Intelligent semantic analysis and instruction analysis method and system

Abstract

The invention discloses an intelligent semantic analysis and instruction analysis method and system, and the method comprises the steps: enabling a plurality of clients to be connected with an RTC, and subscribing a first audio and video channel; the plurality of clients are respectively connected with the business server and are used for realizing real-time data interaction and business logic processing; audio and video screen stream data are obtained through the RTC, and audio signals in the audio and video data are converted into text information; obtaining the text information and context information of the text information, and performing instruction identification on the text information; and calling a corresponding service interface based on the identified instruction to complete the operation or task of the instruction. According to the method, the accuracy, the real-time performance and the flexibility of voice interaction can be remarkably improved, the application space of the intelligent voice technology in a multi-person cooperation scene is expanded, and an innovative interaction mode is brought to the vertical fields of intelligent conferences, online education and the like.

Description

Technical Field

[0001] The present invention relates to the field of audio and video technology, and in particular to an intelligent semantic analysis and instruction parsing method and system. Background Art

[0002] In current audio and video conferencing and other scenarios such as online education, users usually need to interact with the system through voice to achieve operations such as information query and content control. Traditional voice interaction methods mainly rely on predefined keyword matching or simple grammatical analysis, which makes it difficult to accurately understand the user's true intentions, and have poor flexibility and scalability. There are also difficulties in role recognition in multi-person voice interaction and insufficient flexibility in cross-scenario command migration.

[0003] Currently common voice interaction systems, such as smart speakers and voice assistants, can achieve a certain degree of semantic understanding and task execution, but still have the following shortcomings: the accuracy and real-time performance of offline speech recognition cannot meet the needs of streaming processing; the semantic understanding model lacks context awareness and multi-round interaction capabilities, and it is difficult to process complex instructions; the instruction mapping rules rely on domain knowledge, have poor scalability, and cannot flexibly adapt to new application scenarios; there is a lack of instruction parsing and permission management mechanisms for multi-person collaboration.

[0004] Therefore, there is an urgent need for a new intelligent semantic analysis and command parsing method that can achieve streaming semantic understanding for audio and video scenarios and support multi-round interactions and dynamic command expansion in the form of natural language to enhance user experience and application scope. Summary of the invention

[0005] In response to the above technical problems, the present invention provides an intelligent semantic analysis and instruction parsing method and system to solve the shortcomings of existing voice interaction systems in terms of multi-person collaboration, streaming processing, natural language understanding, etc.

[0006] The first embodiment of the present invention provides an intelligent semantic analysis and instruction parsing method, which includes: multiple clients are respectively connected to RTC and all subscribe to a first audio and video channel; multiple clients are respectively connected to a business server to realize real-time data interaction and business logic processing; audio and video stream data is obtained through RTC, and the audio signal in the audio and video data is converted into text information; the text information and the context information of the text information are obtained, and instructions are recognized for the text information; based on the recognized instructions, the corresponding service interface is called to complete the operation or task of the instruction.

[0007] Optionally, the step of converting the audio signal in the audio and video data into text information includes: converting the audio signal into incremental text information in real time, and forming structured meeting record data for storage through semantic analysis and information extraction.

[0008] Optionally, the step of converting the audio signal in the audio and video data into text information includes: using streaming segmentation recognition, the recognition channel inside the speech recognition module will perform parameter analysis on the received audio signal, and dynamically segment the audio signal based on the analyzed parameters, and the segmentation granularity can be adjusted according to voice activity detection or a fixed time window.

[0009] Optionally, the step of obtaining the text information and the context information of the text information and performing instruction recognition on the text information includes: reading the incremental text information from a stored file in real time, and performing word segmentation on the incremental text information, dividing continuous character strings into independent vocabulary units, and matching with keywords to analyze the contextual semantics of the incremental text information, and identifying the semantic structure and expression intention of the text information; matching the identified text information with a predefined instruction library to determine whether the current text information contains executable instructions and instruction parameters.

[0010] Optionally, the client includes a first client and a second client, and the step of performing instruction recognition on the text information includes: recognizing that the text information comes from the first client or the second client; calling a corresponding service interface according to the recognized instruction and instruction parameters to complete the operation or task of the instruction.

