Geological survey data voice structural collection method and system based on large language model

Abstract

The present application is suitable for the technical field of geological survey voice collection, and provides a geological survey data voice structured collection method and system based on a large language model. The present application selects a transfer device and constructs a collection transmission network; performs network transmission through the transfer device; determines a plurality of geological survey objects; performs preprocessing and identification; and performs structured mapping and filling of a plurality of key information. The present application can construct a collection transmission network between a plurality of collection devices, collect personnel voice data to the transfer device, and then perform network transmission. Furthermore, the present application compares and verifies a plurality of distances and personnel voice data, determines a plurality of geological survey objects, divides, preprocesses, identifies, extracts and structures the personnel voice data, and realizes collaborative communication and voice data convergence. Moreover, the present application can avoid the problem of voice transmission interruption, and ensures the formation of unified and complete survey object data.

Description

Technical Field

[0001] This invention belongs to the field of geological survey voice acquisition technology, and particularly relates to a method and system for structured voice acquisition of geological survey data based on a large language model. Background Technology

[0002] Geological surveys are an important means of obtaining basic data on regional geology, lithology, stratigraphy, structure, mineral resources and geological phenomena. During field geological surveys, investigators usually need to record the observed lithological characteristics, geological structures, occurrence information, geographical locations and geological phenomena in real time.

[0003] In existing technologies, voice acquisition in geological surveys typically focuses on the voice acquisition and conversion processing of a single investigator. However, in actual geological surveys, multiple investigators collaborate to conduct the surveys. Furthermore, unstable network signals are common during field geological surveys. Existing technologies, which use a single acquisition device as an independent data source, lack collaborative communication and voice data aggregation mechanisms among multiple investigators. This results in voice data collected by multiple investigators being independent of each other and prone to voice transmission interruptions, making it difficult to form unified and complete survey data. Summary of the Invention

[0004] The purpose of this invention is to provide a method and system for structured voice acquisition of geological survey data based on a large language model, in order to solve the technical problems existing in the prior art mentioned in the background.

[0005] The embodiments of the present invention are implemented as follows:

[0006] A method for structured speech acquisition of geological survey data based on a large language model, the method specifically includes the following steps:

[0007] When multiple personnel conduct field geological surveys simultaneously, the system acquires the location data of multiple geological survey personnel and the equipment signals of multiple acquisition devices, selects relay equipment, and constructs an acquisition and transmission network.

[0008] The voice data of multiple geological surveyors is collected and aggregated through the acquisition and transmission network and transmitted over the network via the relay equipment.

[0009] Based on multiple positioning data, the distance between multiple geological survey personnel is calculated, and the distance between multiple personnel and the voice data of multiple personnel are compared and verified to identify multiple geological survey objects;

[0010] Based on multiple geological survey objects, the voice data of multiple objects are divided, and preprocessed and recognized based on a preset geological language model to obtain the collected text of multiple objects.

[0011] Extract multiple key pieces of information from the collected text of multiple objects, and fill them in according to the preset multiple object collection templates to generate multiple structured collection data.

[0012] As a further limitation of the technical solution of this invention, the steps of acquiring the location data of multiple geological surveyors and the equipment signals of multiple acquisition devices, selecting relay devices, and constructing an acquisition and transmission network when multiple personnel are conducting field geological surveys simultaneously specifically include the following steps:

[0013] When multiple personnel conduct field geological surveys simultaneously, multiple geological survey personnel should be identified;

[0014] Real-time location tracking of multiple geological survey personnel was performed to obtain multiple location data.

[0015] Acquire device signals from the data acquisition equipment of multiple geological surveyors;

[0016] The signals from multiple devices are compared, and a relay device is selected from the multiple acquisition devices;

[0017] Based on the multiple positioning data, a data acquisition and transmission network is constructed by converging multiple other acquisition devices adjacent to each other and converging to the relay device.

[0018] As a further limitation of the technical solution of this embodiment of the invention, the step of collecting and aggregating the voice data of multiple geological surveyors to the relay device through the acquisition and transmission network, and transmitting it through the relay device, specifically includes the following steps:

[0019] Voice data of multiple geological survey personnel were collected using multiple acquisition devices.

