WO2024106782A1

WO2024106782A1 - Safety accident prevention system for apparatus using virtual fences and skeleton images

Info

Publication number: WO2024106782A1
Application number: PCT/KR2023/016433
Authority: WO
Inventors: 이영규
Original assignee: 주식회사 아이티공간
Priority date: 2022-11-14
Filing date: 2023-10-21
Publication date: 2024-05-23
Also published as: KR20240070752A

Abstract

The present invention relates to a safety accident prevention system for an apparatus using virtual fences and skeleton images, the system being characterized by comprising: a virtual area module which constructs and sets a virtual boundary area around the apparatus and detects whether an object enters the boundary area; a skeleton module which captures images around the apparatus and forms the object around the apparatus into a skeleton model on the basis of the image information and collects same; a logic gate which detects and determines the danger level around the apparatus on the basis of information about the virtual area module and skeleton module and transmits the determination information to a control unit; and the control unit which controls the operation of the apparatus on the basis of the determination information of the logic gate.

Description

Device safety accident prevention system using virtual fence and skeleton image

The present invention relates to a safety accident prevention system for devices using virtual fences and skeleton images. More specifically, the logic gate determines the level of risk around the device by combining information measured and collected from the virtual area module and skeleton module, respectively. By clearly judging and detecting, not only can workers be effectively protected by preventing safety accidents in advance, but they can also minimize economic losses that may occur due to unnecessary equipment shutdown, leading to stable operation and management of factory equipment. This is about a safety accident prevention system for devices using virtual fences and skeleton images.

Generally, devices that perform a set process are placed along a process line in a factory, and some or all of these devices perform the process together with the operator.

In this way, when workers perform processes with equipment, safety accidents often occur where workers become trapped (stuck) in the equipment. There are no separate means to prevent such safety accidents, resulting in unfortunate cases where workers are seriously injured or die. There was a problem that occurred.

In addition, even if workers are exposed to the risk of safety accidents due to various noises generated from equipment inside the factory, it is not easy for other workers to raise awareness of the risk of accidents, which causes safety accidents to occur more frequently.

The present invention was proposed to solve the problems described above, and its purpose is to clearly determine and detect the level of risk around the device by combining the information measured and collected from the virtual area module and the skeleton module, respectively. By doing so, it not only effectively protects workers by preventing safety accidents, but also minimizes economic losses that may occur due to unnecessary operation of equipment, leading to stable operation and management of factory equipment. The goal is to provide a safety accident prevention system for devices using fences and skeleton images.

In addition, by dividing the virtual boundary area of the virtual area module into a warning boundary area and a danger boundary area, if there is a possibility that a worker may be exposed to a safety accident, an alert is immediately sent through the alarm unit to encourage the worker to escape the risk of an accident on his or her own. We provide a device safety accident prevention system using virtual fences and skeleton images that can more easily protect the safety of workers.

The device safety accident prevention system using a virtual fence and skeleton image according to the present invention includes a virtual area module that builds and sets a virtual boundary area around the device and detects whether an object enters the boundary area, and an image around the device. A skeleton module that captures images and collects objects around the device by shaping them into a skeleton model based on the image information, detects and determines the level of risk around the device based on the information of the virtual area module and the skeleton module, and makes that determination. It is characterized by comprising a logic gate that transmits information to a control unit, and a control unit that controls the operation of the device based on the judgment information of the logic gate.

In addition, the logic gate detects the entry of an object into the virtual boundary area around the device built in the virtual area module, and at the same time, based on the skeleton model of the object that entered the virtual boundary area through the skeleton module, the object is converted into a human. When detected, the area around the device is judged to be in a dangerous state and an alert is sent to the control unit.

The logic gate detects the entry of an object into the virtual boundary area around the device built in the virtual area module, and detects that the object is not a person based on the skeleton model of the object that entered the virtual boundary area through the skeleton module. When this happens, the surrounding area of the device is determined to be in a safe state and transmitted to the control unit.

The control unit is characterized in that it controls the operation of the device to stop when the logic gate determines that the surroundings of the device are in a dangerous state.

In addition, it further includes a condition setting module that can set body parts that are allowed to enter the virtual boundary area constructed in the virtual area module,

Even if a body part set through the condition setting module enters the virtual boundary area, the logic gate determines that the surrounding situation of the device is in a safe state.

