Written inspection information entry data processing system and method based on speech recognition

Abstract

The present invention relates to the technical field of inspection information data processing, and discloses a written inspection information entry data processing system and method based on voice recognition, wherein the system comprises an inspection information entry module, an inspection information modification module, and an inspection information modification execution module; based on scientifically preset audio parameters of different dialect accent types combined with an intelligent audio matching algorithm and audio parameters of staff accents, accurate analysis of staff accent types is performed to realize adaptive voice recognition entry of inspection information based on staff accent types; staff accent type parameters are combined with an intelligent recognition algorithm and voice recognition parameters of different dialect accent types to perform customized matching of staff personal voice recognition data packet types and accurately collect inspection information, thereby realizing intelligent, accurate and efficient collection of inspection information based on different staff accent features, and improving the convenience and applicability of the inspection information entry system.

Description

Technical Field

[0001] The present invention relates to the technical field of inspection information data processing, and in particular to a written inspection information entry data processing system and method based on voice recognition. Background Art

[0002] With the development of science and technology, digital management has gradually become an important means for various industries to improve work efficiency and quality. In the field of inspection, traditional paper recording methods are still widely used. Although this traditional method is intuitive and simple to operate, it has some obvious shortcomings such as inefficient entry of inspection information and irregular writing. In addition, although the popularity of electronic equipment in inspection work has gradually increased, some staff are still not good at using electronic equipment due to age, skills and other factors, resulting in the inability to fully utilize the convenience brought by modern technology. The existing electronic inspection equipment cannot accurately enter inspection information based on the staff's accent characteristics during use, nor can it facilitate staff to conveniently and efficiently modify the entered inspection information results, reducing the accuracy and efficiency of inspection information entry.

[0003] A Chinese invention patent with announcement number CN104090648B discloses a data entry method and terminal, which respectively extracts data information and identifies user operation gestures through a data capture module and a quick entry module, and simultaneously uses the entry method corresponding to the identified operation gesture to accurately enter the extracted data information into the target area; achieves fast and accurate information entry and improves user experience; however, the above technical solution cannot achieve convenient modification of the input data results. Summary of the invention

[0004] 1. Technical issues to be resolved

[0005] In order to solve the problem that the above-mentioned existing electronic inspection equipment cannot accurately enter inspection information based on the staff's accent characteristics during use, and cannot facilitate the staff to conveniently and efficiently modify the entered inspection information results, thereby reducing the accuracy and efficiency of inspection information entry, the above purpose is to accurately collect staff accent information parameters, accurately analyze staff accent type parameters, and scientifically search for specific accent type voice recognition data packets; intelligently collect inspection information, scientifically digitize and feedback inspection information, accurately collect inspection information entry result feedback instructions, scientifically identify staff inspection information entry result feedback information, accurately collect inspection information identification information and update inspection information, and scientifically execute inspection information modification operations.

[0006] (II) Technical solution

[0007] The present invention is implemented by the following technical solution: a written inspection information input data processing method based on speech recognition, the method comprising the following steps:

[0008] S1. Collect audio data of staff accents;

[0009] S2, performing staff accent type recognition processing based on the staff accent audio data and the audio data of different dialect accent types to generate staff accent type data;

[0010] S3, searching and processing the speech recognition data packet object required for staff inspection information entry according to the staff accent type data and the speech recognition data packets of different dialect accent types, generating the speech recognition data packet required by the staff and collecting inspection text data;

[0011] S4, performing digital identification processing of the inspection information object based on the inspection text data, generating inspection text identification combination data and performing inspection information entry result display operation;

[0012] S5, collecting inspection information entry result feedback instruction data and performing inspection information entry result feedback instruction type identification processing with inspection information entry result feedback instruction type keyword data to generate inspection information entry result feedback instruction type data; when it is completed, ending this inspection information entry operation;

[0013] S6. When it is a modification, collect inspection information modification text object identification data and inspection information update text data;

[0014] S7. Construct inspection information modification feature data and execute inspection information modification operation.

[0015] Preferably, the operation steps of collecting the staff accent audio data are as follows:

[0016] S11. Collect the staff’s personal accent audio information online through the microphone on the electronic inspection equipment and generate staff’s accent audio data .

[0017] The present invention uses a microphone to efficiently and conveniently collect audio parameters of the staff's accent, thereby providing reliable data support for subsequent accurate analysis of the staff's accent feature types.

[0018] Preferably, the staff's accent type recognition processing is performed based on the staff's accent audio data and the audio data of different dialect accent types, and the operation steps of generating the staff's accent type data are as follows:

[0019] S21. Establishing audio data sets of different dialect accent types , ;in Indicates Audio data of different dialect accent types, Indicates the maximum value of the number of different dialect accent types, wherein the different dialect accent type audio data indicates the accent audio data of different local dialects, and the different dialect accent types include northern dialect, Hunan dialect, Gan dialect, Fujian dialect, Cantonese dialect and Hakka dialect;

[0020] S22, using Soundex algorithm to convert the staff accent audio data Audio data set with different dialect accent types Audio data of different dialect accent types described in Perform audio feature matching to search for audio data with the staff member’s accent Match the audio data of the different dialect accent types The dialect accent type is generated through data identification to generate staff accent type data .

[0021] The present invention performs accurate analysis of the staff's accent type based on scientifically preset audio parameters of different dialect accent types in combination with the Soundex algorithm and the staff's accent audio parameters, thereby achieving the effect of inputting inspection information based on adaptive speech recognition of the staff's accent type.