[0020] Based on the acquisition and transmission network, the voice data of multiple individuals are collected and transmitted to the relay device;

[0021] The relay device is used to transmit the voice data of multiple individuals over a network.

[0022] As a further limitation of the technical solution of this invention embodiment, the step of calculating the distance between multiple geological survey personnel based on multiple positioning data, comparing and verifying the multiple distances and the multiple personnel voice data to determine multiple geological survey objects specifically includes the following steps:

[0023] Based on the aforementioned location data, the distance between multiple geological survey personnel is calculated;

[0024] The multiple intervals are compared with a preset distance threshold;

[0025] When the distance between two geological surveyors is less than a distance threshold, it is determined that the two geological surveyors are surveying the same geological object.

[0026] When the distance between two geological surveyors is not less than the distance threshold, two voice sample segments are extracted from the voice data of the two geological surveyors.

[0027] Multiple speech sample segments are subjected to feature recognition and comparison to determine whether the speech descriptions refer to the same geological survey object;

[0028] Record the comparison and judgment results to identify multiple different geological survey objects.

[0029] As a further limitation of the technical solution of this invention, the step of dividing the voice data of multiple geological survey objects into multiple object voice data, and preprocessing and recognizing them based on a preset geological language model to obtain the collected text of multiple objects specifically includes the following steps:

[0030] Based on multiple different geological survey objects, the multiple personnel voice data are divided into multiple object voice data;

[0031] Preprocessing of multiple object speech data with speech noise reduction, echo cancellation and silence removal is performed to obtain multiple preprocessed speech data;

[0032] Perform basic recognition on multiple preprocessed speech data to obtain multiple basic recognition texts;

[0033] Based on a pre-defined geological language model, semantic error correction and standardization are performed on multiple basic identification texts to obtain multiple object collection texts.

[0034] As a further limitation of the technical solution of this embodiment of the invention, the step of extracting multiple key information from multiple object collection texts, and filling in the multiple key information in a structured manner according to multiple preset object collection templates to generate multiple structured collection data specifically includes the following steps:

[0035] Extract multiple key pieces of information from the collected text of multiple objects;

[0036] According to a pre-built field mapping rule library, multiple key pieces of information are converted into multiple standardized attribute field values;

[0037] Match object acquisition templates corresponding to multiple geological survey objects;

[0038] According to the multiple object collection templates, the values ​​of multiple standardized attribute fields are structured and filled in to generate multiple structured collection data.

[0039] Establish a traceability association among multiple sets of structured data, multiple sets of object voice data, and multiple sets of location data.

[0040] A geological survey data speech structured acquisition system based on a large language model is proposed. The system includes a transmission network construction module, a speech relay module, a survey object identification module, a speech preprocessing and recognition module, and a structured mapping and input module, wherein:

[0041] The transmission network construction module is used to acquire the location data of multiple geological surveyors and the equipment signals of multiple acquisition devices when multiple personnel conduct field geological surveys simultaneously, select relay devices, and construct an acquisition and transmission network.

[0042] The voice relay transmission module is used to collect and aggregate the voice data of multiple geological survey personnel to the relay device through the acquisition and transmission network, and then transmit the data over the network through the relay device.

[0043] The survey object determination module is used to calculate the distance between multiple geological survey personnel based on multiple positioning data, compare and verify the multiple distances and the multiple voice data of the personnel, and determine multiple geological survey objects;

[0044] The speech preprocessing and recognition module is used to divide the speech data of multiple geological survey objects into multiple object speech data, and to preprocess and recognize the speech data of multiple objects based on a preset geological big language model to obtain the collected text of multiple objects.

[0045] The structured mapping and filling module is used to extract multiple key information from the collected text of multiple objects, and to perform structured mapping and filling of multiple key information according to the preset multiple object collection templates to generate multiple structured collection data.

[0046] As a further limitation of the technical solution of this embodiment of the invention, the transmission network construction module specifically includes:

[0047] Personnel identification unit, used to identify multiple geological survey personnel when multiple personnel are conducting field geological surveys simultaneously;

[0048] The real-time positioning unit is used to locate multiple geological survey personnel in real time and acquire multiple positioning data.

[0049] The signal acquisition unit is used to acquire equipment signals from the acquisition devices of multiple geological surveyors;

[0050] A relay device selection unit is used to compare signals from multiple devices and select a relay device from the multiple acquisition devices.

