KR102483080B1 - Method of classifying and extracting aircraft noise using artificial intelligence - Google Patents

Abstract

Disclosed is a method of classifying and extracting aircraft noise by using artificial intelligence (AI). A method of classifying and extracting aircraft noise by using AI may comprise: a step of obtaining background noise data generated by a plurality of events, via one or more geographically distributed noise sensors; a step of obtaining first noise data by performing a preprocessing operation on the background noise data, via one or more processors; a step of obtaining second noise data generated by aircraft noise events, by inputting the first noise data to a first AI model trained to identify the aircraft noise events among the plurality of events, via the one or more processors; a step of obtaining third noise data generated by aircraft noise events in a specific field, by inputting the second noise data to a second AI model trained to identify the aircraft noise events in the specific field among the aircraft noise events, via the one or more processors; and a step of extracting and classifying noise data by aircraft type in a specific field, by inputting the third noise data to a third AI model trained to extract the noise data for each aircraft type in the specific field, via the one or more processors. Therefore, based on AI, noise data for each aircraft type in a specific field can be efficiently extracted and classified from noise generated by various events.

Description

Aircraft noise classification and extraction method using artificial intelligence {METHOD OF CLASSIFYING AND EXTRACTING AIRCRAFT NOISE USING ARTIFICIAL INTELLIGENCE}

The present disclosure relates to a method for classifying and extracting noise, and more particularly, to a method for classifying and extracting aircraft noise using artificial intelligence.

The World Health Organization (WHO) defines noise as a pollutant like polluted air or dust, and designates noise as one of the major factors affecting the quality of life.

In addition, various academic circles have revealed the negative effects of noise on the human body, and as people's perception of noise changes, civil complaints related to noise are increasing exponentially. In addition, among civil complaints related to the environment, civil complaints related to noise/vibration account for more than half, and among them, civil complaints related to aircraft noise are steadily increasing.

Accordingly, national interest in the field of aircraft noise has increased, such as a complete survey on aircraft noise in Korea. In order to evaluate the noise generated at the airport, a database related to noise must be established. However, considerable manpower and cost are required to build and process a noise database that is becoming vast.

Japanese Registered Patent Publication No. JP 5863165, 2016.01.08

The present disclosure has been made to solve the above-described problems, and an object of the present disclosure is to provide a method for classifying and extracting aircraft noise using artificial intelligence.

The problems to be solved by the present disclosure are not limited to the problems mentioned above, and other problems not mentioned will be clearly understood by those skilled in the art from the description below.

According to an embodiment of the present disclosure, a method for classifying and extracting aircraft noise using artificial intelligence includes obtaining background noise data generated by a plurality of events through one or more geographically distributed noise sensors. step; obtaining first noise data by performing a pre-processing operation on the background noise data through one or more processors; Through the one or more processors, the first noise data is input to a first artificial intelligence (AI) model learned to identify an aircraft noise event among the plurality of events, so that a second noise event generated by the aircraft noise event is input. 2 acquiring noise data; Through the one or more processors, the second noise data is input to a second AI model learned to identify aircraft noise events in a specific field among the aircraft noise events, and the third noise event generated by the aircraft noise event in the specific field is input. acquiring noise data; And extracting and classifying noise data for each aircraft type in a specific field by inputting the third noise data into a third AI model learned to extract noise data for each aircraft type in a specific field through the one or more processors. can include

In addition, in the method of classifying and extracting aircraft noise using artificial intelligence, through the one or more processors, input noise data for each type of aircraft in a specific field to the fourth AI model learned to extract operation-related information. The method may further include extracting operation-related information for each type of aircraft in the specific field.

Further, the operation-related information for each type of aircraft in the specific field may include at least one of information related to the operation time of each type of aircraft in the specific field, the level of noise generated according to operation, or the frequency.

The obtaining of the first noise data may include obtaining the first noise data by performing a preprocessing operation of dividing the background noise data into predetermined length units.

