The above industrial safety elements refer to objects and places to be monitored for industrial safety, and include construction and transportation equipment, facilities (pipes (water, gas, oil, etc. flowing), pillars, etc.), materials, high It could be the work place of the place, the worker and the worker's safety equipment (helmet, seat belt), etc.

In addition, the dangerous situation in the present invention is a dangerous situation or the possibility of a dangerous situation that may occur due to the industrial safety factor, the possibility of collision and collision between equipment, the possibility of collision and collision between equipment and workers, the possibility of falling and falling of materials, Fall and fall of workers, workers not wearing safety equipment (helmet, seat belt), leakage of media such as gas, water, oil, fire and fire (smoke and flame), equipment toppling and falling (tilting of equipment) ) and so on.

The occupational safety and health education system through artificial intelligence video control according to the present invention includes a plurality of image capture units 20, a plurality of worker location detection units 30, and an occupational safety and health education providing unit 50.

The image capturing unit 20, the worker location detection unit 30, and the occupational safety and health education providing unit 50 are connected through a wired/wireless data communication network 10 to perform mutual data communication.

The wired/wireless data communication network 10 combines at least one of an Internet network including a WiFi network and a local network, a mobile communication network including 3G (3G), 4G, and 5G, and a WiBro network. It could be a data communication network.

The image capture unit 20 is installed at a location where there are industrial safety elements in the industrial site 1 to form a surveillance area 2, and transmits an image of the surveillance area 2 through the wired/wireless data communication network 10. Through this, it is provided to the occupational safety and health education providing unit 50.

The image capture unit 20 may be an image acquisition device such as a Closed Circuit Television (CCTV).

The worker location detection unit 30 measures the current location of the worker, generates worker location information for the measured current location of the worker, and transmits it to the occupational safety and health education providing unit 50 .

The worker position detection unit 30 may be composed of only the mobile terminal 31 of the worker, or may include the mobile terminal 31 and the indoor positioning device 32.

The mobile terminal 31 determines its current location based on location data received from at least three or more GPS satellites through a Global Positioning System (GPS) module when a worker is outdoors in the industrial site 1. After calculation, worker location information will be generated and provided to the occupational safety and health education providing unit 50 .

In addition, when indoors, the mobile terminal 31 applies a general location tracking method using a mobile communication network base station of the wired/wireless data communication network 10 to measure its current location and then generates worker location information to ensure industrial safety. It may be provided to the health education providing unit 50, or the occupational safety and health education providing unit 50 may provide worker location information about the worker's location using the indoor positioning device 32 of the present invention. .

The indoor positioning device 32 is a device whose device identification information and location are mapped and registered in advance in the occupational safety and health education providing unit 50, and may be a beacon that periodically sends a beacon signal including device identification information, It may be an access point (AP).

In the case of an AP, the occupational safety and health education provider 50 maps and stores the MAC address of the worker's mobile terminal with worker information, and collects the MAC address of the mobile terminal 32 connected to the AP to locate the AP. It will be determined as the location of the worker who has the mobile terminal 32 having the collected Mac address.

In addition, the occupational safety and health education providing unit 50 receives the same Mac address from at least three or more APs, and each AP uses the triangulation method based on the received signal strength (RSSI) of the signal received from the mobile terminal 31. It may also be configured to measure the position of the mobile terminal 31, that is, the operator.

The occupational safety and health education providing unit 50 includes a risk detection artificial intelligence model learning unit 60, a risk detection unit 70, a training subject recognition unit 80, an education providing unit 90, and a management unit 100, Processes the overall operation of the occupational safety and health education system through artificial intelligence video control according to the present invention.

That is, the occupational safety and health education providing unit 50 has monitoring area location information about the location of the monitoring area where the video capture unit 20 is installed, and tracks the location of the worker through the worker location detection unit 30. and receives images from the plurality of image capture units 20, and applies at least one dangerous situation detection artificial intelligence model corresponding to the industrial safety elements existing in the surveillance area where each image is received to detect the corresponding industrial safety elements. After monitoring whether a dangerous situation is detected, and when a dangerous situation is detected, a worker in the position of the monitoring area corresponding to the video in which the dangerous situation is detected is selected as a subject for occupational safety and health education, and then the worker selected as a subject for occupational safety and health education Occupational safety and health education contents corresponding to industrial safety elements in the monitoring area and detected risk situations are provided to the mobile terminal 31 of the monitoring area.

More specifically, the danger detection artificial intelligence model learning unit 60 includes a plurality of images including the corresponding dangerous situation (occurrence or prior to occurrence) for each dangerous situation and dangerous situation, which is information labeled for the dangerous situation of each image. The dangerous situation-specific learning data set containing the information is applied to the dangerous situation detection artificial intelligence model corresponding to the corresponding dangerous situation, and the image for each dangerous situation is learned.