In addition, the skeleton module includes an overall image processor that collects the entire image by shooting the entire area around the virtual boundary area built around the device, and a point image collection system that collects point images by shooting around the inside of the virtual boundary area built around the device. It is characterized by comprising a point image processor and a skeleton AI that shapes objects included in the image into a skeleton model based on the overall and point images captured by the overall and point image processor.

In addition, the virtual border area set in the virtual area module is divided into a risk border area built around the device and an alarm border area built outside the risk border area,

It further includes an alarm unit that warns of danger to the surroundings,

When the logic gate detects the entry of an object into the alarm boundary area of the virtual area module and detects that the object entering the alarm boundary area through the skeleton module is a person, it alerts the surrounding area through the alarm unit,

When an object is detected entering the danger boundary area of the virtual area module and the object entering the danger boundary area is detected as a person through the skeleton module, the control unit determines the surrounding area of the device to be in a dangerous state and alerts the device to stop operation. It is characterized by controlling it as much as possible.

According to the device safety accident prevention system using a virtual fence and skeleton image according to the present invention, the logic gate clearly determines and detects the degree of risk around the device by combining the information measured and collected from the virtual area module and the skeleton module, respectively. By doing so, it not only effectively protects workers by preventing safety accidents, but also minimizes economic losses that may occur when equipment stops operating unnecessarily, leading to stable operation and management of factory equipment. There is.

In addition, by dividing the virtual boundary area of the virtual area module into a warning boundary area and a danger boundary area, if there is a possibility that a worker may be exposed to a safety accident, an alert is immediately sent through the alarm unit to encourage the worker to escape the risk of an accident on his or her own. It has the effect of more easily protecting the safety of workers.

Figure 1 is a conceptual diagram of a safety accident prevention system for devices using a virtual fence and skeleton image of the present invention.

Figure 2 is a diagram illustrating a safety accident prevention system for devices using the virtual fence and skeleton image shown in Figure 1

Figures 3 and 4 are diagrams for explaining the diagram of the device safety accident prevention system using the virtual fence and skeleton image shown in Figure 2

Figure 5 is a diagram of the condition setting module in the device safety accident prevention system using the virtual fence and skeleton image shown in Figure 1

Figures 6 and 7 are diagrams for explaining the condition setting module in the device safety accident prevention system using the virtual fence and skeleton image shown in Figure 5.

Figure 8 is a diagram for explaining an example of an alarm unit through a device safety accident prevention system using the virtual fence and skeleton image shown in Figure 1

〈Explanation of symbols for main parts of the drawing〉

10. Virtual area module 11. Virtual boundary area

20. Skeleton module 21. Full video processor

22. Point image processor 23. Skeleton AI

30. Logic gate 40. Control unit

50. Condition setting module 60. Alarm unit

100. Device safety accident prevention system using virtual fence and skeleton image

The present invention relates to a safety accident prevention system for devices using virtual fences and skeleton images, which consists of a virtual area module that builds and sets a virtual boundary area around the device and detects whether an object enters the boundary area; , a skeleton module that captures images around the device, shapes and collects objects around the device into a skeleton model based on the image information, and detects the degree of risk around the device based on the information of the virtual area module and the skeleton module. It is characterized by consisting of a logic gate that makes a decision and transmits the decision information to a control unit, and a control unit that controls the operation of the device based on the decision information of the logic gate.

A safety accident prevention system for devices using a virtual fence and a skeleton image according to a preferred embodiment of the present invention will be described in detail based on the attached drawings. Detailed descriptions of well-known functions and configurations that are judged to unnecessarily obscure the gist of the present invention are omitted.