In addition, the first AI model learns background noise based on labeling information matched to frequency data corresponding to noise according to occurrence of a plurality of aircraft noise events and each of the events excluding the plurality of aircraft noise events. It is an AI model learned to extract noise data according to the occurrence of the plurality of aircraft noise events among data.

In addition, the second AI model is based on labeling information matched to frequency data corresponding to noise according to the occurrence of a civil aircraft noise event and an aircraft noise event in a specific field among a plurality of aircraft noise events, aircraft noise learning data AI model learned to extract noise data generated by aircraft noise events in the specific field.

In addition, the third AI model uses noise learning data for each aircraft model in a specific field based on labeling information matched to frequency data corresponding to noise according to the occurrence of each noise event for each aircraft model in a specific field. It is an AI model trained to extract and classify noise data by aircraft type in a specific field.

And, at least one of the first AI model, the second AI model, the third AI model, or the fourth AI model is a multi-layer perceptron (MLP) using a functional application programming interface (API). , a convolutional neural network (CNN) or a recurrent neural network (RNN).

In addition, the aircraft noise classification and extraction method using artificial intelligence is an end-to-end model of the first AI model, the second AI model, the third AI model, and the fourth AI model. Constructing a fifth AI model by sequentially connecting in the form of; and inputting the background noise data to the fifth model through the one or more processors, and acquiring operation-related information for each type of aircraft in the specific field from the fifth model.

As another embodiment of the present disclosure, in the computing device for classifying and extracting aircraft noise using artificial intelligence, one or more geographically distributed noise sensors; a communication unit including a circuit; and one or more processors, wherein the one or more processors acquire background noise data generated by a plurality of events acquired through the one or more noise sensors through the communication unit, and determine the background noise data. A pre-processing operation is performed to obtain first noise data, and the first noise data is input to a first artificial intelligence (AI) trained to identify an aircraft noise event among the plurality of events to determine the aircraft noise event. Obtaining second noise data generated by, and inputting the second noise data into a second AI model learned to identify aircraft noise events in a specific field among the aircraft noise events, by the aircraft noise event in the specific field Acquiring the generated third noise data, inputting the third noise data to a third AI model learned to extract noise data for each aircraft type in a specific field through the one or more processors, and then inputting the third noise data for each aircraft type in the specific field It can be set to extract and classify data.

In addition to this, another method for implementing the present disclosure, another system, and a computer readable recording medium recording a computer program for executing the method may be further provided.

Other specific details of the disclosure are included in the detailed description and drawings.

According to an embodiment of the present disclosure, a method for classifying and extracting aircraft noise using artificial intelligence may be provided.

According to an embodiment of the present disclosure, it is possible to efficiently extract and classify noise data for each type of aircraft in a specific field based on artificial intelligence from noise generated by various events.

The effects of the present disclosure are not limited to the effects mentioned above, and other effects not mentioned will be clearly understood by those skilled in the art from the description below.

1 is a schematic diagram of a system implementing a method for classifying and extracting aircraft noise in a specific field using artificial intelligence, according to an embodiment of the present disclosure;
2 is a block diagram for explaining the configuration of a computing device for classifying and extracting aircraft noise in a specific field using artificial intelligence, according to an embodiment of the present disclosure;
3 and 4 are block diagrams for explaining an artificial intelligence model that performs an operation of classifying and extracting aircraft noise in a specific field according to an embodiment of the present disclosure;
5 is a flowchart illustrating a method for classifying and extracting aircraft noise in a specific field using artificial intelligence, according to an embodiment of the present disclosure.

Advantages and features of the present disclosure, and methods of achieving them, will become clear with reference to the embodiments described below in detail in conjunction with the accompanying drawings. However, the present disclosure is not limited to the embodiments disclosed below and may be implemented in various different forms, but only the present embodiments make the disclosure of the present disclosure complete, and are common in the art to which the present disclosure belongs. It is provided to fully inform the person skilled in the art of the scope of the present disclosure, which is only defined by the scope of the claims.

Terminology used herein is for describing the embodiments and is not intended to limit the present disclosure. In this specification, singular forms also include plural forms unless specifically stated otherwise in a phrase. As used herein, "comprises" and/or "comprising" does not exclude the presence or addition of one or more other elements other than the recited elements.