[0051] The transmission network construction unit is used to construct a data acquisition and transmission network that connects multiple other data acquisition devices adjacent to each other and converges to the relay device based on multiple positioning data.

[0052] As a further limitation of the technical solution of this embodiment of the invention, the survey object determination module specifically includes:

[0053] The distance calculation unit is used to calculate the distance between multiple geological survey personnel based on the multiple positioning data.

[0054] A distance comparison unit is used to compare multiple distances between each other with a preset distance threshold.

[0055] The object same determination unit is used to determine that the geological survey objects of two geological surveyors are the same when the distance between them is less than the distance threshold.

[0056] The segment extraction unit is used to extract two speech sample segments from the speech data of two geological surveyors when the distance between them is not less than a distance threshold.

[0057] The feature recognition and comparison unit is used to perform feature recognition and comparison on multiple speech sample segments to determine whether the speech description is the same geological survey object.

[0058] The object determination unit is used to record comparison and judgment results and determine multiple different geological survey objects.

[0059] As a further limitation of the technical solution of this embodiment of the invention, the structured mapping filling module specifically includes:

[0060] A key information extraction unit is used to extract multiple key information from the collected text of the multiple objects;

[0061] The field value conversion unit is used to convert multiple key information items into multiple standardized attribute field values ​​according to a pre-built field mapping rule library;

[0062] A template matching unit is used to match object acquisition templates corresponding to multiple geological survey objects;

[0063] The structured mapping filling unit is used to perform structured mapping filling on the values ​​of the standardized attribute fields according to the multiple object collection templates, thereby generating multiple structured collection data.

[0064] The tracing association establishment unit is used to establish tracing associations for multiple sets of structured collected data, multiple sets of object voice data, and multiple sets of positioning data.

[0065] Compared with the prior art, the beneficial effects of the present invention are:

[0066] This invention can construct a data acquisition and transmission network between multiple acquisition devices, collect and aggregate voice data from multiple individuals to a relay device, and then transmit it over the network. By comparing and verifying voice data from multiple individuals at different distances, multiple geological survey objects can be identified. The voice data from multiple individuals is divided, preprocessed, identified, key extracted, and structured and mapped into a structure, achieving collaborative communication and voice data aggregation. It can also avoid the problem of voice transmission interruption and effectively ensure the formation of unified and complete survey object data. Attached Figure Description

[0067] Figure 1 The flowchart illustrates a method for structured speech acquisition of geological survey data based on a large language model, as provided in an embodiment of the present invention.

[0068] Figure 2 The diagram illustrates the application architecture of the geological survey data voice structure acquisition system based on a large language model, as provided in an embodiment of the present invention. Detailed Implementation

[0069] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0070] Understandably, in existing technologies, geological survey voice acquisition usually only involves the voice acquisition and conversion processing of a single investigator. However, in actual geological surveys, multiple investigators collaborate to conduct the surveys. Furthermore, in field geological surveys, network signals are prone to instability. Existing technologies, which use a single acquisition device as an independent data source, lack collaborative communication and voice data aggregation mechanisms among multiple investigators. This results in voice data collected by multiple investigators being independent of each other and prone to voice transmission interruptions, making it difficult to form unified and complete survey data.

[0071] To address the aforementioned issues, this invention discloses a method and system for structured voice acquisition of geological survey data based on a large language model. When multiple personnel conduct simultaneous field geological surveys, the method acquires the location data of multiple geological surveyors and the device signals of multiple acquisition devices, selects a relay device, and constructs an acquisition and transmission network. Through the acquisition and transmission network, the voice data of multiple geological surveyors is collected and aggregated to the relay device, and then transmitted over the network. Based on multiple location data, the distances between multiple geological surveyors are calculated, and the distances and voice data of multiple personnel are compared and verified to identify multiple geological survey objects. Based on these multiple geological survey objects, the voice data of multiple objects is segmented, and preprocessed and recognized based on a preset geological large language model to obtain multiple object acquisition texts. Multiple key pieces of information are extracted from the multiple object acquisition texts, and structured mapping and filling of these key pieces of information are performed according to a preset multiple object acquisition template to generate multiple structured acquisition data. It can construct a data acquisition and transmission network between multiple acquisition devices, collect and aggregate voice data from multiple personnel to a relay device, and then transmit it over the network. By comparing and verifying voice data from multiple personnel at different distances, it can identify multiple geological survey objects, classify, preprocess, identify, extract key information, and fill in structured mapping data from multiple personnel, thereby achieving collaborative communication and voice data aggregation. It can also avoid the problem of voice transmission interruption and ensure the formation of unified and complete survey object data.