The dangerous situation detection artificial intelligence model includes a collision detection artificial intelligence model that detects a collision between two objects, which is an industrial safety element, a leak detection artificial intelligence model that detects media leakage in an industrial safety element, and a fire detection in an industrial safety element. A fire detection artificial intelligence model that detects whether or not safety equipment is worn in an industrial safety element, a fall detection artificial intelligence model that detects a fall of an industrial safety element in an industrial safety element, a collapse of an industrial safety element In addition, a collapse detection artificial intelligence model that detects a fall may be applied.

The danger detection unit 70 includes the learned dangerous situation detection artificial intelligence models for each dangerous situation, selects and sets a dangerous situation detection artificial intelligence model to be applied for each surveillance area, and then applies the dangerous situation detection artificial intelligence model for each surveillance area. It monitors whether a dangerous situation occurs by applying the image corresponding to the surveillance area to the intelligence model, and outputs surveillance area location information and dangerous situation information about the location of the surveillance area when a dangerous situation is detected.

The education target recognition unit 80 tracks the position of the operator through the operator position detection unit 30, receives the monitoring area location information according to the detection of a dangerous situation, and monitors the surveillance area corresponding to the image in which the dangerous situation is detected. After selecting a worker at the location as a target for occupational safety and health education, the worker identification information of the worker selected as a target for training is output.

The training subject recognition unit 80 may also select a manager set in advance to be in charge of managing the dangerous situation and occupational safety factors as a training subject.

The education provider 90 stores and manages occupational safety and health education contents for each risk situation and industrial safety factor, and when the worker identification information of a worker selected as a training target is input from the training target recognition unit 80 Occupational safety and health education contents corresponding to the industrial safety elements of the monitoring area and dangerous situations of the dangerous situation information are provided to the worker's mobile terminal 31 corresponding to the worker identification information.

Since the occupational safety and health education contents are provided directly to the mobile terminal 31 of workers and managers who are in danger in real time, it is preferable to apply microlearning contents having a relatively short runtime. Preferably, the content is a video. Preferably, the runtime is 15 to 30 minutes.

The management unit 100 provides a management means capable of setting the location of a monitoring area for device identification information of the video capture unit 20 from a manager or registering and managing worker information including terminal identification information of a worker's mobile terminal. and provides management means to perform addition and re-learning of artificial intelligence models.

3 is a diagram showing the detailed configuration of the risk detection unit of the occupational safety and health education system according to the present invention.

Referring to FIG. 3, the risk detection unit 70 includes an image acquisition unit 71, an image acquisition location identification unit 72, an artificial intelligence model selection unit 73, and an education target risk detection unit 74. According to an example, a dangerous object detection unit 76 and a danger notification unit 77 may be further included.

The image acquisition unit 71 is connected to a plurality of image capturing units 20 through a wired/wireless data communication network 10 and receives images and capturing unit identification information from the image capturing units 20 in each monitoring area. Corresponding images for each monitoring area and identification information of the capturing unit of the image capturing unit 20 that transmitted the image are output.

The image acquisition location identification unit 72 maps the capturing unit identification information for each of the image capturing units 20 and the surveillance area location information of the surveillance area in which the image capturing unit 20 is installed, and has the location of the surveillance area. A table is provided, and when capturing unit identification information is input from the image acquiring unit 71, the monitoring area location table is referred to to find monitoring area location information corresponding to the capturing unit identification information, and the monitoring area is identified. Area location information is output to the education target recognition unit 80 .

The artificial intelligence model selector 73 selects and sets an artificial intelligence model for detecting a dangerous situation to be applied for each monitoring area.

There may be a manager selection method, an artificial intelligence model selection method, and the like as a method for selecting a dangerous situation detection artificial intelligence model for each monitoring area.

The manager selection method is a method in which a manager directly selects an artificial intelligence model for detecting a dangerous situation for each monitoring area.

In the artificial intelligence model selection method, the artificial intelligence model selection unit has a risk situation table for each industrial safety factor defining one or more dangerous situations that may occur for each industrial safety factor, and the dangerous situation table for each industrial safety factor is referred to. It is a method of selecting a dangerous situation detection artificial intelligence model corresponding to each of at least one or more dangerous situations defined for industrial safety elements existing in the monitoring area. Industrial safety factors for each monitoring area should be defined in advance or configured to identify industrial safety factors in the monitoring area by applying an image of the corresponding monitoring area to a separate industrial safety factor classification artificial intelligence model.