[0022] Preferably, the operation steps of searching and processing the speech recognition data packet object required for staff inspection information entry according to the staff accent type data and the speech recognition data packets of different dialect accent types, generating the speech recognition data packet required by the staff and collecting the inspection text data are as follows:

[0023] S31. Establishing a speech recognition data packet set of different dialect accent types ,in Indicates speech recognition data packets of different dialect accent types, wherein the speech recognition data packets of different dialect accent types represent a language translation database for speech recognition of different local dialect accent types;

[0024] S32: the staff accent type data A collection of speech recognition data packets with different dialect accent types Different dialects and accents described in the speech recognition data package According to the dialect accent type character feature matching, search for the staff accent type data The corresponding speech recognition data packets of different dialect accent types And generate the speech recognition data package required by the staff , execute to generate the speech recognition data packet required by the staff The specific steps are as follows:

[0025] S321, initialization, updating the maximum number of iterations T, in the speech recognition data packet set of different dialect accent types The language data packet is updated in the search space to identify the individual positions of the white shark population. The calculation formula for the language data packet to identify the individual positions of the white shark population is as follows: ,in Indicates language data packets identifying individual white sharks exist The position in the dimensional space, that is, the language data packet identifies the individual white shark In the spatial dimension A collection of speech recognition data packages for different dialect accent types The position in the search space, and Represents a collection of speech recognition data packets in different dialect accent types The upper and lower limits of Indicates value A random number between

[0026] S322, speed update stage, language data packet recognition White Shark in the different dialect accent type speech recognition data packet set Speech recognition data packets based on different dialect accent types in the search space The prey moves to sense its position and update its own speed, that is, in the speech recognition data packet set of different dialect accent types Search the search space for data related to the staff's accent type Speech recognition data packets matching the different dialect accent types , the calculation formula for updating its own speed is as follows: ,in Indicates language data packets identifying individual white sharks exist After the iterations, the speech recognition data packets of different dialects and accents are collected. The speed in the search space, Indicates language data packets identifying individual white sharks exist After the iterations, the speech recognition data packets of different dialects and accents are collected. The speed in the search space, Indicated in After iterations, the language data packet identifies the white shark individual in the different dialects and accent types of speech recognition data packet set The optimal position in the search space, Indicates language data packets identifying individual white sharks exist After iterations, the language data packet identifies the white shark individual in the different dialects and accent types of speech recognition data packet set Position in the search space; Indicates language data packets identifying individual white sharks exist After the iterations, the speech recognition data packets of different dialects and accents are collected. Search space and speed The corresponding position; represents the algorithm shrinkage coefficient, and Respectively and The control coefficient of and Both represent values A random number between

[0027] S323, position update stage, language data packet recognition White Shark through the optimal different dialect accent type speech recognition data packet The prey moves to update itself in the different dialect accent type speech recognition data packet collection The position in the search space, i.e., the set of speech recognition data packets of different dialect accent types Search the search space for data related to the staff's accent type The best matching or sub-optimal matching speech recognition data packets of different dialect accent types , the language data packet recognition white shark position update calculation formula is as follows: ,in Identification of individual white sharks using language packets representing the search for prey during the position update phase exist After the iterations, the speech recognition data packets of different dialects and accents are collected. The position in the search space, Represents a bitwise operator, represents a logical vector, and are binary vectors, Indicates the attraction coefficient of the language data packet to identify the white shark approaching prey, Indicates language data packet identification White Shark sports product wave frequency;

[0028] Language data packet recognition White Shark speech recognition data packet collection in different dialects and accent types In the search space, the white shark position is moved toward the optimal language data packet recognition position to get closer to the optimal speech recognition data packet of different dialect accent types The location of the prey, the language data packet identifies the white shark's location update calculation formula as follows: ,in The language packet representing the location update phase searches for the optimal language packet to identify the white shark location and the language packet to identify the white shark individual exist After the iterations, the speech recognition data packets of different dialects and accents are collected. The position in the search space, , , Both represent values A random number between Indicates the distance between the white shark and the prey identified by the language data packet, that is, the speech recognition data packet set of different dialect accent types Search the search space for data related to the staff's accent type Speech recognition data packets matching the different dialect accent types ; Indicates a sign-returning function. The language data packet identifies the olfactory and visual parameters of the white shark approaching optimal prey;

[0029] S324, fish school behavior stage, language data packet recognition white shark population through feeding behavior to retain the position update stage in the different dialect accent type speech recognition data packet set The optimal language data packet recognition white shark position in the search space is used to update other language data packet recognition white shark individual positions according to the optimal language data packet recognition white shark position to obtain a speech recognition data packet set of different dialect accent types. Search the search space for data related to the staff's accent type The best matching speech recognition data package of the different dialect accent types , the calculation formula for the language data packet to identify the white shark individual position update is as follows: ,in Indicates the stage of school behavior to update other language data packets to identify white shark individuals exist After the iterations, the speech recognition data packets of different dialects and accents are collected. The position in the search space, i.e., the set of speech recognition data packets of different dialect accent types Search the search space for data related to the staff's accent type Speech recognition data packets matching the different dialect accent types ;

[0030] S325: When the maximum number of iterations is met, output the data corresponding to the staff member's accent type. The best matching speech recognition data package of the different dialect accent types And generate the speech recognition data package required by the staff through data identification ;

[0031] S33, the staff collects the inspection work audio data online through the voice recognition device carried by the electronic inspection equipment, and recognizes the data packet according to the voice recognition data required by the staff Translate the inspection work audio data into inspection text information and construct inspection text data .