Like reference numerals throughout the specification refer to like elements, and “and/or” includes each and every combination of one or more of the recited elements. Although "first", "second", etc. are used to describe various components, these components are not limited by these terms, of course. These terms are only used to distinguish one component from another. Accordingly, it goes without saying that the first element mentioned below may also be the second element within the technical spirit of the present disclosure.

In the present disclosure, when a component is said to be "connected", "coupled" or "connected" to another component, this is not only a direct connection relationship, but also an indirect connection relationship between which another component exists. may also be included. Also in this disclosure, the terms "comprises" or "has" specify the presence of a stated feature, step, operation, element and/or component, but not one or more other features, steps, operations, elements, components and/or components. The presence or addition of groups of these is not excluded.

Unless otherwise defined, all terms (including technical and scientific terms) used in this specification may be used with meanings commonly understood by those skilled in the art to which this disclosure belongs. In addition, terms defined in commonly used dictionaries are not interpreted ideally or excessively unless explicitly specifically defined.

The spatially relative terms "below", "beneath", "lower", "above", "upper", etc. It can be used to easily describe a component's correlation with other components. Spatially relative terms should be understood as encompassing different orientations of elements in use or operation in addition to the orientations shown in the drawings.

For example, if you flip a component that is shown in a drawing, a component described as "below" or "beneath" another component will be placed "above" the other component. can Thus, the exemplary term “below” may include directions of both below and above. Components may also be oriented in other orientations, and thus spatially relative terms may be interpreted according to orientation.

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

1 is a diagram for explaining a system for implementing a method for classifying and extracting aircraft noise using artificial intelligence, according to an embodiment of the present disclosure.

As shown in FIG. 1, a system 1000 for implementing a method for classifying and extracting aircraft noise using artificial intelligence according to an embodiment of the present disclosure includes a calculation device 100, a plurality of noise sensors 200 -1, 200-2) and the cloud server 300.

FIG. 1 illustrates a case where each of the computing device 100, the plurality of noise sensors 200-1 and 200-2, and the cloud server is implemented as a separate device to configure one system. However, this is only an example, and the arithmetic device 100 and the plurality of noise sensors 200-1 and 200-2 may be configured as one device, which will be described with reference to FIG. 2 .

Each device included in the system 1000 may perform communication through the network W. Here, the network W may include a wired network and a wireless network. For example, the network may include various networks such as a local area network (LAN), a metropolitan area network (MAN), and a wide area network (WAN).

Also, the network may include the well-known World Wide Web (WWW). However, the network according to an embodiment of the present disclosure is not limited to the networks listed above, and may include at least a part of a known wireless data network, a known telephone network, and a known wired/wireless television network.

The plurality of noise sensors 200-1 and 200-2 may be geographically distributed to acquire (or collect) background noise data generated by a plurality of events. Although FIG. 1 shows a case in which there are two noise sensors, this is merely an example, and the number of noise sensors may be implemented with various values.

Here, the plurality of events may include various events that may generate (or cause) noise around the plurality of noise sensors 200-1 and 200-2 as well as operation of the airplane. That is, noise data generated by a plurality of events may refer to data related to background noise generated by various events occurring around the noise sensors 200-1 and 200-2.

The plurality of noise sensors 200-1 and 200-2 convert the acquired noise data into digital data and transmit it to the computing device 100 through the network W (ie, A (analog) / D in the noise sensor) (if a digital converter is included), but is not limited thereto, and analog type noise data may be transmitted to the arithmetic device 100 as it is.

The computing device 100 includes one or more processors and may perform aircraft noise classification and extraction operations in a specific field. One or more processors included in the associative device 100 may classify and extract aircraft noise in a specific field based on the background noise data received from the plurality of noise sensors 200-1 and 200-2.

Meanwhile, although FIG. 1 illustrates a case in which the computing device 100 is implemented as a desktop PC, it is not limited thereto. The associative device 100 may be implemented in various types of devices such as a tablet PC, a laptop computer, a smart phone, and a wearable device.