[0072] Specifically, Figure 1 The flowchart illustrates a method for structured speech acquisition of geological survey data based on a large language model, as provided in an embodiment of the present invention.

[0073] In a preferred embodiment of the present invention, a method for structured speech acquisition of geological survey data based on a large language model specifically includes the following steps:

[0074] Step S101: When multiple personnel are conducting field geological surveys simultaneously, acquire the location data of multiple geological survey personnel and the equipment signals of multiple acquisition devices, select relay devices, and construct an acquisition and transmission network.

[0075] In this embodiment of the invention, in a scenario where multiple personnel are conducting field geological surveys simultaneously, multiple geological surveyors are identified. By real-time positioning of these multiple geological surveyors, their positioning data is obtained, as well as the device signals of their data acquisition devices. Then, the multiple device signals are compared, and the device with the strongest signal is selected as the relay device. Based on the multiple positioning data, the relay device is used as the aggregation node, and multiple other data acquisition devices are used as access nodes. The adjacent access nodes with the shortest distance are connected via Bluetooth to construct a multi-level or star-shaped data acquisition and transmission network.

[0076] As can be understood, device signal refers to the signal status of the acquisition device used for data communication and network connection. It is used to characterize the current wireless network communication quality and data transmission capability of the acquisition device. The stronger the device signal, the better the network connection quality of the corresponding acquisition device, the fewer packet losses, delays and communication interruptions during network transmission, and the higher the data transmission capability.

[0077] It is understandable that the data acquisition and transmission network is established using a Bluetooth communication connection.

[0078] It is understandable that during the voice acquisition process, the device signal of the relay device is continuously monitored, and when the signal strength of the relay device drops below the preset switching threshold, the relay device is reselected according to the real-time communication quality.

[0079] Specifically, in another preferred embodiment provided by the present invention, the steps of acquiring the location data of multiple geological surveyors and the equipment signals of multiple acquisition devices, selecting relay devices, and constructing an acquisition and transmission network when multiple personnel are conducting field geological surveys simultaneously include the following steps:

[0080] When multiple personnel conduct field geological surveys simultaneously, multiple geological survey personnel should be identified;

[0081] Real-time location tracking of multiple geological survey personnel was performed to obtain multiple location data.

[0082] Acquire device signals from the data acquisition equipment of multiple geological surveyors;

[0083] The signals from multiple devices are compared, and a relay device is selected from the multiple acquisition devices;

[0084] Based on the multiple positioning data, a data acquisition and transmission network is constructed by converging multiple other acquisition devices adjacent to each other and converging to the relay device.

[0085] Furthermore, the method for structured speech acquisition of geological survey data based on a large language model also includes the following steps:

[0086] Step S102: Collect and aggregate the voice data of multiple geological survey personnel to the relay device through the acquisition and transmission network, and transmit the data over the network through the relay device.

[0087] In this embodiment of the invention, multiple acquisition devices are used to collect voice data from multiple geological survey personnel. During the voice data acquisition process, the voice data from multiple personnel is collected and transmitted to a relay device through a data acquisition and transmission network. Then, according to a preset network transmission address, the voice data from multiple personnel is transmitted over the network through the relay device.

[0088] It is understandable that the process of collecting and transmitting voice data from multiple people to a relay device via a data acquisition and transmission network is a Bluetooth transmission process that relies on the data acquisition and transmission network to collect the voice data from multiple people adjacently and step by step to the relay device.

[0089] Specifically, in another preferred embodiment provided by the present invention, the step of collecting and aggregating the voice data of multiple geological surveyors to the relay device through the acquisition and transmission network, and transmitting it over the network through the relay device, specifically includes the following steps:

[0090] Voice data of multiple geological survey personnel were collected using multiple acquisition devices.