The risk detection unit 74 to be trained includes a plurality of risk situation detection units 75 to which the above-described dangerous situation detection artificial intelligence model for each dangerous situation has been applied, and is selected by the artificial intelligence model selection unit 76 It monitors whether a dangerous situation occurs by applying an image corresponding to the surveillance area to the dangerous situation detection artificial intelligence model applied to each surveillance area, and when a dangerous situation is detected, the dangerous situation information on the detected dangerous situation is provided to the education providing unit (90 ) is output as

In FIG. 3, only the collision detection unit 75-1, leakage detection unit 75-2, fire detection unit 75-3, and safety device wearing detection unit 75-4 are used as the dangerous situation detection unit 75. Although shown, more components of the detection unit will be included, such as a fall detection unit applied with an artificial intelligence model for detecting a fall risk situation, and a collapse detection unit applied with an artificial intelligence model for detection of a collapse risk situation.

The collision detection unit 75-1 applies the artificial intelligence model for detecting a risk of collision, and monitors whether a collision with an industrial safety element occurs in the surveillance area by applying the image to the artificial intelligence model for detecting a risk of collision. When a collision with a safety element is detected, the collision detection information, which is dangerous situation information, is output to the danger notification unit 77 or to the education providing unit 77, depending on the embodiment.

The leakage detection unit 75-2 is applied with an artificial intelligence model for detecting dangerous leakage situations, and monitors whether leakage of media for industrial safety elements occurs in the surveillance area by applying the image to the artificial intelligence model for detecting dangerous leakage situations. , When media leakage is detected for industrial safety elements, leakage detection information, which is dangerous situation information, is output to the danger notification unit 77 or to the education providing unit 77 according to the embodiment.

The fire detection unit 75-3 applies the fire risk situation detection artificial intelligence model, and monitors whether a fire occurs for industrial safety elements in the surveillance area by applying the image to the fire risk situation detection artificial intelligence model. When a fire is detected for a safety element, fire detection information, which is dangerous situation information, is output to the danger notification unit 77 or to the education providing unit 77, depending on the embodiment.

The safety equipment detection unit 75-4 applies the artificial intelligence model for detecting dangerous situations in which safety equipment is not worn, and applies the image to the artificial intelligence model for detecting dangerous situations in which safety equipment is not worn, so that safety equipment for industrial safety elements in the monitoring area is applied. It monitors whether non-wearing is detected, and outputs safety equipment non-wearing detection information, which is dangerous situation information about industrial safety elements, to the danger notification unit 77 or to the education providing unit 77 depending on the embodiment when non-wearing of safety equipment is detected. .

The danger notification unit 77 receives monitoring area location information from the image acquisition location identification unit 72 and receives detection information, which is at least one piece of dangerous situation information, from the education target risk detection unit, and converts the surveillance area location information to the education target recognition unit. (80), and the detection information is output to the education providing unit (90).

4 is a diagram showing the configuration of an education subject recognition unit of an occupational safety and health education system according to the present invention.

Referring to FIG. 4 , the education target recognition unit 80 includes a danger occurrence location acquisition unit 81, a worker identification information registration unit 82, a worker location tracking unit 83, and an education target identification unit 84.

The danger occurrence location acquisition unit 81 receives monitoring area location information from the danger detection unit 70 and outputs it to the education subject identification unit 84 .

The worker identification information registration unit 82 provides a means for registering worker information to the mobile terminal 31 of a manager or worker, and stores worker identification information including terminal identification information of the mobile terminal 31 through the worker information registration means. to register

The operator identification information may be terminal identification information or may be separate identification information.

The terminal identification information may include any one or more of a Mac address and phone number of the mobile terminal 31, but preferably includes the phone number.

Worker location tracking unit 83 receives the worker location information including the location information and worker identification information measured through the GPS module of the mobile terminal 31 of the worker from the worker location detection unit 20, and the worker location information of the worker A worker is identified by comparing identification information with worker identification information of worker information for registered workers, and the location of the worker is tracked by the worker's location information for the worker's identification information of the identified worker.

The worker location tracking unit 83 receives monitoring area location information from the education subject identification unit 84, tracks the location of the surveillance area location information, that is, workers existing in the monitoring area, and provides information about the tracked workers. Worker identification information is provided to the training subject identification unit 84 .

When the monitoring area location information is input from the risk occurrence location acquisition unit 81, the education target identification unit 84 transmits the monitoring area location information to the work location tracking unit 83 to monitor the surveillance area location information. Receives worker identification information about a worker in the , selects a training target, and provides training target selection information including the selected worker identification information to the education providing unit 90 .