[0011] The second embodiment of the present invention provides an intelligent semantic analysis and instruction parsing system, which includes: multiple clients, each of which is connected to the RTC and subscribes to a first audio and video channel, and each of which establishes a connection with a business server to realize real-time data interaction and business logic processing; a voice monitoring module, which is used to connect to the first audio channel and obtain audio and video stream data through the RTC; a voice recognition module, which is used to receive the audio and video stream data transmitted by the voice monitoring module, and is used to convert the audio signal in the audio and video data into text information; an intelligent parsing module, which is used to obtain the text information and the context information of the text information, and perform instruction recognition on the text information; an instruction execution module, which is used to call the corresponding service interface based on the recognized instruction to complete the operation or task of the instruction.

[0012] Optionally, the speech recognition module is also used to convert the audio signal in the audio and video data into incremental text information in real time, and form structured meeting record data for storage through semantic analysis and information extraction.

[0013] Optionally, the speech recognition module adopts streaming segmentation recognition, and the recognition channel inside the speech recognition module will perform parameter analysis on the received audio signal and dynamically segment the audio signal based on the analyzed parameters. The segmentation granularity can be adjusted according to voice activity detection or a fixed time window.

[0014] Optionally, the intelligent parsing module is also used to read the incremental text information from the stored file in real time, and perform word segmentation on the incremental text information, dividing continuous character strings into independent vocabulary units, and matching with keywords to analyze the contextual semantics of the incremental text information, and identifying the semantic structure and expression intention of the text information; the instruction recognition module is used to match the recognized text information with a predefined instruction library to determine whether the current text information contains executable instructions and instruction parameters.

[0015] Optionally, the client includes a first client and a second client, and the speech recognition module is used to identify that the text information comes from the first client or the second client; the instruction execution module is used to call the corresponding service interface according to the recognized instruction and the parameters of the instruction to complete the operation or task of the instruction.

[0016] In the technical solution provided by the embodiment of the present invention, the real-time voice in the audio and video conference or its public channel is converted into text form, and the text information is recognized by command, and the corresponding service interface is called based on the recognized command to complete the operation or task of the command. Compared with the existing technology, the present invention can significantly improve the accuracy, real-time performance and flexibility of voice interaction, expand the application space of intelligent voice technology in multi-person collaborative scenarios, and bring innovative interaction modes to vertical fields such as smart meetings and online education. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 The figure is a flow chart of an embodiment of an intelligent semantic analysis and instruction parsing method of the present invention.

[0018] Figure 2 The figure is a flow chart of another embodiment of an intelligent semantic analysis and instruction parsing method of the present invention.

[0019] Figure 3 The figure is a flow chart of converting an audio signal into text information according to the present invention.

[0020] Figure 4 The figure is a flow chart of the present invention for performing instruction recognition on text information. DETAILED DESCRIPTION

[0021] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0022] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0023] In the description of the present invention, it should be noted that, unless otherwise clearly specified and limited, the terms "connected" and "connection" should be understood in a broad sense, for example, it can be directly connected, or indirectly connected through an intermediate medium, or it can be the internal connection of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances. In addition, the technical features involved in the different embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0024] Please refer to Figure 1 , Figure 2 As shown, the present invention provides an intelligent semantic analysis and instruction parsing method, the method comprising the following steps:

[0025] Step S10: Multiple clients are respectively connected to the RTC and subscribe to the first audio and video channel.

[0026] First, the client accesses the audio and video stream acquisition module. The multiple clients include a first client and a second client. For example, the first client is client A and the second client is client B. When the system starts, client A and client B will establish connections with RTC respectively and subscribe to the same audio and video channel, such as the first video channel.

[0027] In one embodiment of the present invention, the client may also be a mobile APP or mini-program, which can access the semantic analysis and instruction execution service of the present invention through built-in voice collection and network communication capabilities.

[0028] Step S20, multiple clients establish connections with the business server respectively to achieve real-time data interaction and business logic processing.

[0029] Step S30, obtaining audio and video stream data through RTC, and converting the audio signal in the audio and video data into text information.