1 to 8 show a safety accident prevention system for devices using a virtual fence and a skeleton image according to an embodiment of the present invention. Figure 1 shows a safety accident prevention system for devices using a virtual fence and a skeleton image of the present invention. , Figure 2 is a diagram to explain the safety accident prevention system of the device using the virtual fence and skeleton image shown in Figure 1, and Figures 3 and 4 are diagrams showing the device using the virtual fence and skeleton image shown in Figure 2. A diagram to explain the diagram of the safety accident prevention system, Figure 5 is a diagram of the condition setting module in the device safety accident prevention system using the virtual fence and skeleton image shown in Figure 1, and Figures 6 and 7 are diagrams. Figure 5 is a diagram to explain the condition setting module in the device safety accident prevention system using the virtual fence and skeleton image shown in Figure 5, and Figure 8 is a diagram showing a device safety accident using the virtual fence and skeleton image shown in Figure 1. Each drawing is shown to explain an example of an alarm unit through a prevention system.

As shown in Figure 1, the device safety accident prevention system 100 using a virtual fence and a skeleton image according to an embodiment of the present invention includes a virtual area module 10, a skeleton module 20, and a logic gate ( 30) and a control unit 40.

The virtual area module 10 establishes a virtual boundary area 11 around the device and detects whether an object enters the boundary area.

Here, the virtual boundary area 11 built around the device can be constructed in various sizes or shapes in consideration of the surrounding conditions or conditions in which the device is used, and this virtual boundary area 11 is a typical 3D area. It will be built based on the LiDAR (Light Detection And Ranging) method, but of course, it is not limited to this method.

That is, the virtual image module 10 constructs a virtual boundary area 11 around the device and detects whether an object enters the constructed boundary area 11. The detection information detected in this way is based on the logic. It serves as the basis for determining the level of risk around the device at the gate 30.

The skeleton module 20 captures images around the device, shapes objects around the device into a skeleton model based on the image information, and collects them.

An overall image processor 21 that collects the entire image by shooting the entire area around the virtual boundary area 11 built around the device, and a point image by shooting the inside area around the virtual boundary area 11 built around the device. It includes a point image processor 22 that collects and a skeleton AI 23 that shapes objects included in the image into a skeleton model based on the overall and point images captured by the overall and

point image processors

21 and 22. This is done.

Here, of course, the overall and

point image processors

21 and 22 can be implemented with any device that can capture images, such as CCTV, IP cameras, infrared cameras, etc.

That is, the overall and

point image processors

21 and 22 capture images around the virtual boundary area 11, and the skeleton AI 23 converts the captured image information into a plurality of consecutive images over time. It is composed of an image frame, and a skeleton model is formed based on key points corresponding to the joint parts of the object on the composed image frame.

Here, the skeleton AI 23 uses the image of the point image processor 22 along with the image of the entire image processor 21 to more precisely identify the part of the object entering the inside of the virtual boundary area 11. For example, in the case of a human object shaped as a skeleton model, the body part that has entered the inside of the virtual boundary area can be detected very precisely (accurately).

Therefore, the logic gate 30 can detect the degree of risk around a highly reliable device based on the skeleton model embodied in the skeleton AI 23, which will be described in detail below.

As shown in FIG. 2, the logic gate 30 detects and determines the degree of risk around the device based on the information of the virtual area module 10 and the skeleton module 20, and transfers the judgment information to the above information, which will be described later. It is transmitted to the control unit 40 to control the device so that the safety of workers is protected.

Here, in order to explain the process of detecting and determining the level of danger around the device through the logic gate 30, the device will be described by implementing it as a kneading device that kneads the ingredients by rotating the stirring blade based on the power source of the motor. Of course, it is not limited to these devices.

First, as shown in FIG. 3, the logic gate 30 detects the entry of an object into the virtual boundary area 11 around the kneading machine built in the virtual area module 10, and at the same time, the skeleton module 20 ), if the object is detected as a person based on the skeleton model of the object that entered the virtual boundary area 11 through ), the area around the kneading machine is judged to be in a dangerous state and an alert is sent to the control unit 40 so that the kneading machine can be stopped immediately. do.

Accordingly, the safety of workers can be effectively protected by preventing safety accidents in which workers entering the virtual boundary area 11 constructed by the virtual area module 10 may become trapped in the kneading machine.

In addition, as shown in FIG. 4, the logic gate detects the entry of an object into the virtual boundary area 11 around the kneading machine built in the virtual area module 10, and detects the virtual boundary area 11 through the skeleton module 20. If the object is detected to be non-human based on the skeleton model of the object entering the boundary area 11, the control unit 40 determines the surrounding area of the device to be in a safe state and transmits the judgment to maintain the operation of the kneading device.