The cloud server 300 is a server storing an artificial intelligence (AI) model used when classifying and extracting aircraft noise in a specific field. For example, the first AI model, the second AI model, the third AI model, the fourth AI model, and the fifth AI model may be stored in the cloud server 300, and descriptions of each AI model will be described later. section to explain.

One or more processors included in the computing device 100 may use the AI model stored in the cloud server 300 . That is, the associative device 100 may utilize a cloud-based AI model.

However, this is only one embodiment, and each AI model may be stored on the computing device 100. That is, the computing device 100 may utilize an on-device based AI model. In this case, the AI model included in the computing device 100 may be learned within the cloud server 300 and transmitted to the computing device 100 or may be learned within the computing device 100 .

2 is a block diagram for explaining the configuration of a computing device for classifying and extracting aircraft noise in a specific field using artificial intelligence, according to an embodiment of the present disclosure.

As shown in FIG. 2 , the arithmetic device 100 may include a noise sensor 110 , a memory 120 , a communication unit 130 and a processor 140 .

However, the configuration shown in FIG. 1 is an exemplary diagram for implementing the embodiments of the present disclosure, and appropriate hardware and software configurations that are obvious to those skilled in the art may be additionally included in the computing device 100 .

The noise sensor 110 may acquire (or collect) background noise data generated by a plurality of events. Noise sensors 110 may be disposed in various positions of the computing device 100, and the number of noise sensors 110 may be one or more. The description of the noise sensor 110 of FIG. 2 corresponds to the plurality of noise sensors 200-1 and 200-2 of FIG. 1, and therefore, overlapping descriptions will be omitted.

The memory 120 may store one or more instructions for performing various operations. The memory 120 may store background noise data acquired through the noise sensor 110 . The memory 120 may store data acquired through a plurality of AI models (eg, a first AI model, a second AI model, a third AI model, a fourth AI model, and a fifth AI model).

The memory 120 may store data required to utilize the AI model (eg, data required to execute the on-device AI model).

The communication unit 130 includes one or more circuits and can communicate with an external device. Connecting the communication unit 130 to an external device may include communication through a third device (eg, a repeater, a hub, an access point, a server, or a gateway). The communication unit 130 may include various communication modules to communicate with external devices.

The communication unit 130 may communicate with a cloud server storing a plurality of AI models. For example, the communication unit 130 may exchange input data to be input to a plurality of AI models and output data acquired through the AI models with the cloud server.

As another example, when the noise sensor 110 is disposed outside the computing device 100 , the communication unit 130 may receive background noise data from the noise sensor 110 .

The processor 140 may be electrically connected to the memory 120 to control overall operations and functions of the arithmetic device 100 . Since the processor 140 of FIG. 2 corresponds to the processor of the arithmetic device 100 of FIG. 1 , overlapping descriptions will be omitted.

Processor 140 may consist of one or more processors. At this time, the one or more processors may include a general-purpose processor such as a CPU (Central Processing Unit), an AP (Application Processor), a DSP (Digital Signal Processor), a graphics-only processor such as a GPU (Graphic Processing Unit), a VPU (Vision Processing Unit), or It can be implemented with an artificial intelligence dedicated processor such as an NPU (Neural Processing Unit). The artificial intelligence-only processor may refer to a processor designed with a hardware structure specialized for processing a specific artificial intelligence model.

The processor 140 may obtain first noise data by performing a preprocessing operation on the background noise data. The pre-processing operation means a preliminary operation operation for separating and removing unnecessary information and processing it in order to process it in a form suitable for the purpose and method of data analysis.

Specifically, before applying the background noise data to the AI model (ie, before performing an inference operation (eg, classification or extraction operation) using the AI model), the processor 140 performs a preprocessing operation on the background noise data. can be done For example, the processor 140 may obtain first noise data by performing a preprocessing operation of dividing the background noise data into preset length units.

As another example, when the background noise data is an analog signal, the processor 140 may perform a preprocessing operation of converting the background noise data into a digital signal.

