Speech recognition system for power industry

Abstract

The invention provides a voice recognition system for the power industry, and the system comprises a voice collection model which is used for carrying out the collection of voice data in a designated region; the voice processing module is used for converting the voice data into a general text for the power industry according to a preset voice dictionary in the database and outputting the general text to the voice recognition module; the voice recognition module is used for carrying out semantic recognition on the general text for the power industry according to input voice key information, recognizing general power industry semantics in the text, carrying out keyword recognition judgment and outputting a recognition result; and the judgment module is used for judging the recognition result according to the input control designation and determining the semantic recognition accuracy according to the judgment result. According to the method, on-line or off-line voice recognition can be directly carried out, voice-to-text conversion can be carried out on different electric power problems and solutions for storage and recording, meanwhile, data processing is carried out on recorded texts, and text key semantics are extracted to establish a model.

Description

Technical Field

[0001] The present invention relates to the field of electric power technology, and in particular to a speech recognition system used in the electric power industry. Background Art

[0002] With the continuous improvement of my country's power infrastructure, the power network is composed of substations and transmission lines of different voltage levels. After a long period of operation, the power network often needs to be repaired, which requires grassroots power maintenance personnel to perform fault maintenance. However, the records of fault maintenance are mostly simple shorthand text, which is not convenient for organization and search, and is not conducive to rapid processing of the same power problem. Summary of the invention

[0003] The purpose of the present invention is to propose a speech recognition system for the power industry to solve the technical problem of how to efficiently and quickly record different power problems to improve the efficiency of fault information processing.

[0004] On the one hand, a speech recognition system for the electric power industry is provided, comprising:

[0005] Voice collection model, used to collect voice data in a specified area;

[0006] A speech processing module, used to convert speech data into common text for the power industry according to a preset speech dictionary in the database, and output it to the speech recognition module;

[0007] A speech recognition module, used to perform semantic recognition on the common text for the electric power industry according to the input speech key information, recognize the common electric power industry semantics in the text, perform keyword recognition and determination, and output the recognition result;

[0008] The determination module is used to determine the recognition result according to the input control specification, and determine the semantic recognition accuracy according to the determination result.

[0009] Preferably, it also includes a learning module, which is used to identify the correct semantic recognition through the determination module based on a preset neural network learning, and record it for automatic semantic determination next time, and divide different industry problems through learning;

[0010] The model generation module is used for semantic recognition of completed judgments, establishing parent-child models according to different levels based on preset power industry requirements, and storing them in the database for rapid extraction through semantic recognition and graded processing according to different issues.

[0011] Preferably, it also includes a database for storing various collected and generated data;

[0012] Image acquisition module, used to record video of the power maintenance solution process.

[0013] Preferably, the voice processing module includes at least one processor, and the processor is used to process the received voice data to obtain corresponding general text for the electric power industry.

[0014] Preferably, it also includes an interactive front end for receiving and displaying the recognition results, and judging whether each item in the recognition results meets the preset standards; if an item meets the standards, it is judged to be appropriate; if an item does not meet the standards, it is judged to be inappropriate.

[0015] Preferably, the interactive front end is also used to input corresponding question information when an item is determined to be inappropriate, and output the question information to the voice processing module.

[0016] Preferably, the voice processing module is also used to replace the question information with corresponding question text information according to a preset voice dictionary in the database, and store the question text information in the database as relevant maintenance record or guidance information.

[0017] In summary, the implementation of the embodiments of the present invention has the following beneficial effects:

[0018] The speech recognition system for the electric power industry provided by the present invention can directly perform speech recognition online or offline, can convert speech into text for storage and recording of different electric power problems and solutions, and can perform data processing on the recorded text to extract key semantics of the text and establish a model. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, without paying creative labor, other drawings obtained based on these drawings still belong to the scope of the present invention.

[0020] Figure 1 Schematic diagram of a speech recognition system for the power industry according to an embodiment of the present invention. DETAILED DESCRIPTION

[0021] In order to make the objectives, technical solutions and advantages of the present invention more clear, the present invention will be further described in detail below with reference to the accompanying drawings.

[0022] like Figure 1 FIG. 1 is a schematic diagram of an embodiment of a speech recognition system for the power industry provided by the present invention. In this embodiment, the system includes:

[0023] The voice collection model is used to collect voice data in the designated area; the voice processing module is used to convert the voice data into a general text for the power industry according to the preset voice dictionary in the database, and output it to the voice recognition module; the voice recognition module is used to perform semantic recognition on the general text for the power industry according to the input voice key information, recognize the general power industry semantics in the text, and perform keyword recognition judgment, and output the recognition result; the judgment module is used to judge the recognition result according to the input control designation, and determine the semantic recognition accuracy according to the judgment result. It can be understood that this system uses a microphone for voice input; the voice processing module processes the input voice and converts the voice input into a general text for the power industry according to the database voice dictionary; the voice recognition module automatically recognizes the input voice key information, performs semantic recognition, recognizes the general power industry semantics in the text, and performs keyword recognition judgment; the judgment module can be connected to the front end for voice recognition confirmation to confirm the correctness of the system semantic recognition. The judgment module is connected to the front end to make the front end pop up a window and perform recognition judgment.