The judgment through the logic gate 30 as described above is made by combining the information collected from the virtual area module 10 and the skeleton module 20, so excellent reliability of the judgment information can be expected, thereby ensuring worker safety. Accidents can be effectively prevented while economic losses that may occur due to unnecessary equipment shutdown can be minimized.

The control unit 40 controls the operation of the device as described above based on the judgment information of the logic gate 30 to safely protect workers.

Meanwhile, as shown in FIG. 1, it further includes a condition setting module 50 that can set body parts that are allowed to enter the virtual boundary area 11 constructed in the virtual area module 10,

Even if the body part set through the condition setting module 50 enters the virtual boundary area 11, the logic gate 30 determines the surrounding situation of the device to be in a safe state.

Depending on the type of device, there may be cases where the worker's body approaches the device and repeatedly enters the virtual boundary area 11. For example, the worker directly inserts materials into the device or retrieves materials from the device. In this case, the worker's body has no choice but to naturally enter the virtual boundary area (11) built around the device.

Therefore, as shown in FIGS. 5 and 6, if the body parts permitted to enter the virtual boundary area 11 are set up to the elbow area including the hand through the condition setting module 50, the logic gate 30 ) detects the entry of a person into the virtual boundary area 11 through the virtual area module 10 and the skeleton module 20, but only enters the hand and elbow areas of the body set in the condition setting module 50. In this case, the surrounding area of the device is judged to be in a safe state.

As shown in FIG. 7, if the forearm or shoulder area beyond the elbow of the worker's body enters the virtual boundary area 11, the logic gate 30 determines the area around the device to be in a dangerous state and the control unit 40 ) of course, ensures the safety of workers by quickly stopping the device by sending an alarm.

Here, the skeleton module 20 includes the entire image processor 21, which images the entire surroundings of the virtual boundary area 11, and the point image processor 22, which intensively images the inside of the virtual boundary area 11. ) By shaping the skeleton model based on the image information captured in Therefore, excellent reliability of the judgment of the logic gate 30, which determines the dangerous state of the device based on the detection information of the skeleton module 20, can be secured.

Meanwhile, as shown in FIG. 1, the virtual border area 11 set in the virtual area module 10 includes a risk border area 11a built around the device and outside of the risk border area 11a. It is divided into a constructed alarm boundary area (11b),

It further includes an alarm unit 60 that warns of danger to the surroundings,

The logic gate 30 detects the entry of an object into the alarm boundary area 11b of the virtual area module 10, and detects the object entering the alarm boundary area 11b through the skeleton module 20 as a person. When detected, an alert is sent to the surroundings through the alarm unit 60,

When an object is detected entering the danger boundary area 11a of the virtual area module 10 and the object entering the danger boundary area 11a through the skeleton module 20 is detected as a person, the control unit 40 The area around the device is judged to be in a dangerous state and an alarm is issued to control the device to stop operation.

That is, as shown in FIG. 8, the logic gate 30 is connected to the alarm boundary area 11b, which is built around the device based on the detection information of the virtual area module 10 and the skeleton module 20. When the entry of is detected, an alert is sent to the surroundings through the alarm unit 60 to guide the person entering the alarm boundary area 11b to escape from danger by leaving the vicinity of the device, and at this time, the device is allowed to operate normally.

Here, the alarm unit 60 is preferably implemented as a normal alarm light that emits a light source to the surroundings, so that the alarm can be effectively delivered to workers even in sites where various noises are generated.

Then, when a person's entry into the danger boundary area 11a built around the device is detected, the logic gate 30 alerts the control unit 40 of a dangerous state and immediately stops the operation of the device and moves it to the danger boundary area. Prevent safety accidents for those entering.

Here, by dividing the virtual boundary area 11 into an alarm boundary area 11b and a danger boundary area 11a, the level of the alarm determined through the logic gate 30 can be set more diversely to ensure worker safety. It protects more effectively and at the same time easily blocks workers from entering the dangerous boundary area 11a to ensure smooth operation of the device.