[0032] The present invention combines the staff accent type parameters with the White Shark optimization algorithm and the speech recognition parameters of different dialect accent types to perform customized matching of the staff individual speech recognition data packet type and accurately collect inspection information, thereby achieving the effect of intelligently, accurately and efficiently collecting inspection information based on the accent characteristics of different staff members.

[0033] Preferably, the operation steps of performing digital identification processing of the inspection information object based on the inspection text data, generating inspection text identification combination data and performing the inspection information entry result display operation are as follows:

[0034] S41, the inspection text data The inspection text character object is digitally identified and a collection of inspection text identification data is generated. , ;in Indicates the inspection text data Middle The inspection text identification data corresponding to the inspection text character object, Indicates the maximum number of inspection text character objects, the inspection text character objects include words, numbers and punctuation marks, and the inspection text identification data indicates an identification code consisting of one or more lowercase numbers;

[0035] S42, the inspection text data and the inspection text identification data set Inspection text identification data described in to Combine data according to the inspection text character object as the reference feature to generate inspection text identification combination data ,in The inspection text identification combined data represents the combined data of the upper and lower structures in which the inspection text identification information is at the top and the inspection text character object is at the bottom;

[0036] S43, combining the inspection text identification data The inspection information entry and result display operations are performed through the display screen of the electronic inspection equipment.

[0037] The present invention performs digital identification processing on the inspection information text, scientifically constructs the inspection text identification combination parameters, and combines the display screen to intuitively and visually display the inspection information entry results, thereby achieving the effect of accurate identification and visual display of the inspection information.

[0038] Preferably, the inspection information entry result feedback instruction data is collected and the inspection information entry result feedback instruction type identification processing is performed with the inspection information entry result feedback instruction type keyword data to generate the inspection information entry result feedback instruction type data; when it is completed, the operation steps for ending this inspection information entry operation are as follows:

[0039] S51, collecting the staff's feedback instruction parameters on the inspection information entry results online through the user interface instruction touch menu of the electronic inspection equipment display screen and generating inspection information entry result feedback instruction data , the inspection information entry result feedback instruction data includes inspection information modification feedback instruction data and inspection information completion feedback instruction data;

[0040] S52. Establish inspection information entry result feedback instruction type keyword data set ,in Indicates the inspection information entry result feedback instruction data Enter the inspection information result feedback instruction type keyword data corresponding to the inspection information modification feedback instruction data, Indicates the inspection information entry result feedback instruction data Inspection information entry result feedback instruction type keyword data corresponding to the inspection information completion feedback instruction data, wherein the inspection information entry result feedback instruction type keyword data includes modification and completion;

[0041] S53, using XGBoost algorithm to input the inspection information into result feedback instruction data Keyword data set related to the inspection information entry result feedback instruction type The inspection information entry result feedback instruction type keyword data is matched according to the inspection information entry result feedback instruction character, and the inspection information entry result feedback instruction data is searched out. The corresponding inspection information entry result feedback instruction type keyword data is generated through data identification to generate inspection information entry result feedback instruction type data ;

[0042] when and If the characters of the inspection information input result feedback command are matched successfully, it means that the inspection information input result needs to be modified, and the inspection information input result feedback command type data is output. To modify;

[0043] when and If the characters of the inspection information input result feedback instruction match successfully, it means that the inspection information input result does not need to be modified, and the inspection information input result feedback instruction type data is output. To complete, end the inspection information entry work at this time.

[0044] The present invention achieves the effect of intelligent and accurate feedback of inspection information entry results by autonomously analyzing inspection information entry result feedback information based on obtaining inspection information entry result feedback instructions in combination with XGBoost algorithm and inspection information entry result feedback instruction type keywords.

[0045] Preferably, when the modification is performed, the operation steps of collecting the inspection information modification text object identification data and the inspection information update text data are as follows:

[0046] S61, when the inspection information input result feedback instruction type data When modifying, the staff collects the identification code audio data of the inspection information modification object online through the voice recognition device carried by the electronic inspection equipment, and recognizes the data packet according to the voice recognition data packet required by the staff. Translate the inspection information modification object identification code audio data into inspection information modification text object identification information, and construct the inspection information modification text object identification data ;

[0047] The staff collects the updated text and audio data of the inspection information modification object online through the voice recognition device carried by the electronic inspection equipment, and recognizes the data packet according to the voice recognition data required by the staff. Translate the updated text audio data of the inspection information modification object into the inspection information update text information, and construct the inspection information update text data .

[0048] The present invention achieves the effect of accurately collecting the modification identification information and text update information of the inspection information by collecting the inspection information modification text object identification parameters and the inspection information update text parameters.

[0049] Preferably, the steps of constructing the inspection information modification characteristic data and executing the inspection information modification operation are as follows:

[0050] S71, modifying the inspection information into text object identification data And the inspection information updates the text data Combine data to construct inspection information modification feature data ,in ;

[0051] S72: Modify the characteristic data according to the inspection information The inspection information modification text object identification parameter and the inspection information update text parameter are combined with the inspection text identification data Inspection text data described in Execute the inspection information modification operation.

[0052] The present invention scientifically constructs inspection information modification characteristic parameters based on the modification identification and text update information of the inspection information and scientifically executes the inspection information modification operation, thereby achieving the effect of timely and accurate acquisition of inspection information modification parameters and accurate and targeted modification operations on the inspection information.