[0030] In one embodiment of the present invention, the audio signal is converted into incremental text information in real time, and structured conference record data is formed and stored through semantic analysis and information extraction.

[0031] After the voice monitoring module obtains the audio and video stream data, it will convert the audio signal into text information in text format through technologies such as acoustic models and language models. In addition, the text information is generated in an incremental manner, that is, each small audio segment corresponds to a piece of text, rather than waiting for the entire audio to end before outputting the complete text. These incremental text information will be persistently stored in the form of files for use in subsequent semantic understanding and instruction parsing.

[0032] The server will start a voice monitoring module and connect to the same audio channel as Client A and Client B. After the voice monitoring module obtains the real-time audio and video data of the channel, it will Figure 2 The process shown handles this.

[0033] Please refer to Figure 3 As shown in the figure, by subscribing to the RTC pipeline, the audio and video mixed stream in the current channel or conference is obtained in real time. RTC adopts a real-time communication protocol and supports multi-terminal real-time audio and video transmission. The audio and video streams obtained from the pipeline will be forwarded to the recognition channel inside the speech recognition module. The recognition channel adopts a streaming processing framework to support real-time distribution and concurrent recognition of audio and video streams. The recognition channel will perform parameter analysis on the received audio, such as sampling rate, frame rate, encoding format, etc., to prepare for subsequent segmented recognition. By analyzing the audio features, the recognition channel can determine the identity of the speaker (such as client A or client B) and achieve speaker separation.

[0034] The present invention adopts streaming segmentation recognition, and the recognition channel inside the speech recognition module dynamically segments the audio signal based on the parsed parameters. The segmentation granularity can be adjusted according to voice activity detection (VAD) or a fixed time window, which not only ensures real-time performance but also takes into account the integrity of recognition. Each segmented audio clip will be sent to the ASR engine for concurrent recognition.

[0035] The ASR engine will perform speech analysis and automatic correction on the recognized text information. Semantic analysis is mainly combined with language models to correct obvious grammatical errors and eliminate some meaningless spoken noise. At the same time, semantic analysis can also standardize some professional vocabulary to improve the standardization of the text. The recognition results will be spliced ​​according to the chronological order and the identity of the speaker to form a complete text record. For example: "Client A: What is the topic today? Client B: Our topic today is to talk about the impact of AI technology on our daily life."

[0036] The final text information, including formatted text and extracted semantic information, will be written to the file storage in an incremental manner. Incremental storage can improve writing efficiency on the one hand, and facilitate real-time display and synchronization on the other hand.

[0037] The present invention adopts a streaming processing framework, obtains multi-person voice signals in real time and transcribes them in segments by subscribing to conference voice channels, which not only ensures the real-time recognition but also improves the processing efficiency of long voices.

[0038] Through the above process, the voice information in the audio and video stream is converted into text information in real time, and through semantic analysis and information extraction, structured meeting record data is formed. These data can provide a variety of applications for smart conferences, such as real-time subtitles, meeting summaries, key decision extraction, intelligent retrieval, etc., greatly improving the information level and management efficiency of the conference.

[0039] At the same time, the speech recognition adopts optimization mechanisms such as streaming concurrent processing, semantic feedback error correction, and incremental storage. While ensuring the real-time recognition, it also improves the recognition accuracy and availability, laying a solid data foundation for intelligent semantic analysis.

[0040] Step S40: acquiring the text information and context information of the text information, and performing instruction recognition on the text information.

[0041] The present invention obtains the context information of the current text information to understand the context and the speaker's intention. Then, text mining and information extraction such as summary, keyword extraction, content classification, etc. in the text information are performed according to business requirements. At the same time, the text information is subjected to instruction recognition based on the large model. By matching with the predefined instruction library, it is determined whether the current text information contains executable instructions and the specific parameters of the instructions. When parsing the instructions, it can distinguish whether the text comes from client A or client B, and when the subsequent instructions are distributed, it can identify which specific client.

[0042] The semantic understanding module of the present invention makes full use of context clues and uses a pre-trained large language model to analyze semantics from multiple levels such as vocabulary, syntax, and pragmatics to extract elements such as intent, slot, and emotion. Different from a single-round question-and-answer interaction, this solution can build a complete context through multiple rounds of accumulation, thereby accurately grasping the speaker's true intent.