On the other hand, if the monitoring area location information is input from the risk occurrence location acquisition unit 81, the education target identification unit 84 selects the monitoring area 2 or monitoring area 2 from a previously registered surveillance area or a list of persons in charge for each industrial safety element. ), and the manager identification information of the manager may be further included in the education subject information and transmitted to the education providing unit 90.

5 is a flowchart showing an occupational safety and health education method through artificial intelligence video control according to the present invention.

Referring to FIG. 5, the occupational safety and health education providing unit 50 is the location (region) of the monitoring area 2 defined by the location of the video recording unit 20 in the industrial site 1, an indoor location such as an AP or a beacon. The location of the confirmation device 32 is registered and defined (S111).

In addition, the occupational safety and health education provider 50 registers worker identification information of workers and managers (hereinafter referred to as "workers" for convenience of description) working in the industrial site 1 (S113). The worker identification information will include terminal identification information and worker information of the mobile terminal 31 of the worker. The terminal identification information includes the phone number of the mobile terminal 31, and may further include other identification information such as a MAC address according to an embodiment. The Mac address may be any one or more of a Wi-Fi Mac address, a Bluetooth beacon Mac address, and the like.

When the location registration and worker registration of the system configuration are completed, the occupational safety and health education provider 50 applies the collected learning data set for each dangerous situation to the corresponding dangerous situation detection artificial intelligence model and inputs the dangerous situation detection artificial intelligence model. It learns to detect the corresponding dangerous situation for the image to be used (S115).

After learning the artificial intelligence model, the occupational safety and health education providing unit 50 starts tracking the location of workers through the mobile terminal 31 of the registered workers through the worker location detection unit 30 (S117).

When the position of the worker starts to be tracked, the occupational safety and health education provider 50 starts acquiring images for each monitoring area through the image capture unit 20 (S119).

When the images for each monitoring area 2 start to be acquired, the occupational safety and health education providing unit 50 monitors whether a dangerous situation is detected in an arbitrary monitoring area 2 (S121). Here, detection of a dangerous situation includes not only detecting the occurrence of a dangerous situation but also detecting the possibility of occurrence of a dangerous situation.

When a dangerous situation is detected, the occupational safety and health education providing unit 50 acquires surveillance area location information for the surveillance area 2 where the dangerous situation is detected (S123), and the location information located in the surveillance area 2 where the dangerous situation is detected Workers are extracted and selected as training subjects (S125).

When the education target is selected, the occupational safety and health education provider 50 determines occupational safety and health education contents corresponding to the detected risk situation and occupational safety elements of the detection area, and maps and stores the education target (S127).

When the occupational safety and health education contents are determined, the occupational safety and health education provision unit 50 notifies the possibility of occurrence of a dangerous situation and the possibility of occurrence of a dangerous situation, and the dangerous situation notification and educational contents requesting the participation of the determined occupational safety and health education contents. The course enrollment request notification message is transmitted to the mobile terminal 31 of the operator selected as the training subject. The dangerous situation notification and the educational content enrollment request notification message may include a link for downloading and viewing the occupational safety and health education content. In addition, when the occupational safety and health education application (or referred to as "app") according to the present invention is installed in the mobile terminal 31, the dangerous situation notification and education content enrollment request notification message is a pop-up message through an app, an instant It may be provided in the form of a message, or may be provided in a mobile communication message such as a short message service (SMS) message, a long message service (LMS) message, or a multimedia message service (MMS) message.

After transmitting the dangerous situation notification and the education content enrollment request notification message, the occupational safety and health education providing unit 50 monitors whether a course enrollment request is generated by the mobile terminal 31 of the worker who is the education target (S131), If the course enrollment request does not occur until a certain time has elapsed (S133), a notification message for an increase in the educational content course enrollment time according to the timeout is sent and the course is re-requested (S135).

When a class enrollment request is generated, the occupational safety and health education provider 50 transmits the corresponding microlearning education content to the mobile terminal 31 of the worker who is the education target (S137).

6 is a flowchart illustrating a risk detection method through video among industrial safety and health education methods through artificial intelligence video control according to an embodiment of the present invention.

Referring to FIG. 6, the risk detection unit 70 of the occupational safety and health education providing unit 50 acquires an image through the image capturing unit 20 (S211), and the image capturing unit included in the acquired image ( The monitoring area 2 is identified by the identification information of the capturing unit of 20) and the location information of the image capturing unit 20 registered in advance (S213).

When the surveillance area is identified, the danger detection unit 70 applies the image of the identified surveillance area 2 to the industrial safety factor classification artificial intelligence model to identify the industrial safety factor of the surveillance area 2 (S215). . Since the industrial safety elements of the monitoring area 2 may change over time, it may be configured to re-identify the industrial safety elements at regular time intervals.