The processor 140 may obtain second noise data generated by the aircraft noise event by inputting the first noise data to the first AI model learned to identify the aircraft noise event among the plurality of events. That is, the first AI model is an artificial intelligence model learned to extract or classify noise generated by aircraft operation among background noise.

The first AI model is based on labeling information matched to frequency data corresponding to noise according to the occurrence of each of the plurality of aircraft noise events and the plurality of aircraft noise events, and a plurality of aircraft among the background noise learning data. It may be learned to extract noise data according to the occurrence of a noise event.

Here, the frequency data means data from which frequency components can be identified, such as a frequency spectrum (eg, an amplitude spectrum, a phase spectrum, an energy spectrum, a power spectrum, etc.).

For example, the first AI model extracts noise data according to the occurrence of an aircraft noise event from learning data including various background noises based on information labeled on a frequency spectrum corresponding to noise according to the occurrence of an aircraft noise event. can be learned

In one embodiment of the present disclosure, the processor 140 identifies aircraft noise through the first AI model, and then uses an aircraft noise evaluation program to determine a Weighted Equivalent Continuous Perceived Noise Level (WECPNL) value and an Lden value. ·Day-evening-night Average Sound Level) can be automatically calculated. The aircraft noise evaluation program is a program that extracts Leq, Lmax, and Leq values for each frequency from the measured noise measurement data, identifies aircraft noise using the first AI model, and calculates WECPNL and Lden.

Here, WECLE means a daily aircraft noise exposure index calculated using the highest noise level of an aircraft, and LDN means a daily aircraft noise level derived by measuring an equivalent noise level of an aircraft.

Here, for the verification of the noise evaluation program, the WECPNL value and the Lden value were compared by inputting new noise data of 15 days not used for learning. The comparison revealed some errors in WECPNL and Lden values.

Meanwhile, the processor 140 may extract noise generated by aircraft operation from background noise and store the remaining noise as non-aircraft noise in a separate database.

The processor 140 acquires third noise data generated by aircraft noise events in a specific field by inputting the second noise data to the second AI model learned to identify aircraft noise events in a specific field among aircraft noise events. can do. That is, the second AI model is an artificial intelligence model learned to extract or classify noise generated by operation of an aircraft in a specific field among a plurality of aircraft.

The second AI model is a specific field among aircraft noise learning data based on labeling information matched to frequency data corresponding to noise according to the occurrence of each of a plurality of aircraft noise events and a civil aircraft noise event and an aircraft noise event in a specific field. It can be learned to extract noise data generated by aircraft noise events of

For example, the second AI model is based on information labeled on a frequency spectrum corresponding to noise caused by the occurrence of an aircraft noise event, in learning data including various aircraft noise, noise data according to the occurrence of an aircraft noise event in a specific field. It can be learned to extract

Meanwhile, the processor 140 may extract noise generated by aircraft operation in a specific field from aircraft noise, and store the remaining noise as aircraft noise in a non-specific field in a separate database.

The processor 140 may extract and classify the noise data for each aircraft type in the specific field by inputting the third noise data to the third AI model learned to extract the noise data for each aircraft type in the specific field. That is, the third AI model is an artificial intelligence model learned to extract and classify noise generated for each type of aircraft in a specific field.

The third AI model is an aircraft in a specific field using noise learning data for each aircraft type in a specific field based on labeling information matched to frequency data corresponding to noise according to the occurrence of each noise event for each aircraft model in a specific field. It can be learned to extract and classify noise data for each aircraft type.

For example, the third AI model is based on information labeled on a frequency spectrum corresponding to noise according to the occurrence of each noise event for each aircraft type in a specific field, in learning data including aircraft noise in various specific fields. It can be learned to extract noise data for each type of aircraft in the field.

The processor 140 inputs the extracted and classified noise data for each type of aircraft in the specific field to the fourth AI model learned to extract the driving related information, and extracts the driving related information for each aircraft type in the specific field. there is.

That is, the fourth AI model means an artificial intelligence model learned to extract operation time, noise level, and frequency-related information from noise data according to the occurrence of aircraft noise events in a specific field.