[0091] Based on the acquisition and transmission network, the voice data of multiple individuals are collected and transmitted to the relay device;

[0092] The relay device is used to transmit the voice data of multiple individuals over a network.

[0093] Furthermore, the method for structured speech acquisition of geological survey data based on a large language model also includes the following steps:

[0094] Step S103: Calculate the distance between multiple geological survey personnel based on the multiple positioning data, compare and verify the multiple distances and the multiple personnel voice data, and determine multiple geological survey objects.

[0095] In this embodiment of the invention, the distances between multiple geological surveyors are calculated based on multiple positioning data. These distances are then compared with a preset distance threshold. If the distance between two geological surveyors is less than the distance threshold, it can be directly determined that the geological survey objects of the two geological surveyors are the same. However, if the distance between two geological surveyors is not less than the distance threshold, whether the geological survey objects are the same needs to be verified. In this case, according to a preset sample segment length, two voice sample segments are extracted from the voice data of the two geological surveyors. These two voice sample segments are then subjected to feature recognition and comparison to determine whether the voice descriptions refer to the same geological survey object. By recording the comparison and judgment results, multiple different geological survey objects are identified.

[0096] Understandably, the process of identifying and comparing features of two speech sample segments to determine whether the speech descriptions refer to the same geological survey object involves identifying the two speech sample segments, extracting semantic features such as geological names, lithological descriptions, stratigraphic descriptions, structural descriptions, location descriptions, and / or survey phenomena, and then performing feature matching and verification comparisons to determine whether the descriptions by different geological surveyors refer to the same geological survey object.

[0097] Specifically, in another preferred embodiment provided by the present invention, the step of calculating the distance between multiple geological survey personnel based on multiple positioning data, comparing and verifying the multiple distances and the multiple personnel voice data to determine multiple geological survey objects specifically includes the following steps:

[0098] Based on the aforementioned location data, the distance between multiple geological survey personnel is calculated;

[0099] The multiple intervals are compared with a preset distance threshold;

[0100] When the distance between two geological surveyors is less than a distance threshold, it is determined that the two geological surveyors are surveying the same geological object.

[0101] When the distance between two geological surveyors is not less than the distance threshold, two voice sample segments are extracted from the voice data of the two geological surveyors.

[0102] Multiple speech sample segments are subjected to feature recognition and comparison to determine whether the speech descriptions refer to the same geological survey object;

[0103] Record the comparison and judgment results to identify multiple different geological survey objects.

[0104] Furthermore, the method for structured speech acquisition of geological survey data based on a large language model also includes the following steps:

[0105] Step S104: Based on the multiple geological survey objects, divide the voice data of multiple objects, and preprocess and recognize them based on the preset geological language model to obtain the collected text of multiple objects.

[0106] In this embodiment of the invention, multiple personnel voice data are classified and divided into multiple object voice data based on multiple different geological survey objects. Then, the multiple object voice data are preprocessed with voice noise reduction, echo cancellation and silence removal to obtain multiple preprocessed voice data. After that, the multiple preprocessed voice data are subjected to basic recognition to obtain multiple basic recognition texts. Then, based on a preset geological big language model, the multiple basic recognition texts are subjected to semantic error correction and standardization processing to obtain multiple object collection texts.

[0107] Understandably, the pre-set geological language model is a large language model optimized and trained for geological professional terms. It supports the accurate recognition of professional terms such as stratigraphic names, rock names, mineral names, mineral deposit names, geological structure terms, geophysical exploration terms, geochemical exploration terms, remote sensing geological terms, hydrogeological terms, environmental geological terms, and geological disaster terms. It can also perform semantic error correction and standardization on professional terms, homophones, and colloquial expressions in the basic recognition text.

[0108] Specifically, in another preferred embodiment provided by the present invention, the step of dividing the voice data of multiple geological survey objects into multiple object voice data, and performing preprocessing and recognition based on a preset geological language model to obtain the collected text of multiple objects specifically includes the following steps:

[0109] Based on multiple different geological survey objects, the multiple personnel voice data are divided into multiple object voice data;

[0110] Preprocessing of multiple object speech data with speech noise reduction, echo cancellation and silence removal is performed to obtain multiple preprocessed speech data;

[0111] Perform basic recognition on multiple preprocessed speech data to obtain multiple basic recognition texts;

[0112] Based on a pre-defined geological language model, semantic error correction and standardization are performed on multiple basic identification texts to obtain multiple object collection texts.