[0043] In one embodiment of the present invention, step S20 specifically includes the following steps:

[0044] The incremental text information is read from the stored file in real time, and the incremental text information is segmented to divide the continuous character string into independent vocabulary units, and the contextual semantics of the incremental text information are analyzed by matching with keywords to identify the semantic structure and expression intention of the text information;

[0045] The recognized text information is matched with the predefined instruction library to determine whether the current text information contains executable instructions and instruction parameters.

[0046] For details, please refer to Figure 4 As shown, the incremental text generated by speech recognition is read from the stored file in real time. Then, the text information is segmented to divide the continuous character string into independent vocabulary units. The segmentation method is based on a combination of dictionaries and statistical models, and the segmented text is matched with predefined system keywords. The keyword library contains professional vocabulary, terminology, entity names, etc. related to specific fields or application scenarios. Through matching, important concepts and entities appearing in the text can be identified to provide support for subsequent semantic understanding. At the same time, the results of keyword matching can also optimize the effect of word segmentation in turn, such as merging matched keyword combinations and correcting misclassified vocabulary.

[0047] Based on keyword matching, the intelligent parsing module further analyzes the contextual semantics of the text. By considering the dependencies and semantic roles between words, combined with domain knowledge graphs and semantic rules, the semantic structure and expression intent of the text can be identified. Common semantic recognition tasks include: intent recognition, slot filling, semantic dependency analysis, etc. Contextual semantic recognition can help understand the speaker's true intention and provide a basis for intelligent question answering and instruction parsing.

[0048] Instruction matching and parameter parsing identify semantic information and match it with the predefined system. The instruction system adopts a structured form and defines a series of standardized operation instructions and corresponding parameter templates. By matching the text semantics with the instruction template, the operation intention implied in the text can be identified and the relevant parameter information can be extracted. The instruction matching process can use a variety of technical means such as keyword matching, regular expressions, semantic rules, etc. to improve the flexibility and accuracy of matching.

[0049] In actual applications, the user's expression may be ambiguous or incomplete, resulting in multiple commands being matched or missing necessary parameters. In this case, command ambiguity resolution is required. Common resolution strategies include: excluding irrelevant commands based on contextual information, inferring default parameters based on user intent and interaction history, and guiding users to supplement information through clarifying questions and answers. Command ambiguity resolution can improve the accuracy of command recognition, and can also enhance the system's fault tolerance and the friendliness of human-computer interaction.

[0050] Step S50: calling a corresponding service interface based on the identified instruction to complete the operation or task of the instruction.

[0051] Identify that the text information originates from the first client or the second client;

[0052] The corresponding service interface is called according to the identified instruction and the parameters of the instruction to complete the operation or task of the instruction.

[0053] The present invention is aimed at audio and video interactions involving multiple people. It needs to distinguish the identities of different speakers from the voice stream, and based on semantic analysis, accurately judge and distribute instructions according to business rules and user permissions, coordinate the task execution of multiple parties, and finally give feedback to the initiator to achieve a complete collaborative closed loop.

[0054] After matching and resolution, clear instructions and parameter information are obtained. The system will route the instructions to the corresponding processing module or service interface according to the type and business logic of the instructions. At the same time, if the instructions involve collaboration between multiple clients or users, the system also needs to distribute and synchronize the instructions. Instruction distribution needs to consider factors such as the client's identity, permissions, and status to ensure that the instructions are accurately delivered to the target recipient.

[0055] The command execution and result feedback will eventually be distributed, triggering the corresponding business logic and operations. The specific execution process may involve multiple links such as data query, status update, and external service call. The execution result will be returned to the command distribution module in an asynchronous manner, and then it will be uniformly managed and notified of the result. For instructions involving multi-party collaboration, the system also needs to coordinate the execution progress of all parties to ensure the final consistency. The execution results will eventually be fed back to the user, and a closed loop of human-computer interaction will be realized through speech synthesis, message push, interface update, etc.

[0056] Through the collaborative work of various modules and the interaction of data streams, a complete closed loop from online speech recognition and understanding to execution feedback is achieved, providing an intelligent interactive experience.

