KR20210065595A - System and method for predicting distribution and moving path of sources in 3d space based on machine learning - Google Patents

Abstract

The present invention relates to a system for predicting a distribution and moving path of a source (light source, electricity, magnetism, radiation) in a three-dimensional (3D) space on the basis of machine learning to simply and correctly perform prediction and a method thereof. According to the present invention, the method comprises: a step of acquiring position information and source information of a source by time while moving the source in a 3D space, and acquiring a data set of response signals by time by a plurality of detectors; a step of performing machine learning of an artificial neural network by inputting the position information and source information by time of the source to an output node and inputting the data set of the response signals by time to an input node, and repetitively selecting and overlapping learning data by time of each single source to predict a multi-source distribution; and a step of inputting the data set of response signals by time received from the plurality of detectors to the machine-learning artificial neural network to predict a moving path of the source.

Description

A system and method for predicting the distribution and movement path of a seafarer in 3D space based on machine learning {SYSTEM AND METHOD FOR PREDICTING DISTRIBUTION AND MOVING PATH OF SOURCES IN 3D SPACE BASED ON MACHINE LEARNING}

The present invention relates to a system and method for predicting the distribution and movement path of a source (light source, electricity, magnetism, radiation) in a machine learning-based three-dimensional space, and in particular, time-wise position information and source of a source while moving the source in a three-dimensional space In addition to acquiring information, a data set of response signals by time is acquired by a plurality of detectors, and location information and crew information by time of the crew are input to the output node of the artificial neural network, and the data set of response signals by time is input to the input node. After input to machine learning, the multi-source distribution is predicted by repeatedly selecting and superimposing the time-based learning data of a single source, and inputting the data set of time-dependent response signals from a plurality of detectors to the machine-learning artificial neural network. To a system for predicting the movement path of a seafarer, and a system for predicting the distribution and movement path of a seafarer in a machine learning-based three-dimensional space, and a method therefor.

In general, a source refers to radiation, a light source, electricity, magnetism, or the like. While radiation is used in many fields such as medical, industrial, educational, public, and military fields, accidents caused by radiation sources are constantly occurring at home and abroad.

In particular, in the medical field, accidents due to over-irradiation in radiation therapy account for most of the accidents, and deaths of patients and over-exposure of patients and medical staff account for the majority. The detailed causes of medical exposure include misidentification of the dose unit, miscalculation of dose, wrong irradiation with photons and electron beams mistaken, overdosing of radioactive isotopes, and forgetting the source in the body.

In the field of non-destructive radiation, accidents mainly occur due to the use or repair of industrial radiation sources, the process of industrial gamma ray photography, and poor management of radiation sources. In many cases, not only workers but also unspecified numbers of people who have nothing to do with the radiation source lead to death, body amputation, and other human casualties such as skin damage.

In Korean Patent Registration No. 1677172 (hereinafter referred to as prior art), a photoconductive layer that absorbs radiation passing through a subject to generate electric charge, an upper electrode formed on the photoconductive layer, and the upper electrode are electrically connected A radiation detector including an automatic exposure control unit for measuring a current generated in the photoconductive layer and generating a control signal for controlling an amount of radiation in response to the current is disclosed.

However, the prior art has the advantage of reducing the exposure dose of the examinee and shortening the imaging time, but there is a problem in that the location and distribution of the radiation source cannot be grasped.

Domestic Patent Registration No. 1677172 (Title of Invention: Radiation detector and radiographic imaging device including same, automatic exposure control method and radiographic imaging method)

Therefore, the present invention has been made to solve the above problems, and an object of the present invention is to make it possible to relatively simply and accurately predict the distribution of multiple sources and predict the movement path of a source, a machine learning-based three-dimensional space image. An object of the present invention is to provide a system and method for predicting the distribution and movement path of a seafarer.

In order to achieve the above object, the system for predicting the distribution and movement path of a source in 3D space based on machine learning according to an embodiment of the present invention is equipped with a source, and moving the source in 3D space by time of the source a mobile seafarer device configured to acquire location information and seafarer information; a plurality of detectors installed in a three-dimensional space and configured to obtain a data set of the response signals for each time of the source; a data collection unit configured to collect time-based location information and source information of a source from the mobile source device, and also collect a data set of time-based response signals of the source from the plurality of detectors; a machine learning unit configured to input the collected time-based location information and source information to an output node of the artificial neural network, and input the collected data set of time-dependent response signals of the seafarer to the input node for machine learning; a multi-source distribution prediction unit configured to predict a distribution of multiple sources by repeatedly performing overlapping time-wise learning data of a set number of single sources from the machine learning unit; and a sailor movement path prediction unit configured to predict a movement path for a sailor by inputting a data set of time-dependent response signals from the detector to the machine-learning artificial neural network of the machine learning unit.