Here, the operation-related information for each type of aircraft in the specific field may include at least one of the operating time of each aircraft type in the specific field, the level of noise generated according to operation, or frequency-related information (eg, frequency spectrum, etc.).

On the other hand, at least one of the first AI model, the second AI model, the third AI model, or the fourth AI model is a multi-layer perceptron (MLP) using a functional API (application programming interface), It may be composed of a combination of at least one of a convolutional neural network (CNN) and a recurrent neural network (RNN).

Functional API means a tool that defines each layer as a kind of function. Functional APIs layer like functions and can directly handle the input and output of tensors. When using the functional API, layer objects can be reused multiple times to share weights and learn together. In other words, functional APIs can be the tools you need to build advanced deep neural network architectures.

CNN is a type of multilayer artificial neural network used to analyze visual images, and refers to a deep learning algorithm that learns photos, videos, and voice problems. A CNN can be configured to convert an input image into a mathematical matrix and then obtain several feature maps by filters in a convolutional layer. The feature map obtained from the convolutional layer is further compressed while preserving the most important part of the image through a pooling layer, and the pooled image can be made into a single column through flattening.

Data input to the CNN is generally image data. In order to input noise data into a CNN, noise data (i.e., numbers) can be converted into images. However, this is only an example, and the noise data may be changed to data located between 0 and 1 (MinMaxScaler), and the changed data may be input to the CNN.

RNN refers to a type of neural network that specializes in sequential data processing. CNNs cannot remember the results of previous data, but RNNs can use not only current input data but also experiences over time as input data by remembering previous states.

In an embodiment of the present disclosure, an AI model (a first AI model, a second AI model, a third AI model, a fourth AI model, and a fifth AI model) composed of a combination of at least one of MLP, RNN, and CNN is dense It is composed of at least one of a dense layer, an LSTM layer, and a Conv2D layer, and an activation function may use a ReLU function or/and a sigmoid function. Binary_crossentropy may be used as a loss function used when performing a learning operation for each AI model, and Adam may be used as an optimizer. Here, the batch size may be set to 20, and an Early Stopping callback may be used to prevent overfitting. As a result of learning, the accuracy of the test data of the AI model was 99.55%, and the loss value was 0.016. However, examples of functions that can be applied in relation to the AI model are not limited to the above examples, and various functions may be used.

Meanwhile, a fifth AI model may be configured by sequentially connecting the first AI model, the second AI model, the third AI model, and the fourth AI model in the form of an end-to-end model. An embodiment related to this will be described with reference to FIG. 4 .

3 and 4 are block diagrams for explaining an artificial intelligence model performing an operation of classifying and extracting aircraft noise in a specific field, according to an embodiment of the present disclosure.

The computing device may obtain first noise data by inputting the background noise data obtained through the noise sensor to the preprocessing module 310 . The first noise data refers to data on which a preprocessing operation for performing an artificial intelligence-based classification and extraction operation is performed on the background noise data.

The computing device may obtain second noise data by inputting the first noise data to the first AI model 320 . Here, the second noise data means aircraft noise data among background noise.

The calculation device may obtain third noise data by inputting the second noise data to the second AI model 330 . Here, the third noise data means aircraft noise data of a specific field among aircraft noise.

The calculation device may input the third noise data to the third AI model 340 to obtain noise data for each aircraft type in a specific field. The computing device may input noise data for each type of aircraft in a specific field to the fourth AI model 350 to obtain operation-related information for each type of aircraft in a specific field.

4 is a first AI model 320, a second AI model 330, a third AI model 340, and a fourth AI model 350 sequentially connected in the form of an end-to-end model to form one AI model ( That is, it shows an embodiment in which the fifth AI model 400) is configured. The fifth AI model 400 may be trained to extract operation-related information for each aircraft in a specific field based on the first noise data preprocessed for the background noise data.

As shown in FIG. 4, it is assumed that the fifth AI model 400 is configured. The computing device may acquire operation-related information for each aircraft in a specific field by inputting the first noise data into the fifth AI model 400 .

