KR102191083B1 - Method for controlling bus information terminal(BIT) and bus information terminal for performing the method - Google Patents

Abstract

The present invention relates to a control method of a bus information providing terminal and a bus information providing terminal performing the method thereof. The control method of the bus information providing terminal may include: a step of receiving, by a bus information providing terminal, a sound signal; a step of determining, by the bus information providing terminal, text information to be analyzed based on the sound signal; and a step of providing, by the bus information providing terminal, information on a target bus corresponding to the text information to be analyzed.

Description

A method for controlling a bus information providing terminal and a bus information providing terminal for performing such a method {Method for controlling bus information terminal (BIT) and bus information terminal for performing the method}

The present invention relates to a control method of a bus information providing terminal and a bus information providing terminal performing the method. More specifically, it relates to a method for controlling a bus information terminal (BIT) that provides bus-related information for a disabled person such as a visually impaired person, and an apparatus for performing the method.

Currently operating city bus stops may have a rather narrow waiting space due to the central lane system for buses operating in some areas of Seoul, so if you look at the answers to the inconveniences, including these inconveniences, the most answers. It accounts for 27.4% of the buses that do not stop within the Lee' stop area. Among the responses by type of disability, both the physically impaired and the brain lesion disabled group had the highest answer of being uncomfortable due to the bus that does not stop within the bus stop area, with the highest 26% and 25.5%. Bus information terminal) is the reason that there are many stops that do not provide guidance, accounting for 31.1% and 30.5%, respectively, so it can be seen that more information is needed.

In the case of the visually impaired group, 29.5% of the respondents answered that it was due to the bus that does not stop within the bus stop area, indicating the importance of information guidance as well as the inaccuracy of the city bus stop area. Also, the visually impaired group responded that it was very inconvenient that there was no voice guidance for the arriving city bus.

That is, at present, visually impaired people are inconvenient to board a bus because bus information is not accurately transmitted at a bus stop, and development of a bus information providing terminal (BIT) for accurately transmitting such bus information is required.

An object of the present invention is to solve all of the above-described problems.

In addition, an object of the present invention is to implement a voice-based bus information guidance system for the visually impaired.

In addition, an object of the present invention is to provide accurate bus information by recognizing the motion of the visually impaired and by removing noise from the input sound.

A typical configuration of the present invention for achieving the above object is as follows.

According to an embodiment of the present invention, a method of controlling a bus information providing terminal includes the steps of receiving, by the bus information providing terminal, a sound signal, and determining, by the bus information providing terminal, text information to be analyzed based on the sound signal. And providing, by the bus information providing terminal, information on a target bus corresponding to the text information to be analyzed.

Meanwhile, the analysis target text information may be determined based on a trigger word on candidate text information extracted based on the bus information terminal.

In addition, the analysis target text information is divided into a main analysis text and a sub analysis text based on the trigger word, and the bus information providing terminal may determine the target bus information by prioritizing analysis on the main analysis text. have.

According to another embodiment of the present invention, the bus information providing terminal includes a voice input unit implemented to receive a sound signal and a processor operatively connected to the voice input unit, wherein the processor analyzes the sound signal based on the sound signal. It may be implemented to determine target text information and provide information on a target bus corresponding to the analysis target text information.

Meanwhile, the analysis target text information may be determined based on a trigger word on candidate text information extracted based on the bus information terminal.

Also, the analysis target text information is divided into a main analysis text and a sub analysis text based on the trigger word, and the processor may determine the target bus information by prioritizing analysis on the main analysis text.

According to the present invention, a voice-based bus information guidance system for the visually impaired can be implemented.

In addition, according to the present invention, accurate bus information can be provided by accurately recognizing the voice of the visually impaired by recognizing the motion of the visually impaired and removing noise from the input sound.

1 is a conceptual diagram showing a bus information providing terminal according to an embodiment of the present invention.
2 is a conceptual diagram showing a speech analysis operation of a speech analysis unit according to an embodiment of the present invention.
3 is a conceptual diagram showing the operation of a speech analysis unit according to an embodiment of the present invention.
4 is a conceptual diagram showing an operation of an image analysis unit according to an embodiment of the present invention.
5 is a conceptual diagram showing an object motion analysis algorithm according to an embodiment of the present invention.
6 is a conceptual diagram showing a noise removal method according to an embodiment of the present invention.