When the industrial safety factor is identified, the risk detection unit 70 selects at least one dangerous situation detection artificial intelligence model corresponding to the identified industrial safety factor (S217).

When at least one dangerous situation detection artificial intelligence model is selected, the danger detection unit 70 applies the image of the surveillance area 2 to the selected danger situation detection artificial intelligence model (S219), and checks whether a dangerous situation is detected in the image. is monitored (S221).

When a dangerous situation is detected, the danger detecting unit 70 outputs surveillance area location information and dangerous situation information (sensing information) to the training subject recognition unit 80 and the education providing unit 90 (S223).

On the other hand, it is common knowledge in the art that the present invention is not limited to the above-described typical preferred embodiments, but can be implemented by various improvements, changes, substitutions, or additions within the scope of the present invention. If you have the , you can easily understand. If the implementation by such improvement, change, substitution or addition falls within the scope of the appended claims below, the technical idea should also be regarded as belonging to the present invention.

1: Industrial site 2: Surveillance area
10: wired/wireless data communication network 20: video recording unit
30: operator position detection unit 31: mobile terminal
32: indoor positioning device 50: occupational safety and health education provision unit
60: risk detection artificial intelligence model learning unit
70: danger detection unit 71: image acquisition unit
72: image acquisition location identification unit 73: artificial intelligence model selection unit
74: Risk detection unit for education 75: Risk situation detection unit
76: dangerous object detection unit 77: danger notification unit
80: training subject recognition unit 81: danger location acquisition unit
82: worker identification information registration unit 83: worker location tracking unit
84: Education target identification unit 90: Education provision unit
100: management unit

Claims (16)