[0057] The present invention converts voice into structured command information by subscribing to real-time voice streams in scenarios such as audio and video conferences or live teaching, and uses technologies such as voice recognition, semantic understanding, and command mapping, and performs authority judgment and task distribution according to business rules, ultimately achieving intelligent voice interaction in a multi-person collaborative environment to solve the deficiencies of existing voice interaction systems in multi-person collaboration, streaming processing, and natural language understanding. The present invention can significantly improve the accuracy, real-time performance, and flexibility of voice interaction, expand the application space of intelligent voice technology in multi-person collaborative scenarios, and bring innovative interaction modes to vertical fields such as smart conferences and online education.

[0058] The present invention also provides an intelligent semantic analysis and instruction parsing system, which includes multiple clients, a voice monitoring module, a voice recognition module, an intelligent parsing module and an instruction execution module. The multiple clients are respectively connected to the RTC and subscribe to the first audio and video channel. The multiple clients are respectively connected to the business server to realize real-time data interaction and business logic processing; the voice monitoring module is used to connect to the first audio channel and obtain audio and video stream data through the RTC; the voice recognition module is used to receive the audio and video stream data transmitted by the voice monitoring module, and is used to convert the audio signal in the audio and video data into text information; the intelligent parsing module is used to obtain the text information and the context information of the text information, and perform instruction recognition on the text information; the instruction execution module is used to call the corresponding service interface based on the recognized instruction to complete the operation or task of the instruction.

[0059] In one embodiment of the present invention, the speech recognition module is also used to convert the audio signal in the audio and video data into incremental text information in real time, and form structured meeting record data for storage through semantic analysis and information extraction.

[0060] In one embodiment of the present invention, the speech recognition module adopts streaming segmentation recognition, and the recognition channel in the speech recognition module dynamically segments the audio signal based on the parsed parameters, and the segmentation granularity can be adjusted according to voice activity detection or a fixed time window.

[0061] In one embodiment of the present invention, the intelligent parsing module is further used to read the incremental text information from the stored file in real time, and perform word segmentation processing on the incremental text information, dividing the continuous character string into independent vocabulary units, and matching and analyzing the context semantics of the incremental text information with keywords, and identifying the semantic structure and expression intention of the text information;

[0062] The instruction recognition module is used to match the recognized text information with a predefined instruction library to determine whether the current text information contains executable instructions and instruction parameters.

[0063] In one embodiment of the present invention, the client includes a first client and a second client, and the speech recognition module is used to identify whether the text information originates from the first client or the second client; the instruction execution module is used to call the corresponding service interface according to the recognized instruction and the parameters of the instruction to complete the operation or task of the instruction.

[0064] The intelligent semantic analysis and instruction parsing system described in the present invention is used to execute the intelligent semantic analysis and instruction parsing method described in any of the above embodiments.

[0065] The intelligent semantic analysis and instruction analysis technology of the present invention can convert real-time voice in audio and video conferences or their public channels into text form, and perform semantic understanding and instruction parsing, so that participants can realize intelligent operations such as information retrieval and content control through natural language. At the same time, the client-to-client data flow design enables the recognition results to be fed back to the client in real time, improving the interactivity and real-time performance of the system. In addition, semantic analysis based on contextual information can understand the speaker's intention and improve the accuracy of instruction recognition. The predefined extensible instruction set enables the system to have personalized customization capabilities and can be flexibly adapted to different business scenarios. This process comprehensively uses technologies in multiple fields such as natural language processing, knowledge engineering, and business logic to achieve the conversion from unstructured text to structured instructions, so that the system can automatically understand the user's intentions and perform corresponding operations.

[0066] It should be understood that the "one embodiment" mentioned in the specification means that the specific features, structures or characteristics related to the embodiment are included in at least one embodiment of the present application. Therefore, the "one embodiment" appearing in various places throughout the specification does not necessarily refer to the same embodiment. In addition, these specific features, structures or characteristics can be combined in one or more embodiments in any suitable manner. It should be understood that in various embodiments of the present application, the size of the sequence number of the above-mentioned processes does not mean the order of execution, and the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present application. The above-mentioned sequence numbers of the embodiments of the present application are only for description and do not represent the advantages and disadvantages of the embodiments.