[0024] In one embodiment, the system further includes a learning module for learning the correct semantic recognition determined by the judgment module based on a preset neural network, and recording it for the next automatic semantic judgment, and dividing different industry problems through learning; a model generation module for the semantic recognition that has been completed, and establishing a parent-child model according to different levels according to the preset power industry requirements, and storing it in a database, so as to be quickly extracted through semantic recognition, and to be graded according to different problems. It can be understood that the automatic learning module has a built-in learning chip based on a neural network artificial intelligence (AI) chip, which has the ability of autonomous learning and reasoning. The automatic learning module automatically learns the correct semantic recognition determined by the judgment module, and records it, so as to facilitate the next automatic semantic judgment, and to divide different industry problems through learning; a model generation module is set, and the semantic recognition that has been completed is established according to different levels according to the power industry requirements, and stored in a database, so as to be quickly extracted through semantic recognition, and graded according to different problems, such as line problems are divided into short circuits, open circuits, etc.

[0025] In one embodiment, the system further includes a database for storing various data collected and generated; an image acquisition module for recording the video of the power maintenance solution process. The database records the voice dictionary and assists in carrying the semantic model; and the camera can simultaneously record the video of the power maintenance solution process.

[0026] In one embodiment, the voice processing module includes at least one processor, which is used to process the received voice data to obtain corresponding general text for the power industry. It can be understood that the voice processing module includes a circuit board and a processing chip, the circuit board is connected to a microphone to access the voice, and the processing chip processes the voice and converts the voice into text.

[0027] In one embodiment, the system further includes an interactive front end for receiving and displaying the recognition results, and determining whether each recognition in the recognition results meets a preset standard; if an item meets the standard, the item is determined to be appropriate, and if an item does not meet the standard, the item is determined to be inappropriate. The interactive front end is also used to input corresponding question information when an item is determined to be inappropriate, and output the question information to the speech processing module. It is understandable that after the semantic recognition is completed, feedback is provided through the front end to confirm whether the recognition is correct. When the recognition is inappropriate, multiple semantic options appear and the correct selection is made. That is, when the maintenance staff completes a maintenance work, they directly input voice offline or online through the microphone, and the voice processing module converts the input voice into common text for the power industry according to the database voice dictionary. The voice recognition module automatically recognizes the key information of the input voice and performs semantic recognition. When the semantic recognition is completed, feedback is given through the front end to confirm whether the recognition is correct. The semantics of the completed voice are learned through the learning module, and the voice-related information is stored in the database. A camera can be set to synchronously record the maintenance work video. Through the model generation module, the semantic recognition of the completed judgment is carried out, and the parent-child model is established according to different levels according to the requirements of the power industry, and stored in the database for quick retrieval.

[0028] In one embodiment, the voice processing module is also used to replace the problem information with corresponding problem text information according to the preset voice dictionary in the database, and store the problem text information as relevant maintenance record or guidance information in the database. It is understandable that when the maintenance staff needs to perform maintenance, according to the problem found, the problem is input through the microphone, and the voice processing module and the voice recognition module perform semantic extraction, and after semantic confirmation, the relevant maintenance record or guidance in the database model is retrieved to assist the maintenance work.

[0029] In summary, the implementation of the embodiments of the present invention has the following beneficial effects:

[0030] The speech recognition system for the electric power industry provided by the present invention can directly perform speech recognition online or offline, can convert speech into text for storage and recording of different electric power problems and solutions, and can perform data processing on the recorded text to extract key semantics of the text and establish a model.

[0031] The above disclosure is only the preferred embodiment of the present invention, which certainly cannot be used to limit the scope of the present invention. Therefore, equivalent changes made according to the claims of the present invention are still within the scope of the present invention.

Claims

1. A speech recognition system for the power industry, characterized in that: include: Voice collection model, used to collect voice data in a specified area; A speech processing module, used to convert speech data into common text for the power industry according to a preset speech dictionary in the database, and output it to the speech recognition module; A speech recognition module, used to perform semantic recognition on the common text for the electric power industry according to the input speech key information, recognize the common electric power industry semantics in the text, perform keyword recognition and determination, and output the recognition result; The determination module is used to determine the recognition result according to the input control specification, and determine the semantic recognition accuracy according to the determination result.

2. The system according to claim 1, characterized in that Also includes, The learning module is used to identify the correct semantic recognition through the judgment module based on the preset neural network learning, and record it for the next automatic semantic judgment, and divide different industry problems through learning; The model generation module is used for semantic recognition of completed judgments, establishing parent-child models according to different levels based on preset power industry requirements, and storing them in the database for rapid extraction through semantic recognition and graded processing according to different issues.

3. The system according to claim 2, characterized in that Also includes, Database, used to store various collected and generated data; Image acquisition module, used to record video of the power maintenance solution process.

4. The system according to claim 3, characterized in that The voice processing module includes at least one processor, and the processor is used to process the received voice data to obtain corresponding general text for the electric power industry.

5. The system according to claim 3, characterized in that Also includes, The interactive front end is used to receive and display the recognition results, and determine whether each recognition in the recognition results meets the preset standards; If an item meets the requirements, it is judged to be appropriate; if an item does not meet the requirements, it is judged to be inappropriate.

6. The system according to claim 5, characterized in that The interactive front end is also used to input corresponding question information when a certain item is determined to be inappropriate, and output the question information to the voice processing module.

7. The system according to claim 6, characterized in that The speech processing module is also used to replace the question information with corresponding question text information according to the speech dictionary preset in the database, and store the question text information in the database as relevant maintenance record or guidance information.

Images