DETAILED DESCRIPTION OF THE INVENTION The detailed description of the present invention to be described later refers to the accompanying drawings, which illustrate specific embodiments in which the present invention may be practiced. These embodiments are described in detail sufficient to enable those skilled in the art to practice the present invention. It is to be understood that the various embodiments of the present invention are different from each other, but need not be mutually exclusive. For example, specific shapes, structures, and characteristics described herein may be changed from one embodiment to another and implemented without departing from the spirit and scope of the present invention. In addition, it should be understood that the positions or arrangements of individual elements in each embodiment may be changed without departing from the spirit and scope of the present invention. Therefore, the detailed description to be described later is not intended to be limited, and the scope of the present invention should be taken as encompassing the scope claimed by the claims of the claims and all scopes equivalent thereto. Like reference numerals in the drawings indicate the same or similar elements over several aspects.

Hereinafter, various preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings in order to enable those skilled in the art to easily implement the present invention.

1 is a conceptual diagram showing a bus information providing terminal according to an embodiment of the present invention.

In FIG. 1, a bus information providing terminal is disclosed for analyzing movement of a visually impaired person in a bus information providing terminal, removing noise from an input sound, and extracting necessary voice information.

Referring to FIG. 1, the bus information providing terminal includes an image capture unit 100, an image analysis unit 110, an audio input unit 120, an audio analysis unit 130, a bus information providing unit 140, and a processor 150. It may include.

The image capture unit 100 may be implemented to capture an image of a person waiting for boarding located at a bus stop. Specifically, the image capture unit 100 includes image information (ignition) for detecting whether a blind person ignites at the bus information providing terminal among the waiting for boarding, and a target bus to which the boarding waiter (visually impaired) wants to ride is boarded upon arrival. It may be implemented to capture image information (boarding) for notifying boarding to a waiting person (visually impaired).

The image analysis unit 110 may be implemented to analyze an image captured by the image pickup unit. The image analysis unit 110 may analyze image information (speech) and image information (boarding) among image information captured as described above. The image analysis unit 110 extracts an image in which a person waiting for boarding (visually impaired) approaches the bus information providing terminal from the imaged images and makes an utterance, and determines the image information (speech), and determines the target that the occupant (visually impaired) wishes to board. After the arrival and arrival of the bus, image information (boarding), which is an image for confirming whether the occupant (visually impaired) boards the target bus, may be generated.

The image analysis unit 110 may perform image analysis to remove noise from the input voice by extracting the utterance time from the image information (speech), and analyze the image information (boarding) to analyze the motion of the occupant (visually impaired). When the target bus of the occupant (visually impaired) arrives by tracking, information for boarding the visually impaired may be provided.

The voice input unit 120 may be implemented to receive a sound signal input to a bus information providing terminal.

The voice analysis unit 130 may be implemented to remove a noise signal from among sound signals transmitted to the voice input unit, and extract and analyze only a voice signal to be analyzed. The voice analysis unit 130 may perform a separate noise determination procedure in order to extract an analysis target voice signal for inquiring about bus information. In addition, the voice analysis unit 130 is implemented to extract the voice signal to be analyzed by removing noise by analyzing the utterance time of the occupant (visually impaired) through the image analysis unit 110 in conjunction with the image analysis unit 110 Can be.

The bus information providing unit 140 may be implemented to provide target bus information based on an analysis target voice signal analyzed by the voice analysis unit 130. For example, when the voice signal to be analyzed is'Bus XXX', information (waiting time, arrival or not) on bus XXX, which is a target bus, may be transmitted by the bus information provider.

The processor 150 may be implemented to control the operation of the image capture unit 100, the image analysis unit 110, the audio input unit 120, the audio analysis unit 130, and the bus information providing unit 140.

In FIG. 1, for convenience of explanation, it is assumed that each component such as the image analysis unit 110 and the audio analysis unit 130 is included in the bus information providing terminal. However, the terminal is implemented in a separate location to provide bus information through a wired or wireless communication network. It may be implemented in a structure that can communicate with.

2 is a conceptual diagram showing a speech analysis operation of a speech analysis unit according to an embodiment of the present invention.

In FIG. 2, a method for extracting a voice signal to be analyzed by analyzing a voice input to a voice analysis unit is disclosed.

Referring to FIG. 2, the speech analysis unit may determine candidate text information 200 from an input sound signal, and may determine an analysis target text 240 by removing noise text information 220 from the candidate text information 200. .