5 is a flowchart illustrating a method for classifying and extracting aircraft noise in a specific field using artificial intelligence, according to an embodiment of the present disclosure. If a description of terms and operations included in the flowchart disclosed in FIG. 5 is duplicated with the previous description, it will be omitted.

Background noise data generated by a plurality of events may be acquired through one or more geographically dispersed noise sensors (S510).

The first noise data may be obtained by performing a preprocessing operation on the background noise data through one or more processors (S520).

Second noise data generated by the aircraft noise event may be obtained by inputting the first noise data to the first AI model learned to identify the aircraft noise event among the plurality of events through one or more processors (S530). .

Through one or more processors, the third noise data generated by the aircraft noise event in the specific field is acquired by inputting the second noise data into the second AI model learned to identify the aircraft noise event in the specific field among the aircraft noise events. It can be (S540).

Through one or more processors, noise data for each aircraft type in a specific field may be extracted and classified by inputting third noise data to a third AI model learned to extract noise data for each aircraft type in a specific field (S550). .

Through one or more processors, the extracted and classified noise data for each aircraft type in a specific field is input to the fourth AI model learned to extract operation-related information, so that operation-related information for each aircraft type in a specific field can be extracted. Yes (S560).

Meanwhile, the above-described embodiments of the present disclosure can be written as a program that can be executed on a computer, and the written program can be stored in a medium.

The medium may continuously store programs executable by a computer or temporarily store them for execution or download. In addition, the medium may be various recording means or storage means in the form of a single or combined hardware, and is not limited to a medium directly connected to a certain computer system, and may be distributed on a network.

Examples of the medium include magnetic media such as hard disks, floppy disks and magnetic tapes, optical recording media such as CD-ROM and DVD, magneto-optical media such as floptical disks, and ROM, RAM, flash memory, etc. configured to store program instructions. In addition, examples of other media include recording media or storage media managed by an app store that distributes applications, a site that supplies or distributes various other software, and a server.

Above, the embodiments described in the present disclosure and the accompanying drawings merely illustrate some of the technical ideas included in the present disclosure by way of example. Therefore, since the embodiments disclosed in this specification are intended to explain rather than limit the technical spirit of the present disclosure, it is obvious that the scope of the technical spirit of the present disclosure is not limited by these embodiments.

All modifications and specific examples that can be easily inferred by those skilled in the art within the scope of the technical idea included in the specification and drawings of the present disclosure should be construed as being included in the scope of the present disclosure.

Claims (10)