A plurality of image capture units for transmitting images obtained by capturing different surveillance areas of an industrial site including at least one or more industrial safety elements;
a worker location detection unit installed at least one for each surveillance area and transmitting worker location information including worker identification information about a worker located in the surveillance region and location information about the location of the worker; and
It has surveillance area location information about the location of the surveillance area where the video capture unit is installed, tracks the location of the worker through the operator location detection unit, receives images from the plurality of image capture units, and receives each image. At least one dangerous situation detection artificial intelligence model corresponding to industrial safety elements existing in the monitoring area is applied to monitor whether a dangerous situation is detected for the corresponding industrial safety factor, and when a dangerous situation is detected, respond to the video in which the dangerous situation is detected. After selecting a worker in the position of the monitoring area as a target for occupational safety and health education, the mobile terminal of the worker selected as the subject for occupational safety and health education responds to occupational safety elements and detected risk situations in the monitoring area. An occupational safety and health education system through artificial intelligence video control, comprising an occupational safety and health education provider providing educational contents.
According to claim 1,
The occupational safety and health education provision department,
By applying learning data for each dangerous situation, including a plurality of images including the corresponding dangerous situation for each dangerous situation and dangerous situation information, which is labeling information included in each image, to the dangerous situation detection artificial intelligence model corresponding to the dangerous situation, each image A dangerous situation detection artificial intelligence model learning unit that learns images for each dangerous situation;
After selecting and setting the dangerous situation detection artificial intelligence model to be applied to each surveillance area, including the learned dangerous situation detection artificial intelligence model for each danger situation, the dangerous situation detection artificial intelligence model applied for each surveillance area is applied to the corresponding surveillance area. a danger detection unit that monitors whether a dangerous situation occurs by applying a corresponding image, and outputs monitoring area location information and dangerous situation information about the location of the surveillance area when a dangerous situation is detected;
The position of the worker is tracked through the worker position detection unit, and the position information of the surveillance region is input according to the detection of the dangerous situation, and the worker in the position of the surveillance region corresponding to the image in which the dangerous situation is detected is made a subject of occupational safety and health education. After selecting the education target recognition unit for outputting the worker identification information of the worker selected as the training target; and
When the worker identification information of the worker selected as the training target is input from the training target recognition unit, the worker's mobile terminal corresponding to the worker identification information corresponds to the occupational safety element of the monitoring area and the risk situation of the risk situation information. Industrial safety and health education system through artificial intelligence video control, characterized in that it comprises an education provider providing safety and health education contents.
According to claim 2,
The danger detection unit,
an image acquiring unit that receives images and capturing unit identification information from image capturing units of each monitoring area and outputs corresponding images and capturing unit identification information of the image capturing unit that has transmitted the images for each surveillance area;
A monitoring region location table having mapping unit identification information for each of the image capturing units and surveillance region location information of a surveillance region in which the image capturing unit is installed is provided, and when the capturing unit identification information is input from the image acquisition unit, an image acquisition location identification unit for identifying the position of the surveillance region by looking for surveillance region location information corresponding to the photographing unit identification information with reference to the surveillance region location table, and then outputting the surveillance region location information to the training target recognition unit;
an artificial intelligence model selection unit that selects and sets an artificial intelligence model for detecting a dangerous situation to be applied to each surveillance area; and
Including the learned dangerous situation detection artificial intelligence model for each dangerous situation, and applying the image corresponding to the surveillance area to the dangerous situation detection artificial intelligence model selected by the artificial intelligence model selector and applied to each surveillance area, the dangerous situation is detected. Occupational safety and health education system through artificial intelligence video control, characterized in that it includes a risk detection unit for monitoring whether a dangerous situation occurs and outputting dangerous situation information on the detected dangerous situation to the education providing unit when detecting a dangerous situation.
According to claim 3,
The education target risk detection unit,
A collision risk situation detection artificial intelligence model is applied, and the image is applied to the collision risk situation detection artificial intelligence model to monitor whether a collision with an industrial safety element occurs in the surveillance area, and dangerous situation information when a collision with an industrial safety element is detected. a collision detection unit that outputs collision detection information;
An artificial intelligence model for detecting a risky leakage situation is applied, and the image is applied to the artificial intelligence model for detecting a risky leakage situation to monitor whether leakage of media for industrial safety elements occurs in the surveillance area, and to detect media leakage for industrial safety elements. a leakage detection unit outputting leakage detection information, which is information about a dangerous situation;
A fire risk situation detection artificial intelligence model is applied, and the image is applied to the fire risk situation detection artificial intelligence model to monitor whether a fire occurs for industrial safety elements in the surveillance area, and a dangerous situation when a fire is detected for industrial safety elements. a fire detection unit outputting fire detection information, which is information; and
A dangerous situation detection artificial intelligence model for not wearing safety equipment is applied, and the above image is applied to the artificial intelligence model for detecting a dangerous situation without wearing safety equipment to monitor whether the unwearing of safety equipment for industrial safety elements is detected in the surveillance area, and the unwearing of safety equipment An occupational safety and health education system through artificial intelligence video control, characterized in that it includes a safety equipment wearing detection unit that outputs safety equipment non-wearing detection information, which is dangerous situation information on industrial safety elements when detected.
According to claim 4,
The education target risk detection unit,
A dangerous object detection artificial intelligence model is applied, and the image is applied to the dangerous object detection artificial intelligence model to detect a dangerous object in which a dangerous situation may occur in the image, and to obtain information on the type and location of the detected dangerous object. Further comprising a dangerous object detection unit provided to the collision detection unit,
The education target risk detection unit,