The candidate text information 200 may be information obtained by converting a sound signal transmitted to the voice input unit into text. The speech analysis unit may include a text conversion module that converts the sound signal into text.

The candidate text information 200 may include not only a sound signal asking for a bus number, but also textual information of various other sound signals. Information obtained by textualizing a sound signal asking for a bus number may be text information to be analyzed, and other text may be noise text information 220.

The speech analysis unit may separate the noise text information 220 and the analysis target text information 240 from the candidate text information 200. A procedure of separating the noise text information from the text information to be analyzed 240 from the candidate text information 200 may be expressed in terms of a noise removal procedure. The noise removal procedure may be performed based on a trigger word or may be performed by additionally considering image information. The noise removal procedure will be described later.

When the text information to be analyzed 240 is extracted, the target bus 260 is determined by comparing the text information to be analyzed 240 and the bus information stored in the bus database 250, and the information on the target bus 260 is It can be delivered through the bus information provider.

If the matching degree between the analysis target text information 240 and the bus information stored in the bus database 250 is equal to or greater than a threshold value, information on the matched bus may be provided as information on the target bus 260. When the matching degree between the analysis target text information 240 and the bus information stored in the bus database 250 is less than the threshold value, an additional inquiry for the target bus 260 may be performed.

3 is a conceptual diagram showing the operation of a speech analysis unit according to an embodiment of the present invention.

In FIG. 3, a noise removal procedure based on a trigger word is disclosed.

Referring to FIG. 3, the trigger word 300 may be a word used when inquiring about a bus number.

When inquiring about the bus number, a sound signal including'number','bus', and'number word (thousand, hundred twenty four)' may be input. Words such as'burn','bus', and'number words (thousand, hundred twenty four)' may be trigger words, and candidate text information including these words may be extracted as the text to be analyzed 310. A plurality of trigger words 300 may be included in the analysis target text 310. For example, in the case of a bus number one hundred forty-two, a trigger word may be set as a first trigger word and a second trigger word as'one hundred forty-two' and'bus', and the analysis target text 310 may be determined.

The analysis target text 310 may be classified into a main analysis text 330 and a sub analysis text 320 and 340 based on a trigger word. The main analysis text 330 is a text to be primarily analyzed based on the trigger word 300, and the sub-analysis texts 320 and 340 are texts to be analyzed secondary after the analysis of the main analysis text 330 I can.

The main analysis text 330 may be set differently according to the trigger word 300. For example, when the trigger word 300 is'time', the part corresponding to the number in front of the'time' may be the main analysis text 330, and the range of the main analysis text 330 is'time' It may be n texts adjacent to'. In this case, n is a value that is adaptively changed according to the trigger word 300, and may be determined to be a value equal to or greater than the first threshold value in which a probability of including valid information based on the existing trigger word. This probability value may be determined based on learning of the text to be analyzed 310 including the existing trigger word 300.

When a plurality of trigger words 300 are present in succession, the main analysis text 330 may be set based on the trigger word 300 located preferentially among the trigger words 300.

As texts other than the main analysis text 330, x texts that are temporally in front of the main analysis text 330 and y texts that are temporally behind the main analysis text 330 may be sub-analysis texts 320 and 340.

The x texts present temporally in front of the main analysis text 330 are the first sub-analysis texts 320, and y texts that are temporally behind the main analysis text 330 are the second sub-analysis. It may be text 340.

The sizes of x and y are values that are adaptively changed according to the trigger word 300, and may be determined as a value that is equal to or greater than the second threshold value of a probability of including valid information based on the existing trigger word 300. This probability value may be determined based on learning of the text to be analyzed 310 including the existing trigger word 300.

Priority of the first sub-analysis text 320 and the second sub-analysis text 340 may be set for each trigger word 300. For example, in the case of the trigger word A, the first sub-analysis text 320 has a relatively high priority and is set as the first sub-analysis text (first priority), and the second sub-analysis text 340 is relatively It may be set as the second sub-analysis text (second priority) with a low priority. Conversely, in the case of the trigger word B, the first sub-analysis text 320 has a relatively low priority and is set as the first sub-analysis text (second priority), and the second sub-analysis text 340 has a relatively high priority. It may be set as a second sub-analysis text (first priority) with a priority.

First, when the reliability of the first bus information determined based on the main analysis text 330 is greater than or equal to the threshold, the bus information providing unit may provide information on the target bus without analyzing the sub-analysis texts 320 and 340. I can.