In the aircraft noise classification and extraction method using artificial intelligence,
obtaining background noise data generated by a plurality of events via one or more geographically distributed noise sensors;
obtaining first noise data by performing a pre-processing operation on the background noise data through one or more processors;
Through the one or more processors, the first noise data is input to a first artificial intelligence (AI) model learned to identify an aircraft noise event among the plurality of events, so that a second noise event generated by the aircraft noise event is input. 2 acquiring noise data;
Through the one or more processors, the second noise data is input to a second AI model learned to identify aircraft noise events in a specific field among the aircraft noise events, and the third noise event generated by the aircraft noise event in the specific field is input. acquiring noise data;
extracting and classifying noise data for each aircraft type by inputting the third noise data to a third AI model learned to extract noise data for each aircraft type through the one or more processors; and
Inputting the extracted and classified noise data for each aircraft type into a fourth AI model learned to extract driving-related information through the one or more processors, and extracting driving-related information for each of the aircraft types; ,
At least one of the first AI model, the second AI model, the third AI model, or the fourth AI model,
Constructed by a combination of a multi-layer perceptron (MLP), a convolutional neural network (CNN), and a recurrent neural network (RNN) using a functional application programming interface (API),
the processor,
After identifying the second noise data through the first AI model, it is configured to calculate a Weckle value and an Eldin value using an aircraft noise evaluation program,
The aircraft noise evaluation program,
It is configured to extract an equivalent noise level value of the aircraft, a maximum noise level value of the aircraft, and an equivalent noise level value of the aircraft for each frequency from the background noise data,
The Weckle value is,
A daily aircraft noise exposure index calculated using the highest noise level value of the aircraft,
The eldien value is
It is the daily aircraft noise level derived using the equivalent noise level value of the aircraft,
The aircraft noise evaluation program,
15 new noise data not used for learning by the first AI model is input to the aircraft noise evaluation program and configured to be verified by comparing the Weckle value with the Eldin value. delete According to claim 1,
Operation-related information for each of the aircraft types,
A method comprising at least one of the operating time of each aircraft type, noise level generated according to operation, or frequency-related information. According to claim 1,
Obtaining the first noise data,
And obtaining the first noise data by performing a preprocessing operation of dividing the background noise data into predetermined length units. According to claim 1,
The first AI model,
Based on labeling information matched to frequency data corresponding to noise according to occurrence of a plurality of aircraft noise events and each of the events excluding the plurality of aircraft noise events, generation of the plurality of aircraft noise events among the background noise learning data An AI model trained to extract noise data according to, Method. According to claim 1,
The second AI model,
Based on the labeling information matched to the frequency data corresponding to the noise according to the occurrence of each of the civil aircraft noise event and the aircraft noise event in the specific field among the plurality of aircraft noise events, the aircraft noise event in the specific field among the aircraft noise learning data An AI model trained to extract noise data generated by the method. According to claim 1,
The third AI model,
Noise data for each aircraft type in the specific field is obtained by using the noise learning data for each aircraft type in the specific field based on the labeling information matched to the frequency data corresponding to the noise caused by the occurrence of each noise event for each aircraft type in the specific field. A method, an AI model trained to extract and classify. delete According to claim 1,
Configuring a fifth AI model by sequentially connecting the first AI model, the second AI model, the third AI model, and the fourth AI model in the form of an end-to-end model ; and
Further comprising, through the one or more processors, inputting the background noise data to the fifth AI model, and acquiring operation-related information for each aircraft type in the specific field from the fifth AI model. In the calculation device for classifying and extracting aircraft noise using artificial intelligence,
one or more geographically distributed noise sensors;
a communication unit including a circuit; and
includes one or more processors;
The one or more processors,
Obtaining background noise data generated by a plurality of events acquired through the one or more noise sensors through the communication unit;
Obtaining first noise data by performing a preprocessing operation on the background noise data;
Obtaining second noise data generated by the aircraft noise event by inputting the first noise data to a first artificial intelligence (AI) trained to identify an aircraft noise event among the plurality of events;
Obtaining third noise data generated by the aircraft noise event in the specific field by inputting the second noise data into a second AI model learned to identify aircraft noise events in a specific field among the aircraft noise events;
Through the one or more processors, the third noise data is input to a third AI model learned to extract noise data by aircraft type in a specific field to extract and classify noise data by aircraft type in a specific field,
Through the one or more processors, the extracted and classified noise data for each aircraft type in the specific field is input to a fourth AI model learned to extract operation-related information, and operation-related information for each aircraft type is extracted,
At least one of the first AI model, the second AI model, the third AI model, or the fourth AI model,
Constructed by a combination of a multi-layer perceptron (MLP), a convolutional neural network (CNN), and a recurrent neural network (RNN) using a functional application programming interface (API),
the processor,
After identifying the second noise data through the first AI model, it is configured to calculate a Weckle value and an Eldin value using an aircraft noise evaluation program,
The aircraft noise evaluation program,
It is configured to extract an equivalent noise level value of the aircraft, a maximum noise level value of the aircraft, and an equivalent noise level value of the aircraft for each frequency from the background noise data,
The Weckle value is,
A daily aircraft noise exposure index calculated using the highest noise level value of the aircraft,
The Eldien value is,
It is the daily aircraft noise level derived using the equivalent noise level value of the aircraft,
The aircraft noise evaluation program,
An operation device configured to input new noise data of 15 days not used for learning by the first AI model to the aircraft noise evaluation program and verify by comparing the Weckle value and the Eldin value.

Images