[0053] A written inspection information entry data processing system based on speech recognition, used to implement the written inspection information entry data processing method based on speech recognition, the system includes an inspection information entry module, an inspection information modification module, and an inspection information modification execution module;

[0054] The inspection information entry module includes a staff accent audio parameter collection unit, a different accent type audio parameter storage unit, a staff accent type recognition unit, a different accent type speech recognition data packet storage unit, a specific accent type speech recognition data packet search unit, and an inspection information collection unit;

[0055] The staff accent audio parameter acquisition unit acquires staff accent audio data through a microphone; the different accent type audio parameter storage unit is used to store audio data of different dialect accent types; the staff accent type recognition unit performs staff accent type recognition processing based on the staff accent audio data and audio data of different dialect accent types to generate staff accent type data; the different accent type speech recognition data packet storage unit is used to store speech recognition data packets of different dialect accent types; the specific accent type speech recognition data packet search unit performs speech recognition data packet object search processing required for staff inspection information entry based on the staff accent type data and speech recognition data packets of different dialect accent types to generate speech recognition data packets required for staff; the inspection information acquisition unit collects inspection text data based on the speech recognition data packets required by the staff through a speech recognition device;

[0056] The inspection information modification module includes an inspection information identification display unit, an inspection information entry result feedback instruction collection unit, an inspection information entry result feedback instruction type keyword storage unit, an inspection information entry result feedback instruction type identification unit, an inspection information identification information collection unit, and an inspection information update information collection unit;

[0057] The inspection information identification display unit performs inspection information object digital identification processing based on the inspection text data, generates inspection text identification combination data and uses a display screen to perform inspection information entry result display operations; the inspection information entry result feedback instruction acquisition unit acquires inspection information entry result feedback instruction data through the user interface instruction touch menu of the display screen; the inspection information entry result feedback instruction type keyword storage unit is used to store inspection information entry result feedback instruction type keyword data; the inspection information entry result feedback instruction type identification unit performs inspection information entry result feedback instruction type identification processing based on the inspection information entry result feedback instruction data and the inspection information entry result feedback instruction type keyword data to generate inspection information entry result feedback instruction type data; the inspection information identification information acquisition unit acquires inspection information modification text object identification data through a voice recognition device based on the voice recognition data packet required by the staff; the inspection information update information acquisition unit acquires inspection information update text data through a voice recognition device based on the voice recognition data packet required by the staff;

[0058] The inspection information modification execution module includes an inspection information modification information collection unit and an inspection information modification job execution unit;

[0059] The inspection information modification information collection unit constructs inspection information modification characteristic data; the inspection information modification operation execution unit executes the inspection information modification operation according to the inspection information modification characteristic data.

[0060] (III) Beneficial effects

[0061] The present invention provides a written inspection information input data processing system and method based on voice recognition. It has the following beneficial effects:

[0062] 1. Efficiently and conveniently collect staff accent audio parameters through microphones to provide data support for subsequent accurate analysis of staff accent feature types; accurately analyze staff accent types based on scientifically preset audio parameters of different dialect accent types combined with intelligent audio matching algorithms and staff accent audio parameters, and realize adaptive voice recognition based on staff accent types to enter inspection information; combine staff accent type parameters with intelligent recognition algorithms and voice recognition parameters of different dialect accent types to perform customized matching of staff personal voice recognition data packet types and accurately collect inspection information, realizing intelligent, accurate and efficient collection of inspection information based on different staff accent characteristics, and improving the convenience and applicability of the inspection information entry system.

[0063] 2. Through digital identification of inspection information objects on inspection information texts, scientifically construct inspection text identification combination parameters and combine with display screen to intuitively visualize inspection information entry results; based on obtaining inspection information entry result feedback instructions, combine intelligent search algorithms with inspection information entry result feedback instruction type keywords to autonomously analyze inspection information entry result feedback information; accurately collect inspection information modification text object identification parameters and inspection information update text parameters, realize accurate collection of inspection information modification identification information and text update information, and improve the flexibility and efficiency of data processing of inspection information entry system.

[0064] 3. By scientifically constructing inspection information modification characteristic parameters based on the modification identification and text update information of inspection information, timely and accurate acquisition of inspection information modification parameters is achieved; based on the inspection information modification characteristic parameters, the inspection information modification operation is scientifically performed to achieve accurate and targeted modification of inspection information, thereby improving the quality and accuracy of data collection in the inspection information entry system. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] Figure 1 A module diagram of a written inspection information entry data processing system based on speech recognition provided by the present invention;

[0066] Figure 2 The present invention provides a flowchart of a method for processing written inspection information entry data based on speech recognition. DETAILED DESCRIPTION

[0067] 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 ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0068] The embodiments of the written inspection information input data processing system and method based on voice recognition are as follows:

[0069] Example 1: Please refer to Figure 1 - Figure 2 , a written inspection information input data processing method based on speech recognition, the method comprises the following steps:

[0070] S1. Collect audio data of staff accents;

[0071] S2, performing staff accent type recognition processing based on the staff accent audio data and the different dialect accent type audio data to generate staff accent type data;

[0072] S3, searching and processing the speech recognition data packet object required for staff inspection information entry according to the staff accent type data and the speech recognition data packets of different dialect accent types, generating the speech recognition data packet required by the staff and collecting inspection text data;

[0073] S4. Perform digital identification processing on the inspection information object based on the inspection text data, generate inspection text identification combination data and perform inspection information entry result display operation;

[0074] S5, collecting inspection information entry result feedback instruction data and performing inspection information entry result feedback instruction type identification processing with inspection information entry result feedback instruction type keyword data to generate inspection information entry result feedback instruction type data; when it is completed, ending this inspection information entry operation;

[0075] S6. When it is a modification, collect inspection information modification text object identification data and inspection information update text data;

[0076] S7. Construct inspection information modification feature data and execute inspection information modification operation.