The reliability may be determined in consideration of whether bus information matching the main analysis text 330 exists and whether a target bus and a similar bus number exist. The similar bus number is a case where the number included in the target bus number and the number included in the target bus number are equal to or greater than a threshold number (eg, two).

First, when the reliability of the first bus information determined based on the main analysis text 330 is less than the threshold value, the second bus information may be extracted based on the text further including the first priority among the sub-analysis texts.

If the first bus information and the second bus information are the same, a bus corresponding to the same first bus information and the second bus information is set as a target bus, and the bus information providing unit can provide information on the target bus. have.

Conversely, when the first bus information and the second bus information are not the same, the third bus information may be extracted by additionally including a sub-analysis text of the next order.

When bus information corresponding to a majority of the first bus information, the second bus information, and the third bus information exists, a bus corresponding to the corresponding bus information may be set as a target bus and information on the target bus may be transmitted.

When the first bus information, the second bus information, and the third bus information are all different, the main analysis text 330 may be re-extracted through a procedure of inquiring for bus information again.

4 is a conceptual diagram showing an operation of an image analysis unit according to an embodiment of the present invention.

In FIG. 4, a method for extracting a main analysis text by removing noise based on an image is disclosed.

Referring to FIG. 4, based on the image capture unit 470 located in the bus information providing terminal 460, the access of the occupant (especially, the occupant (visually impaired)) is analyzed and the text analysis time is set to extract the main analysis text. A method for doing is disclosed.

The image analysis unit determines the utterance start point 410 by extracting image information (speech) 400 through analysis of the object (person) approaching the image pickup unit 470, and determines the utterance start point 410 based on the utterance start point 410 The main analysis text 450 may be determined.

The criterion of the utterance start time point 410 may be set in various ways. The utterance start time point 410 may be determined as a time point when a sound signal having a threshold level or higher among input sounds is generated after the object moves within a threshold distance based on the bus information providing terminal 460.

Alternatively, the image analysis unit may determine the generation time of a sound signal having a threshold level or more among input sounds as the utterance start time 410 based on a time point when a mouth of an object approaching within a threshold distance from the bus information providing terminal 460 moves. .

The text generated after the utterance start time 410 may be determined as the main analysis text 450. When the speech start time point 410 is additionally considered, the first sub-analysis text positioned before the main analysis text 450 may not be separately extracted.

In addition, according to an embodiment of the present invention, the main analysis text 450 may be determined only with the utterance start time point 410, but if the utterance end time point 420 can be determined through image analysis, a separate first sub-analysis text Also may not be extracted.

That is, when the utterance start point 410 and the utterance end point 420 are clearly determined through image analysis, the sound signals input to the utterance start point 410 and the utterance end point 420 without the above-described sub-analysis text The target bus can be determined by setting the as the main analysis text.

When the reliability of the utterance start time point 410 and the utterance end time point 420 is low as a result of the image analysis, the target bus additionally including the first sub-analysis text and the second sub-analysis text based on the above-described main analysis text 450 Decisions can be made.

As a result of image analysis, the reliability of the utterance start point 410 and the utterance end point 420 can be determined in consideration of the movement of the mouth of the object and the start point of the sound input to the bus information providing terminal 460 and the end point of the sound. have. Speech based on the degree of synchronization between the start point of the mouth movement and the start point of the sound input to the bus information providing terminal 460, the end point of the mouth movement, and the end point of the sound input to the bus information providing terminal 460 Reliability of the start point 410 and the end point of speech 420 may be determined.

Alternatively, according to an embodiment of the present invention, candidate text information may be extracted by performing a search for a trigger word based on the utterance start time point 410. In the method described above in FIGS. 2 and 3, a trigger word search section for searching for a trigger word based on image information (speech) may be set, and an analysis target text may be determined by searching for a trigger word on the trigger word search section. .

In addition, according to an embodiment of the present invention, after the video is uttered, the boarding waiter (visually impaired) matching the utterance is tracked, and the target bus to which the boarding waiter (visually impaired) wants to ride is allowed to board the boarding standby (visually impaired) upon arrival. I can tell. After the ignition, the waiting for boarding (visually impaired) is set as a tracking object, and the target bus of the waiting for boarding (visually impaired) notifies the waiting for boarding (visually impaired) upon arrival, and based on the captured image, the waiting for boarding (visually impaired) You can analyze whether you are on board. After the arrival of the target bus through the captured image, when it is analyzed that the boarding waiter (visually impaired) does not board the target bus, a signal notifying the boarding waiter (visually impaired) to board the target bus may be transmitted. .