Industrial safety and health education system through artificial intelligence video control, characterized in that for monitoring whether a dangerous situation occurs in the object detected by the dangerous object detection unit.
According to claim 4,
The artificial intelligence model selection unit,
It includes a risk situation table for each industrial safety factor that defines one or more dangerous situations that may occur for each industrial safety factor, and has an artificial intelligence model selection artificial intelligence model, and selects the artificial intelligence model from the image input from the image acquisition unit. It is applied to the artificial intelligence model to classify industrial safety factors that exist in the monitoring area of the video, recognize one or more dangerous situations corresponding to the classified industrial safety factors from the table of dangerous situations for each industrial safety factor, and recognize the recognized dangerous situations. An occupational safety and health education system through artificial intelligence video control, characterized in that for selecting a dangerous situation detection artificial intelligence model corresponding to, and providing information of the selected dangerous situation detection artificial intelligence model to the training target risk detection unit.
According to claim 6,
The artificial intelligence model selection unit,
Include a risk situation table for each occupational safety factor that defines one or more hazardous situations that may occur for each occupational safety factor;
With reference to the risk situation table for each industrial safety factor, a dangerous situation detection artificial intelligence model corresponding to each of at least one or more dangerous situations defined for the industrial safety elements existing in the monitoring area is selected, and the selected dangerous situation detection artificial intelligence An occupational safety and health education system through artificial intelligence image control, characterized in that it provides model information to the risk detection unit for education.
According to claim 3,
The medium is any one of smoke, gas, water, and petroleum,
The safety equipment is a safety helmet,
The leakage detection unit,
detecting leakage of one or more of the media;
The safety equipment non-wearing detection unit,
An occupational safety and health education system through artificial intelligence video control, characterized in that for outputting the non-wearing detection information of the safety equipment when detecting a worker not wearing the safety helmet.
According to claim 1,
The occupational safety and health education contents,
Industrial safety and health education system through artificial intelligence video control, characterized in that it is a short-time microlearning video content.
According to claim 2,
The position detection unit,
Outdoors, a beacon obtains location information by calculating a location based on location data received from GPS satellites, and periodically transmits its own location information by installing at least one surveillance area indoors where location data is not received from GPS satellites. A worker's mobile terminal that obtains the location information from and transmits the location information and worker location information including worker identification information for a worker working in the surveillance area to the training subject recognition unit,
Education Target Recognition Division,
a danger occurrence location acquisition unit for receiving and outputting surveillance area location information about a surveillance area in which a dangerous situation occurs from the danger detection unit;
A worker identification information registration unit for receiving and registering worker identification information including worker information for each worker performing work at the industrial site and terminal identification information of a mobile terminal used by each worker;
Receives operator location information from the location detection unit, tracks the location the operator is working by referring to the registered operator identification information, and outputs operator identification information located in the monitoring area of the monitoring area location information upon receiving surveillance area location information. Worker location tracking unit to; and
When monitoring area location information is input from the danger occurrence location acquisition unit, the monitoring area location information is transmitted to the work location tracking unit to receive worker identification information for workers in the surveillance area of the surveillance area location information and select training targets. And an occupational safety and health education system through artificial intelligence video control, characterized in that it comprises an education subject identification unit for providing education subject selection information including the selected worker identification information to the education providing unit.
An image providing process in which a plurality of image capture units installed in different surveillance areas of an industrial site transmit images captured in the corresponding surveillance area including at least one industrial safety element to an occupational safety and health education providing unit;
A worker location providing process in which at least one worker location detection unit installed for each surveillance area transmits worker location information including worker identification information about a worker located in the surveillance region and location information about the location of the worker; and
An occupational safety and health education providing unit having surveillance area location information about the location of a surveillance area in which the image capture unit is installed tracks the location of the worker through the worker position detection unit, and receives images from the plurality of image capture units, At least one dangerous situation detection artificial intelligence model corresponding to the industrial safety element existing in the surveillance area where each video is received is applied to monitor whether a dangerous situation is detected for the corresponding industrial safety element, and when a dangerous situation is detected, a dangerous situation is detected. After selecting a worker at the location of the monitoring area corresponding to the detected image as a subject for occupational safety and health education, the occupational safety elements of the monitoring area and the detected risk situation are monitored through the mobile terminal of the worker selected as a subject for occupational safety and health education. An occupational safety and health education method through artificial intelligence video control, comprising a process of providing occupational safety and health education that provides corresponding occupational safety and health education contents.
According to claim 11,
The process of providing occupational safety and health education,
The risk situation detection artificial intelligence model learning unit of the occupational safety and health education providing unit converts learning data for each risk situation including a plurality of images including the corresponding risk situation for each risk situation and dangerous situation information, which is labeling information included for each image, to the corresponding risk A dangerous situation detection artificial intelligence model learning step of applying the dangerous situation detection artificial intelligence model corresponding to the situation and learning an image for each dangerous situation;
The risk detection unit of the occupational safety and health education providing unit includes the learned risk situation detection AI model for each risk situation, selects and sets the risk situation detection AI model to be applied for each monitoring area, and then sets the risk applied for each monitoring area. A risk detection step of monitoring whether a dangerous situation occurs by applying an image corresponding to the surveillance area to the situation detection artificial intelligence model, and outputting surveillance area location information and dangerous situation information about the location of the surveillance area when a dangerous situation is detected;
The education subject recognition unit of the industrial safety and health education providing unit tracks the position of the worker through the worker position detection unit, receives the location information of the surveillance area according to the detection of the dangerous situation, and monitors the surveillance area corresponding to the image in which the dangerous situation is detected. A training target recognition step of selecting a worker at the location as an occupational safety and health training target and then outputting worker identification information of the worker selected as the training target; and