In the machine learning-based three-dimensional spatial source distribution and movement path prediction system according to the embodiment, the source may include a light source, electricity, magnetism, and radiation.

In the machine learning-based three-dimensional spatial source distribution and movement path prediction system according to the embodiment, the source information may include energy, intensity, and particle type.

In order to achieve the above object, a method for predicting the distribution and movement path of a source in 3D space based on machine learning according to another embodiment of the present invention is a method of predicting the distribution and movement path of a source in 3D space while the mobile source device moves in 3D space, position information and source information for each time of the source obtaining, by a plurality of detectors, obtaining a data set of time-dependent response signals of the source; collecting, by a data collection unit, location information and source information of a source by time from the mobile source device, and collecting a data set of response signals of the source by time from the detector; performing machine learning by inputting, by a machine learning unit, the location information and the source information of the seafarer collected by time to an output node of an artificial neural network, and inputting the collected data set of the response signals by time of the seafarer to the input node; predicting the distribution of multiple sources by repeatedly performing, by a multi-source distribution predicting unit, selecting and overlapping time-wise learning data of a set number of single sources from the machine learning unit; and predicting, by the crew member movement path prediction unit, the movement path of the crew member by inputting a data set of time-dependent response signals from the detector to the machine-learned artificial neural network of the machine learning unit.

According to the machine learning-based three-dimensional space source distribution and movement path prediction system and method according to the embodiments of the present invention, the mobile seafarer device acquires the position information and the seafarer information by time while moving in the three-dimensional space In addition, a plurality of detectors acquire a data set of time response signals of the source; a data collection unit collects location information and source information of a crew member by time from the mobile source device, and collects a data set of response signals of the crew member by time from the detector; the machine learning unit inputs the collected time-based location information and source information of the seafarer to an output node of the artificial neural network, and inputs the collected data set of the time-wise response signals of the seafarer to the input node to perform machine learning; predicting the distribution of multiple sources by repeatedly performing, by a multi-source distribution predicting unit, selecting and overlapping time-wise learning data of a set number of single sources from the machine learning unit; The crew movement path prediction unit is configured to predict the movement path of the source by inputting a data set of response signals by time from the detector to the machine-learned artificial neural network of the machine learning unit, thereby predicting the distribution of multiple sources and the movement path of the source. It has an excellent effect of making predictions relatively simple and accurate.

1 is a block diagram of a machine learning-based three-dimensional space source distribution and movement path prediction system according to an embodiment of the present invention.
FIG. 2 is a diagram illustrating a two-dimensional example (a) and a three-dimensional example (b) of a data collection process for each detector as the mobile source device of FIG. 1 moves.
FIG. 3 is a diagram illustrating a data set of time-based location information and source information of a seafarer collected by the data collection unit in FIG. 1, and time-dependent response signals of a seafarer.
4 is a diagram illustrating an artificial neural network used in the machine learning unit of FIG. 1 .
FIG. 5 is a diagram illustrating a process of selecting and superimposing learning data for each time of a single source in the multi-source distribution prediction unit of FIG. 1 .
FIG. 6 is a diagram illustrating a process of predicting a movement path for a seafarer in the sailor movement path prediction unit of FIG. 1 .
7 is a flowchart illustrating a method for predicting a distribution and movement path of a source in a machine learning-based three-dimensional space according to an embodiment of the present invention.

In describing an embodiment of the present invention, if it is determined that a detailed description of a known technology related to the present invention may unnecessarily obscure the gist of the present invention, the detailed description thereof will be omitted. In addition, the terms described below are terms defined in consideration of functions in the present invention, which may vary according to intentions or customs of users and operators. Therefore, the definition should be made based on the content throughout this specification. The terminology used in the detailed description is for the purpose of describing embodiments of the present invention only, and should in no way be construed as limiting. Unless explicitly used otherwise, expressions in the singular include the meaning of the plural. In this description, expressions such as “comprising” or “comprising” are intended to indicate certain features, numbers, steps, acts, elements, some or a combination thereof, and one or more other than those described. It should not be construed as excluding the presence or possibility of other features, numbers, steps, acts, elements, or any part or combination thereof.

In each system shown in the figures, elements in some instances may each have the same reference number or a different reference number to suggest that the represented element may be different or similar. However, elements may have different implementations and work with some or all of the systems shown or described herein. The various elements shown in the drawings may be the same or different. Which one is referred to as the first element and which is referred to as the second element is arbitrary.

In this specification, when any one component 'transmits', 'transfers' or 'provides' data or signal to another component, one component directly transmits data or signal to another component, as well as, and transmitting data or signals to the other component via at least one further component.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings.