5 is a conceptual diagram showing an object motion analysis algorithm according to an embodiment of the present invention.

In FIG. 5, an operation of an image capture unit for generating an image for analyzing an image related to an utterance of an object is started.

Referring to FIG. 5, the image pickup unit may change an image pickup focus for more accurate image analysis.

The image pickup unit may perform image capturing by switching three focal points. Specifically, the first focus 510 may be a default focus as a focus on the entire area of the stop. The second focus 520 may be a focus for tracking a target object moving to the bus information providing terminal as a converted focus when there is a target object moving to the set bus information providing terminal.

The third focal point 530 is a converted focal point when the target object moving to the bus information providing terminal approaches within a critical distance from the bus information providing terminal, and displays an image of a specific part of the target object, such as a mouth, that makes an utterance. It may be a focal point for imaging.

An utterance start time point and an utterance end time point may be determined based on the third focus 530.

Additionally, image analysis may be performed by setting different regions for analyzing images without such a change in focus.

The primary image analysis area 515 may be an entire imaging area in order to detect an object (person) in the entire captured image area.

The secondary image analysis area 525 may set an object moving to the bus information providing terminal as a target object and may be an area corresponding to the target object.

What is the third image analysis area 535? When the object moves within a threshold distance from the bus information providing terminal, it may be an area corresponding to the mouth of the target object.

6 is a conceptual diagram showing a noise removal method according to an embodiment of the present invention.

In FIG. 6, a method for determining a noise text to separate a noise text and an analysis target text from candidate text information is disclosed.

Referring to FIG. 6, noise text may be determined based on sound signals input through a plurality of microphones.

A plurality of microphones may be installed inside the bus stop, and the bus information providing terminal may perform a noise removal procedure by analyzing sound signals received by the plurality of microphones.

For example, microphone 1, microphone 2, microphone 3, and microphone 4 may be present, and one microphone among a plurality of microphones may be set as a speech reference microphone 600, and the other microphones may be set as a noise reference microphone 650. .

The speech reference microphone 600 may be a microphone that is set closest to the bus information providing terminal and receives the loudest sound signal uttered by the target object. The noise reference microphone 650 may be a microphone located relatively farther from the utterance reference microphone 600 with respect to a bus information providing terminal.

A sound signal corresponding to the text to be analyzed may be input relatively louder to the speech reference microphone 600, and a sound signal corresponding to the noise text may be input relatively louder to the noise reference microphone 650.

Specifically, among the sounds input to the speech reference microphone 600, a sound that is relatively louder than the sound input to the noise reference microphone 650 may be the text to be analyzed, and the speech reference microphone among sounds input to the noise reference microphone 650 A sound that is relatively louder than the sound input to 600 may be a noise text.

In this way, the text to be analyzed and the noise text may be classified based on the size of the sound signal input to the speech reference microphone 600 and the noise reference microphone 650.

The analysis target text extracted based on the microphone may be expressed in terms of the analysis target text (microphone), and the noise text extracted based on the microphone may be expressed as the noise text (microphone).

The analysis target text (microphone) and noise text (microphone) are used to determine the target bus together with the analysis target text (trigger) and noise text (trigger) extracted based on the above-described trigger, or independently determine the target bus. Can be used for

For example, the target bus is determined by combining the text to be analyzed (microphone) and the text to be analyzed (trigger), or the target bus determined based only on the text to be analyzed (microphone) and the target determined based only on the text to be analyzed (trigger). The bus can also be determined individually.

In addition, according to an embodiment of the present invention, the noisy text (microphone) can be used for learning about the noise signal.

The noisy text (microphone) can be collected and used later for learning to filter out the noisy text. The noise text (microphone) is learned based on the learning of the noisy text (microphone), and may be used later to separate the noise text and the analysis target text from the candidate text information.

The embodiments according to the present invention described above may be implemented in the form of program instructions that can be executed through various computer components and recorded in a computer-readable recording medium. The computer-readable recording medium may include program instructions, data files, data structures, etc. alone or in combination. The program instructions recorded in the computer-readable recording medium may be specially designed and configured for the present invention or may be known and usable to those skilled in the computer software field. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical recording media such as CD-ROMs and DVDs, and magnetic-optical media such as floptical disks. medium), and a hardware device specially configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of the program instructions include not only machine language codes such as those produced by a compiler but also high-level language codes that can be executed by a computer using an interpreter or the like. The hardware device can be changed to one or more software modules to perform the processing according to the present invention, and vice versa.