[0077] For further information, see Figure 1 - Figure 2 ,The steps for collecting staff accent audio data are as follows:

[0078] S11. Collect the staff’s personal accent audio information online through the microphone on the electronic inspection equipment and generate staff’s accent audio data .

[0079] The steps for performing staff accent type recognition processing based on staff accent audio data and different dialect accent type audio data to generate staff accent type data are as follows:

[0080] S21. Establishing audio data sets of different dialect accent types , ;in Indicates Audio data of different dialect accent types, Indicates the maximum value of the number of different dialect accent types. The different dialect accent type audio data indicates the accent audio data of different local dialects. The different dialect accent types include northern dialect, Hunan dialect, Gan dialect, Fujian dialect, Cantonese dialect and Hakka dialect.

[0081] S22, using Soundex algorithm to convert the staff accent audio data Audio data collection with different dialect accent types Audio data of different dialect accent types Perform audio feature matching to search for audio data with the staff’s accent Matching audio data of different dialect accent types The dialect accent type is generated through data identification to generate staff accent type data .

[0082] The following are the steps to search and process the speech recognition data package objects required for staff inspection information entry based on the staff accent type data and the speech recognition data packages of different dialect accent types, generate the speech recognition data package required by the staff, and collect the inspection text data:

[0083] S31. Establishing a speech recognition data packet set of different dialect accent types ,in Indicates Different dialect accent type speech recognition data packets, the different dialect accent type speech recognition data packets represent language translation databases used for speech recognition of different local dialect accent types;

[0084] S32. Staff accent type data A collection of speech recognition data packages with different dialect accent types Speech recognition data package for different dialect accent types According to the character feature matching of dialect accent type, search for the staff accent type data Corresponding speech recognition data packages for different dialect accent types And generate the speech recognition data package required by the staff , execute to generate the speech recognition data packet required by the staff The specific steps are as follows:

[0085] S321, initialization, update the maximum number of iterations T, speech recognition data packet set in different dialect accent types The language data packet is updated in the search space to identify the individual positions of the white shark population. The calculation formula for the language data packet to identify the individual positions of the white shark population is as follows: ,in Indicates language data packets identifying individual white sharks exist The position in the dimensional space, that is, the language data packet identifies the individual white shark In the spatial dimension A collection of speech recognition data packages for different dialect accent types The position in the search space, and Respectively represent the speech recognition data package collection of different dialect accent types The upper and lower limits of Indicates value A random number between

[0086] S322, speed update stage, language data packet recognition White Shark speech recognition data packet collection in different dialect accent types Speech recognition data packets based on different dialect accent types in the search space The prey moves to sense its position and update its own speed, that is, a collection of speech recognition data packets in different dialects and accents Search space to find data related to the staff's accent type Speech recognition data packages matching different dialect accent types , the calculation formula for updating its own speed is as follows: ,in Indicates language data packets identifying individual white sharks exist The speech recognition data set of different dialects and accents after iterations The speed in the search space, Indicates language data packets identifying individual white sharks exist The speech recognition data set of different dialects and accents after iterations The speed in the search space, Indicated in After iterations, the language data packet is identified in the speech recognition data packet set of white shark individuals in different dialects and accent types. The optimal position in the search space, Indicates language data packets identifying individual white sharks exist After iterations, the language data packet is identified in the speech recognition data packet set of white shark individuals in different dialects and accent types. Position in the search space; Indicates language data packets identifying individual white sharks exist The speech recognition data set of different dialects and accents after iterations Search space and speed The corresponding position; represents the algorithm shrinkage coefficient, and Respectively and The control coefficient of and Both represent values A random number between

[0087] S323, position update stage, language data packet recognition White Shark through the optimal different dialect accent type speech recognition data packet The prey moves to update its own speech recognition data set in different dialects and accents The position in the search space, i.e. the collection of speech recognition data packets in different dialect accent types Search space to find data related to the staff's accent type Best or sub-best matching speech recognition data package for different dialect accent types , the language data packet recognition white shark position update calculation formula is as follows: ,in Identification of individual white sharks using language packets representing the search for prey during the position update phase exist The speech recognition data set of different dialects and accents after iterations The position in the search space, Represents a bitwise operator, represents a logical vector, and are binary vectors, Indicates the attraction coefficient of the language data packet to identify the white shark approaching prey, Indicates language data packet identification White Shark sports product wave frequency;

[0088] Language data packet recognition White Shark speech recognition data packet collection in different dialects and accents In the search space, the white shark position is moved towards the optimal language data packet recognition to get closer to the optimal speech recognition data packet of different dialect accent types The location of the prey, the language data packet identifies the white shark's location update calculation formula as follows: ,in The language packet representing the location update phase searches for the optimal language packet to identify the white shark location and the language packet to identify the white shark individual exist The speech recognition data set of different dialects and accents after iterations The position in the search space, , , Both represent values A random number between Represents the distance between the white shark and its prey identified by the language data packet, i.e., a collection of speech recognition data packets in different dialects and accents Search space to find data related to the staff's accent type Speech recognition data package matching different dialect accent types ; Indicates a sign-returning function. The language data packet identifies the olfactory and visual parameters of the white shark approaching optimal prey;