1 is a block diagram of a machine learning-based three-dimensional space source distribution and movement path prediction system according to an embodiment of the present invention.

As shown in FIG. 1, the machine learning-based three-dimensional spatial source distribution and movement path prediction system according to an embodiment of the present invention includes a mobile source device 100, a plurality of detectors 110, and a data collection unit ( 200 ), a machine learning unit 300 , a multi-source distribution prediction unit 400 , and a source movement path prediction unit 500 .

The mobile source device 100 is equipped with a source including a light source, electricity, magnetism, and radiation, and serves to obtain positional information and source information of the source by time while moving the source in a three-dimensional space. 2 shows a path along which a source moves by the mobile source device 100 .

A plurality of detectors 110 are installed in a three-dimensional space, and serve to acquire a data set of time-dependent response signals of a source. Fig. 2 (a) shows a two-dimensional example of the installation position of the detector 110 and the data collection process for each detector 110, and Fig. 2 (b) shows the 3 data collection process for each detector 110 A dimensional example is shown.

The data collection unit 200 collects the position information and the source information by time of the crew from the mobile seafarer device 100 and collects a data set of the response signals of the crew by time from the plurality of detectors 110 to collect the machine learning unit 300 ) serves to provide

,

,…

)가 입력된 함수이고, R은 시간(t)에서의 디텍터(

)(110)에서 반응한 신호들의 데이터 집합이다. 선원의 정보(S)에는 에너지, 강도, 입자의 종류 등의 여러 정보가 포함될 수 있다.FIG. 3 shows a data set of positional information and source information of a source by time, which the data collection unit 200 collects, and response signals of a source by time. Here, f is the source information (S) and spatial position information (

,

,…

) is the input function, and R is the detector (

) is a data set of signals reacted in 110 . The source information (S) may include various types of information such as energy, strength, and particle types.

The machine learning unit 300 inputs the hourly position information and the source information of the seafarer collected from the data collection unit 200 to the output node of the artificial neural network, and inputs the collected data set of the hourly response signals of the seafarer to the input node. It works for machine learning.

,

,…

) 및 선원의 정보(S)가 입력되고, 인공신경망의 입력노드에는 시간(t)에서의 디텍터(110)에서 반응한 신호들의 데이터 집합 R(

)이 입력된다. 4 is a diagram illustrating an artificial neural network used in the machine learning unit 300 . Here, the output node of the artificial neural network contains spatial location information (

,

,…

) and source information (S) are input, and the data set R (

) is entered.

The multi-source distribution prediction unit 400 serves to predict the distribution of multiple sources by repeatedly performing overlapping by selecting and overlapping time-based learning data of a set number of single sources from the machine learning unit 300 .

FIG. 5 is a diagram illustrating a process in which the multi-source distribution prediction unit 400 selects and overlaps learning data for each time of a single source. When the time (t) is 1, 300, 500 (see Fig. 5 a), select the training data [see Fig. 5 b, here three training data (i.e., blue, black, and red curves)], When overlapped, the curve of FIG. 5(c) is obtained. In the case of repeating such overlapping of the learning data, the position and strength of the source can be predicted through the learning result when a plurality of sources exist without direct experimentation.

The crew movement path prediction unit 500 inputs a data set of response signals for each time from the detector 110 to the machine-learned artificial neural network of the machine learning unit 300 to predict the movement path of the crew member.

FIG. 6 is a diagram illustrating a process of predicting a movement path for a seafarer in the sailor movement path prediction unit of FIG. 1 . 6 is a two-dimensional diagram (a of FIG. 6) and a three-dimensional diagram showing the position and movement path of a sailor at a future time (t = 5) predicted from movement information of time (t = 1, 2, 3, 4) Figure (b) is shown.

A method for predicting a source distribution and movement path using a machine learning-based three-dimensional space source distribution and movement path prediction system according to an embodiment of the present invention configured as described above will be described with reference to the drawings.

7 is a flowchart for explaining a method for predicting the distribution and movement path of a source in 3D space based on machine learning according to an embodiment of the present invention, where S means a step.

First, while the mobile seafarer apparatus 100 moves in a three-dimensional space (S10), it acquires positional information and source information of a seafarer by time (S20).

At this time, the plurality of detectors 110 acquires a data set of the response signals for each time of the source (S30).

Subsequently, the data collection unit 200 collects the position information and the source information by time of the seafarer obtained in step S20 from the mobile seafarer device 100 and the data of the response signals of the seafarer obtained in step S30 by time A set is collected from the detector 110 (S40).

Next, the machine learning unit 300 inputs the position information and the source information by time of the seafarer collected in step S40 to the output node of the artificial neural network, and inputs the data set of the response signals by time of the collected seafarer to the input node. machine learning (S50).