[0113] Furthermore, the method for structured speech acquisition of geological survey data based on a large language model also includes the following steps:

[0114] Step S105: Extract multiple key information from the multiple object collection texts, and fill in the multiple key information in a structured mapping according to the preset multiple object collection templates to generate multiple structured collection data.

[0115] In this embodiment of the invention, multiple key information is extracted from the collected text of multiple objects, and then converted into multiple standardized attribute field values ​​according to a pre-built field mapping rule library. These key information values ​​are matched with the object collection templates corresponding to multiple geological survey objects, and then the standardized attribute field values ​​are structured and mapped and filled in according to the multiple object collection templates to generate multiple structured collection data. Furthermore, a traceability association is established between the multiple structured collection data, multiple object voice data, and multiple location data to ensure that each structured field in the multiple structured collection data can be traced back to the corresponding voice source and location.

[0116] Understandably, multiple key pieces of information include basic information about multiple geological survey objects, location parameters, lithological information, stratigraphic information, structural information, occurrence information, and / or geological phenomena.

[0117] Understandably, the pre-built field mapping rule library contains the mapping relationship between each form field and standardized attribute values. For the extracted key information, the corresponding form field is first matched, and then the key information is converted into standardized attribute field values ​​according to the field mapping rule library. For enumerated fields, the extracted key information is matched with the field's preset standardized attribute values, and the standardized value with the highest matching degree is selected and filled in. For numeric fields, the extracted numeric information is format-validated and unit-converted before being filled in. The converted standardized field values ​​are automatically filled into the corresponding field options of the data collection form, and the filling results are displayed in real time on the data collection terminal interface.

[0118] Specifically, in another preferred embodiment provided by the present invention, the step of extracting multiple key pieces of information from multiple object collection texts, and filling in the multiple key pieces of information in a structured manner according to multiple preset object collection templates to generate multiple structured collection data specifically includes the following steps:

[0119] Extract multiple key pieces of information from the collected text of multiple objects;

[0120] According to a pre-built field mapping rule library, multiple key pieces of information are converted into multiple standardized attribute field values;

[0121] Match object acquisition templates corresponding to multiple geological survey objects;

[0122] According to the multiple object collection templates, the values ​​of multiple standardized attribute fields are structured and filled in to generate multiple structured collection data.

[0123] Establish a traceability association among multiple sets of structured data, multiple sets of object voice data, and multiple sets of location data.

[0124] Furthermore, Figure 2 The diagram illustrates the application architecture of the geological survey data voice structure acquisition system based on a large language model, as provided in an embodiment of the present invention.

[0125] Specifically, in another preferred embodiment provided by the present invention, the geological survey data speech structure acquisition system based on a large language model includes:

[0126] The transmission network construction module 101 is used to acquire the location data of multiple geological surveyors and the equipment signals of multiple acquisition devices when multiple personnel conduct field geological surveys simultaneously, select relay devices, and construct an acquisition and transmission network.

[0127] In this embodiment of the invention, in a scenario where multiple personnel are conducting field geological surveys simultaneously, the transmission network construction module 101 identifies multiple geological survey personnel, obtains their location data by real-time positioning, and acquires the device signals of their data acquisition devices. Then, it compares the multiple device signals, selects the acquisition device with the strongest signal from among the multiple acquisition devices, and identifies it as a relay device. Based on the multiple location data, with the relay device as the aggregation node and multiple other acquisition devices as access nodes, the adjacent access nodes with the shortest distance are connected via Bluetooth to construct a multi-level or star-shaped data acquisition and transmission network.

[0128] Specifically, in another preferred embodiment provided by the present invention, the transmission network construction module 101 specifically includes:

[0129] Personnel identification unit, used to identify multiple geological survey personnel when multiple personnel are conducting field geological surveys simultaneously;

[0130] The real-time positioning unit is used to locate multiple geological survey personnel in real time and acquire multiple positioning data.

[0131] The signal acquisition unit is used to acquire equipment signals from the acquisition devices of multiple geological surveyors;

[0132] A relay device selection unit is used to compare signals from multiple devices and select a relay device from the multiple acquisition devices.