[0089] S324, fish school behavior stage, language data packet recognition white shark population through feeding behavior to retain the position update stage in different dialects and accent types of speech recognition data packet collection The optimal language data packet recognition white shark position in the search space is used to update the individual positions of other language data packets recognition white sharks according to the optimal language data packet recognition white shark position to obtain a set of speech recognition data packets in different dialect accent types. Search space to find data related to the staff's accent type The most matching speech recognition data package for different dialect accent types , the calculation formula for the language data packet to identify the white shark individual position update is as follows: ,in Indicates the stage of school behavior to update other language data packets to identify white shark individuals exist The speech recognition data set of different dialects and accents after iterations The position in the search space, i.e. the collection of speech recognition data packets in different dialect accent types Search space to find data related to the staff's accent type Speech recognition data package matching different dialect accent types ;

[0090] S325: When the maximum number of iterations is met, output the data corresponding to the staff member's accent type The most matching speech recognition data package for different dialect accent types And generate the speech recognition data package required by the staff through data identification ;

[0091] S33. The staff collects the audio data of the inspection work online through the voice recognition device on the electronic inspection equipment, and recognizes the data packets according to the voice recognition data required by the staff. Translate the inspection work audio data into inspection text information and construct inspection text data .

[0092] Through the staff accent audio parameter collection unit, a microphone is used to efficiently and conveniently collect the staff accent audio parameters, providing data support for the subsequent accurate analysis of the staff accent feature types; the different accent type audio parameter storage units and the staff accent type recognition units cooperate with each other, based on the scientific preset audio parameters of different dialect accent types combined with the intelligent audio matching algorithm and the staff accent audio parameters to perform accurate analysis of the staff accent type, and realize adaptive voice recognition based on the staff accent type to enter the inspection information; the specific accent type voice recognition data packet search unit and the inspection information collection unit cooperate with each other, combining the staff accent type parameters with the intelligent recognition algorithm and the voice recognition parameters of different dialect accent types to perform customized matching of the staff's personal voice recognition data packet type and accurately collect inspection information, realizing intelligent, accurate and efficient collection of inspection information based on different staff accent characteristics, and improving the convenience and applicability of the inspection information entry system.

[0093] For further information, see Figure 1 - Figure 2 The steps for performing digital identification processing of inspection information objects based on inspection text data, generating inspection text identification combination data and performing inspection information entry result display operations are as follows:

[0094] S41, inspection text data The inspection text character object is digitally identified and a collection of inspection text identification data is generated. , ;in Indicates inspection text data Middle The inspection text identification data corresponding to the inspection text character object, Indicates the maximum number of inspection text character objects. Inspection text character objects include letters, numbers and punctuation marks. Inspection text identification data indicates an identification code consisting of one or more lowercase numbers.

[0095] S42, inspection text data And inspection text identification data set Inspection text identification data to Combine data according to the inspection text character object as the reference feature to generate inspection text identification combination data ,in The inspection text identification combined data represents the combined data of the upper and lower structures in which the inspection text identification information is at the top and the inspection text character object is at the bottom;

[0096] S43, combining inspection text identification data The inspection information entry and result display operations are performed through the display screen of the electronic inspection equipment.

[0097] Collect the inspection information entry result feedback instruction data and perform inspection information entry result feedback instruction type identification processing with the inspection information entry result feedback instruction type keyword data to generate inspection information entry result feedback instruction type data; when it is completed, the operation steps to end this inspection information entry operation are as follows:

[0098] S51, collecting the staff's feedback instruction parameters on the inspection information entry results online through the user interface instruction touch menu of the electronic inspection equipment display screen and generating inspection information entry result feedback instruction data , the inspection information entry result feedback instruction data includes the inspection information modification feedback instruction data and the inspection information completion feedback instruction data;

[0099] S52. Establish inspection information entry result feedback instruction type keyword data set ,in Indicates inspection information entry result feedback instruction data Enter the inspection information result feedback instruction type keyword data corresponding to the inspection information modification feedback instruction data, Indicates inspection information entry result feedback instruction data Inspection information entry result feedback instruction type keyword data corresponding to the inspection information completion feedback instruction data, the inspection information entry result feedback instruction type keyword data including modification and completion;

[0100] S53, using XGBoost algorithm to input inspection information into result feedback instruction data Inspection information input result feedback instruction type keyword data set In the inspection information input result feedback instruction type keyword data, match the inspection information input result feedback instruction characters to search for the inspection information input result feedback instruction data The corresponding inspection information entry result feedback instruction type keyword data is generated through data identification and inspection information entry result feedback instruction type data ;

[0101] when and If the characters of the inspection information input result feedback command are matched successfully, it means that the inspection information input result needs to be modified, and the inspection information input result feedback command type data is output. To modify;

[0102] when and If the characters of the inspection information input result feedback instruction match successfully, it means that the inspection information input result does not need to be modified, and the inspection information input result feedback instruction type data is output. To complete, end the inspection information entry work at this time.

[0103] When modifying, the steps for collecting inspection information to modify text object identification data and updating inspection information text data are as follows:

[0104] S61, when the inspection information is entered and the result feedback instruction type data is received When modifying, the staff collects the identification code audio data of the inspection information modification object online through the voice recognition device carried by the electronic inspection equipment, and recognizes the data packet according to the voice recognition data packet required by the staff. Translate the inspection information modification object identification code audio data into inspection information modification text object identification information, and construct the inspection information modification text object identification data ;

[0105] The staff collects the updated text and audio data of the inspection information modification object online through the voice recognition device installed in the electronic inspection equipment, and recognizes the data packet according to the voice recognition data required by the staff. Translate the updated text audio data of the inspection information modification object into the inspection information update text information, and construct the inspection information update text data .