When the education providing unit of the occupational safety and health education providing unit inputs worker identification information of a worker selected as a training target from the education target recognition unit, the occupational safety element of the monitoring area and the An occupational safety and health education method through artificial intelligence video control, comprising an education provision step of providing occupational safety and health education contents corresponding to dangerous situations of dangerous situation information.
According to claim 12,
The risk detection step,
An image acquisition step in which the image acquisition unit of the danger detection unit receives images and capturing unit identification information from image capturing units of each monitoring area and outputs corresponding images and capturing unit identification information of the image capturing unit that transmitted the images for each monitoring area;
The image acquisition location identification unit of the danger detection unit is provided with a surveillance area location table that each of the image capture units has by mapping capture unit identification information and surveillance area location information of a surveillance area in which the image capture unit is installed, and the image acquisition unit When the identification information of the recording unit is input from the monitoring area location table, the location of the monitoring area is identified by referring to the monitoring area location table, and then the location information of the monitoring area, which is the location information of the monitoring area, is generated and the image acquisition position is output to the training target recognition unit. identification step;
an artificial intelligence model selection step in which the artificial intelligence model selection unit of the risk detection unit selects and sets an artificial intelligence model for detecting a dangerous situation to be applied for each surveillance area; and
The risk detection unit to be trained in the danger detection unit includes the learned dangerous situation detection artificial intelligence model for each dangerous situation, and the dangerous situation detection artificial intelligence model selected by the artificial intelligence model selection unit and applied to each surveillance area is applied to the corresponding surveillance area. An artificial intelligence image comprising a risk detection step for education subject to monitor whether a dangerous situation occurs by applying a corresponding image and output dangerous situation information on the detected dangerous situation to the education providing unit when a dangerous situation is detected. Occupational safety and health education method through control.
According to claim 13,
In the step of detecting the risk of the education target,
The collision detection unit of the training subject risk detection unit to which the collision risk situation detection artificial intelligence model is applied applies the image to the collision risk situation detection artificial intelligence model to monitor whether a collision with an industrial safety element occurs in the surveillance area, and to monitor whether a collision with an industrial safety element occurs. A collision detection step of outputting collision detection information that is dangerous situation information upon detection of a collision with respect to;
The leak detection unit of the education target risk detection unit to which the leakage dangerous situation detection artificial intelligence model is applied applies the image to the leakage dangerous situation detection artificial intelligence model to monitor whether leakage of media for industrial safety elements occurs in the surveillance area, A leakage detection step of outputting leakage detection information, which is dangerous situation information, when media leakage is detected for industrial safety elements;
The fire detection unit to which the fire hazard detection artificial intelligence model is applied applies the image to the fire hazard detection artificial intelligence model to monitor whether a fire occurs for industrial safety elements in the surveillance area, and when a fire is detected for industrial safety elements A fire detection step of outputting fire detection information that is dangerous situation information; and
The safety equipment detection unit to which the safety equipment detection artificial intelligence model is applied applies the image to the artificial intelligence model to detect the dangerous situation without safety equipment, and monitors whether or not the safety equipment for industrial safety elements is detected in the surveillance area. Selectively performing the step of applying the artificial intelligence model selected in the artificial intelligence model selection step among the safety equipment detection steps of outputting safety equipment non-wear detection information, which is dangerous situation information on industrial safety elements when non-wearing of safety equipment is detected Occupational safety and health education method through artificial intelligence image control.
According to claim 14,
The risk detection step for education is,
The dangerous object detection unit to which the dangerous object detection artificial intelligence model is applied applies the image to the dangerous object detection artificial intelligence model to detect a dangerous object that may cause a dangerous situation in the image, and determines the type and location of the detected dangerous object. Further comprising a dangerous object detection step of providing information about the collision to the collision detection unit,
In the education target risk detection step, the collision detection unit,
Industrial safety and health education method through artificial intelligence video control, characterized in that for monitoring whether a dangerous situation occurs in the object detected by the dangerous object detection unit.
According to claim 12,
The position detection unit,
Outdoors, a beacon obtains location information by calculating a location based on location data received from GPS satellites, and periodically transmits its own location information by installing at least one surveillance area indoors where location data is not received from GPS satellites. A worker's mobile terminal that obtains the location information from and transmits the location information and worker location information including worker identification information for a worker working in the surveillance area to the training subject recognition unit,
The education target recognition step,
Identification of a worker for which the worker identification information registration unit of the training target recognition unit receives and registers worker identification information including worker information for each worker performing work at the industrial site and terminal identification information of a mobile terminal used by each worker information registration step;
a danger occurrence location acquisition step of receiving and outputting surveillance area location information about a surveillance area in which a dangerous situation occurs from the danger detection unit by the location acquisition unit of the education subject recognition unit;
The operator location tracking unit of the training target recognition unit receives the location information of the operator from the location detection unit and tracks the location where the operator is working by referring to the registered operator identification information. Worker location tracking step of outputting worker identification information located in the monitoring area; and
When the education subject identification unit of the education target recognition unit transmits the monitoring area location information to the work location tracking unit when monitoring area location information is input from the danger occurrence location acquisition unit, the operator for the operator in the monitoring area of the surveillance area location information Occupational safety and health through artificial intelligence video control, comprising a step of identifying training subjects by receiving identification information, selecting training subjects, and providing training subject selection information including the selected worker identification information to the education providing unit. educational method.