In the above, the present invention has been described by specific matters such as specific elements and limited embodiments and drawings, but this is provided only to help a more general understanding of the present invention, and the present invention is not limited to the above embodiments. Anyone with ordinary knowledge in the technical field to which the invention belongs can make various modifications and changes from these descriptions.

Accordingly, the spirit of the present invention is limited to the above-described embodiments and should not be defined, and all ranges equivalent to or equivalently changed from the claims to be described later as well as the claims to be described later are the scope of the spirit of the present invention. It will be said to belong to.

Claims (6)

The control method of the bus information providing terminal is:
Receiving, by the bus information providing terminal, a sound signal;
Determining, by the bus information providing terminal, text information to be analyzed based on the sound signal; And
Including the step of providing, by the bus information providing terminal, information on a target bus corresponding to the text information to be analyzed,
The analysis target text information is determined based on a trigger word on candidate text information extracted based on the bus information terminal,
The analysis target text information is divided into a main analysis text and a sub analysis text based on the trigger word,
The bus information providing terminal determines the target bus information by preferentially analyzing the main analysis text,
The main analysis text is set differently according to the trigger word,
The main analysis text is n texts whose probability of including valid information based on the trigger word is greater than or equal to a first threshold value,
When the reliability of the first bus information determined based on the main analysis text is less than a threshold, the second text further includes a sub-analysis text having a relatively high priority among the first and second sub-analysis texts. Determine the bus information,
And determining the target bus based on whether the first bus information and the second bus information are the same. The method of claim 1,
The first sub-analysis text is x texts that exist in front of the main analysis text in time,
The second sub-analysis text is y texts that exist temporally behind the main analysis text,
The size of the x and y is a value that is adaptively changed according to the trigger word, and the probability of including valid information based on the existing trigger word is determined to be a value equal to or greater than a second threshold. The method of claim 2,
When the first bus information and the second bus information are not the same, the third bus information is additionally included in the first sub-analysis text and the second sub-analysis text, which has a lower priority. Extracted,
When bus information corresponding to a majority of the first bus information, the second bus information, and the third bus information exists, a bus corresponding to the bus information corresponding to the majority is set as the target bus,
Whether to extract the first sub-analysis text or the second sub-analysis text is determined according to a reliability level of a starting point and a ending point of a speech based on image analysis. In the bus information providing terminal,
A voice input unit implemented to receive a sound signal; And
A processor operatively connected to the voice input unit,
The processor determines text information to be analyzed based on the sound signal,
It is implemented to provide information on a target bus corresponding to the text information to be analyzed,
The analysis target text information is determined based on a trigger word on candidate text information extracted based on the bus information terminal,
The analysis target text information is divided into a main analysis text and a sub analysis text based on the trigger word,
The bus information providing terminal determines the target bus information by preferentially analyzing the main analysis text,
The main analysis text is set differently according to the trigger word,
The main analysis text is n texts whose probability of including valid information based on the trigger word is greater than or equal to a first threshold value,
When the reliability of the first bus information determined based on the main analysis text is less than a threshold, the second text further includes a sub-analysis text having a relatively high priority among the first and second sub-analysis texts. Determine the bus information,
And determining the target bus based on whether the first bus information and the second bus information are the same. The method of claim 4,
The first sub-analysis text is x texts that exist in front of the main analysis text in time,
The second sub-analysis text is y texts that exist temporally behind the main analysis text,
The magnitudes of x and y are values that are adaptively changed according to the trigger word, and the probability of including valid information based on an existing trigger word is determined to be a value equal to or greater than a second threshold. Provided terminal. The method of claim 5,
When the first bus information and the second bus information are not the same, the third bus information is additionally included in the first sub-analysis text and the second sub-analysis text, which has a lower priority. Extracted,
When bus information corresponding to a majority of the first bus information, the second bus information, and the third bus information exists, a bus corresponding to the bus information corresponding to the majority is set as the target bus,
The bus information providing terminal, wherein whether or not the first sub-analysis text and the second sub-analysis text are to be extracted is determined according to a reliability level of an utterance start time and an utterance end time based on image analysis.

Images