[0106] Through the inspection information identification display unit, the inspection information text is digitally identified, the inspection text identification combination parameters are scientifically constructed, and the inspection information entry results are intuitively and visually displayed in combination with the display screen; the inspection information entry result feedback instruction collection unit and the inspection information entry result feedback instruction type recognition unit cooperate with each other, and the inspection information entry result feedback information is autonomously analyzed based on the acquisition of the inspection information entry result feedback instruction combined with the intelligent search algorithm and the inspection information entry result feedback instruction type keywords; the inspection information identification information collection unit and the inspection information update information collection unit cooperate with each other to accurately collect the inspection information modification text object identification parameters and the inspection information update text parameters, so as to realize the accurate collection of the inspection information modification identification information and text update information, thereby improving the flexibility and efficiency of the inspection information entry system data processing.

[0107] For further information, see Figure 1 - Figure 2 The steps to construct inspection information modification feature data and perform inspection information modification operations are as follows:

[0108] S71, modify the inspection information text object identification data Update text data with inspection information Combine data to construct inspection information modification feature data ,in ;

[0109] S72. Modify characteristic data according to inspection information The inspection information modification text object identification parameter and the inspection information update text parameter are combined with the inspection text identification data Inspection text data Execute the inspection information modification operation.

[0110] Through the cooperation of the inspection information modification information collection unit and the inspection information modification operation execution unit, the inspection information modification characteristic parameters are scientifically constructed based on the modification identification and text update information of the inspection information, so as to realize the timely and accurate acquisition of the inspection information modification parameters; the inspection information modification operation is scientifically executed based on the inspection information modification characteristic parameters, so as to realize the precise and targeted modification operation of the inspection information, and improve the quality and accuracy of data collection of the inspection information entry system.

[0111] Example 2: Please refer to Figure 1 - Figure 2 , a written inspection information input data processing system based on speech recognition, used to implement a written inspection information input data processing method based on speech recognition, the system includes an inspection information input module, an inspection information modification module, and an inspection information modification execution module;

[0112] The inspection information entry module includes a staff accent audio parameter collection unit, a different accent type audio parameter storage unit, a staff accent type recognition unit, a different accent type voice recognition data packet storage unit, a specific accent type voice recognition data packet search unit, and an inspection information collection unit;

[0113] A staff accent audio parameter collection unit, which collects staff accent audio data through a microphone; a different accent type audio parameter storage unit, which is used to store audio data of different dialect accent types; a staff accent type recognition unit, which performs staff accent type recognition processing based on staff accent audio data and audio data of different dialect accent types, and generates staff accent type data; a different accent type voice recognition data packet storage unit, which is used to store voice recognition data packets of different dialect accent types; a specific accent type voice recognition data packet search unit, which performs voice recognition data packet object search processing required for staff inspection information entry based on staff accent type data and voice recognition data packets of different dialect accent types, and generates voice recognition data packets required by staff; an inspection information collection unit, which collects inspection text data based on voice recognition data packets required by staff through a voice recognition device;

[0114] The inspection information modification module includes an inspection information identification display unit, an inspection information entry result feedback instruction collection unit, an inspection information entry result feedback instruction type keyword storage unit, an inspection information entry result feedback instruction type identification unit, an inspection information identification information collection unit, and an inspection information update information collection unit;

[0115] The inspection information identification display unit performs digital identification processing of the inspection information object based on the inspection text data, generates inspection text identification combination data and uses the display screen to perform the inspection information entry result display operation; the inspection information entry result feedback instruction collection unit collects the inspection information entry result feedback instruction data through the user interface instruction touch menu of the display screen; the inspection information entry result feedback instruction type keyword storage unit is used to store the inspection information entry result feedback instruction type keyword data; the inspection information entry result feedback instruction type identification unit performs type identification processing of the inspection information entry result feedback instruction based on the inspection information entry result feedback instruction data and the inspection information entry result feedback instruction type keyword data to generate the inspection information entry result feedback instruction type data; the inspection information identification information collection unit collects the inspection information modification text object identification data through the voice recognition device according to the voice recognition data packet required by the staff; the inspection information update information collection unit collects the inspection information update text data through the voice recognition device according to the voice recognition data packet required by the staff;

[0116] The inspection information modification execution module includes an inspection information modification information collection unit and an inspection information modification job execution unit;

[0117] The inspection information modification information collection unit constructs inspection information modification characteristic data; the inspection information modification operation execution unit executes the inspection information modification operation according to the inspection information modification characteristic data.