[0067] The above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An intelligent semantic analysis and instruction parsing method, characterized in that: The method comprises: Multiple clients are connected to the RTC respectively and subscribe to the first audio and video channel; Multiple clients establish connections with the business server respectively to achieve real-time data interaction and business logic processing; Get audio and video stream data through RTC, and convert the audio signal in the audio and video data into text information; Acquire the text information and context information of the text information, and perform instruction recognition on the text information; Based on the identified instruction, the corresponding service interface is called to complete the operation or task of the instruction.

2. The intelligent semantic analysis and instruction parsing method according to claim 1, characterized in that: The step of converting the audio signal in the audio and video data into text information comprises: The audio signal is converted into incremental text information in real time, and structured meeting record data is formed and stored through semantic analysis and information extraction.

3. The intelligent semantic analysis and instruction parsing method according to claim 1, characterized in that: The step of converting the audio signal in the audio and video data into text information comprises: Using streaming segmented recognition, the recognition channel inside the speech recognition module will perform parameter analysis on the received audio signal and dynamically segment the audio signal based on the analyzed parameters. The segmentation granularity can be adjusted according to voice activity detection or a fixed time window.

4. The intelligent semantic analysis and instruction parsing method according to claim 2, characterized in that: The step of obtaining the text information and the context information of the text information and performing instruction recognition on the text information includes: The incremental text information is read from the stored file in real time, and the incremental text information is segmented to divide the continuous character string into independent vocabulary units, and the contextual semantics of the incremental text information are analyzed by matching with keywords to identify the semantic structure and expression intention of the text information; The recognized text information is matched with the predefined instruction library to determine whether the current text information contains executable instructions and instruction parameters.

5. The intelligent semantic analysis and instruction parsing method according to claim 4, characterized in that: The client includes a first client and a second client, and the step of performing instruction recognition on the text information includes: Identify that the text information originates from the first client or the second client; The corresponding service interface is called according to the identified instruction and the parameters of the instruction to complete the operation or task of the instruction.

6. An intelligent semantic analysis and instruction parsing system, characterized in that: The system comprises: Multiple clients are respectively connected to the RTC and subscribe to the first audio and video channel. Multiple clients are respectively connected to the business server to realize real-time data interaction and business logic processing; A voice monitoring module, used to connect to the first audio channel and obtain audio and video stream data through RTC; A speech recognition module is used to receive the audio and video stream data transmitted by the speech monitoring module, and to convert the audio signal in the audio and video data into text information; An intelligent parsing module, used to obtain the text information and context information of the text information, and perform instruction recognition on the text information; The instruction execution module is used to call the corresponding service interface based on the identified instruction to complete the operation or task of the instruction.

7. The intelligent semantic analysis and instruction parsing system according to claim 6, characterized in that: The speech recognition module is also used to convert the audio signal in the audio and video data into incremental text information in real time, and form structured meeting record data for storage through semantic analysis and information extraction.

8. The intelligent semantic analysis and instruction parsing system according to claim 6, characterized in that: Using streaming segmented recognition, the recognition channel inside the speech recognition module will perform parameter analysis on the received audio signal and dynamically segment the audio signal based on the analyzed parameters. The segmentation granularity can be adjusted according to voice activity detection or a fixed time window.

9. The intelligent semantic analysis and instruction parsing system according to claim 7, characterized in that: The intelligent parsing module is also used to read the incremental text information from the stored file in real time, and perform word segmentation on the incremental text information, dividing the continuous character string into independent vocabulary units, and matching and analyzing the contextual semantics of the incremental text information with keywords, and identifying the semantic structure and expression intention of the text information; The instruction recognition module is used to match the recognized text information with a predefined instruction library to determine whether the current text information contains executable instructions and instruction parameters.

10. The intelligent semantic analysis and instruction parsing system according to claim 9, characterized in that: The client includes a first client and a second client, the speech recognition module is used to identify whether the text information comes from the first client or the second client; the instruction execution module is used to call the corresponding service interface according to the recognized instruction and the parameters of the instruction to complete the operation or task of the instruction.

Images