Then, the multi-source distribution prediction unit 400 selects and overlaps the time-based learning data of a set number of single sources from the machine learning unit 300 repeatedly to predict the distribution of the multi-source sources (S60).

Thereafter, the movement path prediction unit 500 for the sailor predicts the movement path for the sailor by inputting a data set of response signals for each time from the detector 110 to the machine-learned artificial neural network of the machine learning unit 300 (S70). .

According to the system and method for predicting the distribution and movement path of a seafarer in three-dimensional space based on machine learning according to an embodiment of the present invention, the mobile seafarer device acquires the position information and the seafarer information by time of the seafarer while moving in the three-dimensional space, In addition, a plurality of detectors acquire a data set of the time-dependent response signals of the source; a data collection unit collects location information and source information of a crew member by time from the mobile source device, and collects a data set of response signals of the crew member by time from the detector; the machine learning unit inputs the collected time-based location information and source information of the seafarer to an output node of the artificial neural network, and inputs the collected data set of the time-wise response signals of the seafarer to the input node to perform machine learning; predicting the distribution of multiple sources by repeatedly performing, by a multi-source distribution predicting unit, selecting and overlapping time-wise learning data of a set number of single sources from the machine learning unit; The crew movement path prediction unit is configured to predict the movement path of the source by inputting a data set of response signals by time from the detector to the machine-learned artificial neural network of the machine learning unit, thereby predicting the distribution of multiple sources and the movement path of the source. It has an excellent effect of making predictions relatively simple and accurate.

In the drawings and specification, the best embodiment is disclosed, and specific terms are used, but these are used only for the purpose of describing the embodiments of the present invention, and are used to limit the meaning or limit the scope of the present invention described in the claims it didn't happen Therefore, it will be understood by those skilled in the art that various modifications and equivalent other embodiments are possible therefrom. Therefore, the true technical protection scope of the present invention should be determined by the technical spirit of the appended claims.

100: mobile sailor device
110: detector
200: data collection unit
300: machine learning unit
400: multi-source distribution prediction unit
500: Sailor movement path prediction unit

Claims (6)

a mobile crew member equipped with a crew member and configured to acquire positional information and crew member information by time of the crew member while moving the crew member in a three-dimensional space;
a plurality of detectors installed in a three-dimensional space and configured to acquire a data set of response signals of the source by time;
a data collection unit configured to collect time-based location information and source information of a source from the mobile source device and collect a data set of time-based response signals of the source from the plurality of detectors;
a machine learning unit configured to input the collected time-based location information and source information to an output node of an artificial neural network, and input the collected data set of time-dependent response signals of the seafarer to the input node for machine learning;
a multi-source distribution prediction unit configured to predict a distribution of multiple sources by repeatedly performing overlapping time-wise learning data of a set number of single sources from the machine learning unit; and
Sailor movement path prediction unit configured to predict the movement path of the sailor by inputting a data set of response signals by time from the detector to the machine-learned artificial neural network of the machine learning unit; Distribution and movement path prediction system.
The method of claim 1,
The source is a machine learning-based three-dimensional space source distribution and movement path prediction system including a light source, electricity, magnetism, and radiation.
The method of claim 1,
The source information is a machine learning-based three-dimensional space source distribution and movement path prediction system including energy, intensity, and particle types.
A method for predicting the distribution and movement of a source using the system for predicting the distribution and movement path of a source in a three-dimensional space based on machine learning according to claim 1,
obtaining, by the mobile source device, positional information and source information by time of a source while moving in a three-dimensional space, and obtaining, by a plurality of detectors, a data set of response signals by time of the source;
collecting, by a data collection unit, time-based location information and source information of a source from the mobile source device, and collecting a data set of time-based response signals of the source from the detector;
performing machine learning by inputting, by a machine learning unit, the location information and source information of the seafarer collected by time to an output node of an artificial neural network, and inputting the collected data set of the response signals by time of the seafarer to the input node;
predicting the distribution of multiple sources by repeatedly performing, by a multi-source distribution predicting unit, selecting and overlapping time-wise learning data of a set number of single sources from the machine learning unit; and
Predicting the movement path of the sailor by inputting the data set of the response signals for each time from the detector to the machine-learned artificial neural network of the machine learning unit by the sailor movement path prediction unit; Distribution and movement path prediction methods.
5. The method of claim 4,
The source is a method of predicting distribution and movement path of a source in a machine learning-based three-dimensional space including a light source, electricity, magnetism, and radiation.
5. The method of claim 4,
The source information is a machine learning-based three-dimensional space source distribution and movement path prediction method including energy, intensity, and particle types.

Images