[0118] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A written inspection information input data processing method based on speech recognition, characterized in that: The method comprises the following steps: S1, collecting staff accent audio data, said S1 comprises the following steps: S11. Collect the staff’s personal accent audio information online through the microphone on the electronic inspection equipment and generate staff’s accent audio data ; S2, performing staff accent type recognition processing based on the staff accent audio data and the different dialect accent type audio data to generate staff accent type data, wherein S2 includes the following steps: S21. Establishing audio data sets of different dialect accent types , ;in Indicates Different dialect accent types audio data, Indicates the maximum number of different dialect accent types; S22, using Soundex algorithm to convert the staff accent audio data Audio data set with different dialect accent types Audio data of different dialect accent types described in Perform audio feature matching to search for audio data with the staff member’s accent Match the audio data of the different dialect accent types The dialect accent type is generated through data identification to generate staff accent type data ; S3, performing a search process on the speech recognition data packet object required for staff inspection information entry according to the staff accent type data and the speech recognition data packets of different dialect accent types, generating the speech recognition data packet required by the staff and collecting inspection text data, wherein S3 comprises the following steps: S31. Establishing a speech recognition data packet set of different dialect accent types ,in Indicates Speech recognition data packages of different dialect accent types; S32: the staff accent type data A collection of speech recognition data packets with different dialect accent types Different dialects and accents described in the speech recognition data package According to the dialect accent type character feature matching, search for the staff accent type data The corresponding speech recognition data packets of different dialect accent types And generate the speech recognition data package required by the staff , execute to generate the speech recognition data packet required by the staff The specific steps are as follows: S321, initialization, update the maximum number of iterations T, in the The updated generated language data packets in the search space identify the individual locations of the white shark population; S322, speed update stage, the language data packet identifies the white shark in the In the search space, according to The prey moves to sense its position and update its own speed, that is, Search the search space for Matching the ; S323, position update stage, language data packet identification white shark by moving toward the optimal The prey moves to update itself in the The position in the search space, i.e., Search the search space for The best match or the second best match ; Language data packet identification White Shark in the The search space is moved towards the optimal language packet recognition white shark position to get closer to the optimal The location of the prey; S324, fish school behavior stage, language data packet identification white shark population through feeding behavior to retain the position update stage described in The optimal language data packet in the search space identifies the white shark position, and updates other language data packets to identify the white shark individual positions according to the optimal language data packet to obtain the white shark individual positions in the search space. Search the search space for The best match for ; S325. When the maximum number of iterations is met, output The best match for And generate the speech recognition data package required by the staff through data identification ; S33, staff collects inspection audio data online through the voice recognition device on the electronic inspection equipment, and Translate the inspection work audio data into inspection text information and construct inspection text data ; S4, performing digital identification processing of the inspection information object based on the inspection text data, generating inspection text identification combination data and performing inspection information entry result display operation, wherein S4 includes the following steps: S41, the inspection text data The inspection text character object is digitally identified and a collection of inspection text identification data is generated. , ;in Indicates the inspection text data Middle The inspection text identification data corresponding to the inspection text character object, Indicates the maximum number of inspection text character objects, wherein the inspection text identification data indicates an identification code consisting of one or more lowercase digits; S42, the inspection text data and the inspection text identification data set Inspection text identification data described in to Combine data according to the inspection text character object as the reference feature to generate inspection text identification combination data ,in ; S43, combining the inspection text identification data Perform inspection information entry and result display operations through the display screen of the electronic inspection equipment; S5, collecting inspection information entry result feedback instruction data and performing inspection information entry result feedback instruction type identification processing with inspection information entry result feedback instruction type keyword data to generate inspection information entry result feedback instruction type data; when it is completed, ending this inspection information entry operation, said S5 includes the following steps: S51, collecting the staff's feedback instruction parameters on the inspection information entry results online through the user interface instruction touch menu of the electronic inspection equipment display screen and generating inspection information entry result feedback instruction data , the inspection information entry result feedback instruction data includes inspection information modification feedback instruction data and inspection information completion feedback instruction data; S52. Establish inspection information entry result feedback instruction type keyword data set ,in Indicates the Enter the inspection information result feedback instruction type keyword data corresponding to the inspection information modification feedback instruction data, Indicates the Inspection information entry result feedback instruction type keyword data corresponding to the inspection information completion feedback instruction data, wherein the inspection information entry result feedback instruction type keyword data includes modification and completion; S53, using XGBoost algorithm to With the The inspection information entry result feedback instruction type keyword data is matched according to the inspection information entry result feedback instruction character, and the inspection information entry result feedback instruction character is searched. The corresponding inspection information entry result feedback instruction type keyword data is generated through data identification to generate inspection information entry result feedback instruction type data ; when and If the characters of the inspection information input result feedback instruction match successfully, the inspection information input result feedback instruction type data will be output To modify; when and If the characters of the inspection information input result feedback instruction match successfully, the inspection information input result feedback instruction type data will be output To complete, end the inspection information entry operation at this time; S6. When it is a modification, collect inspection information modification text object identification data and inspection information update text data; S7. Construct inspection information modification feature data and execute inspection information modification operation.

2. The method for processing written inspection information input data based on speech recognition according to claim 1 is characterized in that: The S6 comprises the following steps: S61, when the When modifying, the staff collects the identification code audio data of the inspection information modification object online through the voice recognition device carried by the electronic inspection equipment, and Translate the inspection information modification object identification code audio data into inspection information modification text object identification information, and construct the inspection information modification text object identification data ; The staff collects the updated text and audio data of the inspection information modification object online through the voice recognition device installed in the electronic inspection equipment. Translate the updated text audio data of the inspection information modification object into the inspection information update text information, and construct the inspection information update text data .

3. The method for processing written inspection information input data based on speech recognition according to claim 2 is characterized in that: The S7 comprises the following steps: S71, the and stated Combine data to construct inspection information modification feature data ,in ; S72. According to the The inspection information modification text object identification parameter and the inspection information update text parameter are As stated in Execute the inspection information modification operation.

4. A written inspection information input data processing system based on speech recognition, used to implement the written inspection information input data processing method based on speech recognition as described in any one of claims 1 to 3, characterized in that: The system comprises an inspection information input module, an inspection information modification module and an inspection information modification execution module.

Images