The logic gate 30 of the device safety accident prevention system 100 using the virtual fence and skeleton image of the present invention configured as described above is measured and collected from the virtual area module 10 and the skeleton module 20, respectively. By combining a variety of information to clearly determine and detect the level of risk around the device, not only can workers be effectively protected by preventing safety accidents in advance, but it can also provide economical savings that can occur when the device stops operating unnecessarily. It has the effect of minimizing losses and leading to stable operation and management of factory equipment.

In addition, by constructing the virtual boundary area 11 of the virtual area module 10 by dividing it into an alarm boundary area 11b and a danger boundary area 11a, if there is a possibility that a worker may be exposed to a safety accident, the alarm unit ( 60), it has the effect of making it easier to protect the safety of workers by guiding them to escape the risk of accidents by providing an alarm.

The present invention has been described with reference to the embodiments shown in the accompanying drawings, but these are illustrative and are not limited to the above-described embodiments, and various modifications and equivalent embodiments can be made by those skilled in the art. You will understand the point. In addition, it goes without saying that modifications can be made by those skilled in the art without impairing the spirit of the present invention. Accordingly, the scope claimed in the present invention will not be limited by the scope of the detailed description, but will be limited by the claims and technical ideas described later.

Claims

In a system that prevents safety accidents that may occur in devices,

A virtual area module 10 that constructs and sets a virtual boundary area 11 around the device and detects whether an object enters the boundary area.

A skeleton module 20 that captures images around the device and collects objects around the device by forming them into a skeleton model based on the image information;

A logic gate (30) that detects and determines the degree of risk around the device based on information from the virtual area module (10) and the skeleton module (20) and transmits the determination information to the control unit (40); and

A safety accident prevention system for devices using a virtual fence and skeleton image, comprising: a control unit 40 that controls the operation of the device based on the judgment information of the logic gate 30.
According to claim 1,

The logic gate 30 detects the entry of an object into the virtual boundary area 11 around the device built in the virtual area module 10, and simultaneously enters the virtual boundary area 11 through the skeleton module 20. If the object is detected as a person based on the skeleton model of the entered object, the surrounding area of the device is judged to be in a dangerous state and an alert is sent to the control unit 40,

The logic gate 30 detects the entry of an object into the virtual boundary area 11 around the device constructed in the virtual area module 10, and enters the virtual boundary area 11 through the skeleton module 20. If the object is detected to be non-human based on the skeleton model of the object, the surrounding area of the device is determined to be in a safe state and transmitted to the control unit 40.

The control unit 40 is a device safety accident prevention system using a virtual fence and skeleton image, characterized in that the logic gate 30 controls the operation of the device to stop when the device's surroundings are judged to be in a dangerous state.
According to claim 2,

It further includes a condition setting module (50) that can set body parts that are allowed to enter the virtual boundary area (11) constructed in the virtual area module (10),

A virtual fence and skeleton image, characterized in that even if a body part set through the condition setting module 50 enters the inside of the virtual boundary area 11, the logic gate 30 determines the surrounding situation of the device to be in a safe state. Safety accident prevention system for devices using .
According to claim 3,

The skeleton module 20 is,

an overall image processor (21) that collects the entire image by capturing the entire area around the virtual boundary area (11) built around the device;

A point image processor 22 that collects point images by photographing the inside of a virtual boundary area 11 built around the device;

A virtual machine comprising a skeleton AI (Artificial Intelligence) (23) that shapes objects included in the image into a skeleton model based on the overall and point images captured by the overall and point image processors (21, 22). Safety accident prevention system for devices using fence and skeleton images.
According to claim 2,

The virtual border area 11 set in the virtual area module 10 is divided into a risk border area 11a built around the device and a warning border area 11b built outside the risk border area 11a. Separated,

It further includes an alarm unit 60 that warns of danger to the surroundings,

The logic gate 30 detects the entry of an object into the alarm boundary area 11b of the virtual area module 10, and detects the object entering the alarm boundary area 11b through the skeleton module 20 as a person. When detected, an alert is sent to the surroundings through the alarm unit 60,

When an object is detected entering the danger boundary area 11a of the virtual area module 10 and the object entering the danger boundary area 11a through the skeleton module 20 is detected as a person, the control unit 40 A safety accident prevention system for devices using a virtual fence and skeleton image, which determines the surrounding area around the device to be in a dangerous state and controls the operation of the device to stop.