[0133] The transmission network construction unit is used to construct a data acquisition and transmission network that connects multiple other data acquisition devices adjacent to each other and converges to the relay device based on multiple positioning data.

[0134] Furthermore, the geological survey data speech structure acquisition system based on a large language model also includes:

[0135] The voice relay transmission module 102 is used to collect and aggregate the voice data of multiple geological survey personnel to the relay device through the acquisition and transmission network, and then transmit the data over the network through the relay device.

[0136] In this embodiment of the invention, the voice relay transmission module 102 collects voice data of multiple geological survey personnel through multiple acquisition devices. During the voice data acquisition process, the voice data of multiple personnel is collected and transmitted to the relay device through the acquisition and transmission network. Then, according to the preset network transmission address, the voice data of multiple personnel is transmitted over the network through the relay device.

[0137] The survey object determination module 103 is used to calculate the distance between multiple geological survey personnel based on multiple positioning data, compare and verify the multiple distances and the multiple personnel voice data, and determine multiple geological survey objects.

[0138] In this embodiment of the invention, the survey object determination module 103 calculates the distance between multiple geological surveyors based on multiple positioning data, and then compares the multiple distances with a preset distance threshold. If the distance between two geological surveyors is less than the distance threshold, it can be directly determined that the geological survey objects of the two geological surveyors are the same. However, if the distance between two geological surveyors is not less than the distance threshold, whether the geological survey objects are the same needs to be verified. At this time, according to the preset sample segment length, two voice sample segments are extracted from the voice data of the two geological surveyors, and then feature recognition and comparison are performed on the two voice sample segments to determine whether the voice description is the same geological survey object. By recording the comparison and judgment results, multiple different geological survey objects are determined.

[0139] Specifically, in another preferred embodiment provided by the present invention, the survey subject determination module 103 specifically includes:

[0140] The distance calculation unit is used to calculate the distance between multiple geological survey personnel based on the multiple positioning data.

[0141] A distance comparison unit is used to compare multiple distances between each other with a preset distance threshold.

[0142] The object same determination unit is used to determine that the geological survey objects of two geological surveyors are the same when the distance between them is less than the distance threshold.

[0143] The segment extraction unit is used to extract two speech sample segments from the speech data of two geological surveyors when the distance between them is not less than a distance threshold.

[0144] The feature recognition and comparison unit is used to perform feature recognition and comparison on multiple speech sample segments to determine whether the speech description is the same geological survey object.

[0145] The object determination unit is used to record comparison and judgment results and determine multiple different geological survey objects.

[0146] Furthermore, the geological survey data speech structure acquisition system based on a large language model also includes:

[0147] The speech preprocessing and recognition module 104 is used to divide the speech data of multiple geological survey objects into multiple object speech data, and perform preprocessing and recognition based on a preset geological language model to obtain multiple object collected texts.

[0148] In this embodiment of the invention, the speech preprocessing and recognition module 104 classifies multiple personnel speech data into multiple object speech data based on multiple different geological survey objects as classification criteria. Then, it performs speech noise reduction, echo cancellation, and silence removal preprocessing on the multiple object speech data to obtain multiple preprocessed speech data. After that, it performs basic recognition on the multiple preprocessed speech data to obtain multiple basic recognition texts. Then, based on a preset geological big language model, it performs semantic error correction and standardization processing on the multiple basic recognition texts to obtain multiple object collection texts.

[0149] The structured mapping and filling module 105 is used to extract multiple key information from the multiple object collection texts, and to perform structured mapping and filling on the multiple key information according to the preset multiple object collection templates to generate multiple structured collection data.

[0150] In this embodiment of the invention, the structured mapping filling module 105 extracts multiple key information from multiple object collection texts, converts the multiple key information into multiple standardized attribute field values ​​according to a pre-built field mapping rule library, matches the object collection templates corresponding to multiple geological survey objects, and then performs structured mapping filling on the multiple standardized attribute field values ​​according to the multiple object collection templates to generate multiple structured collection data. It also establishes a traceability association between multiple structured collection data, multiple object voice data, and multiple location data to ensure that each structured field in the multiple structured collection data can be traced back to the corresponding voice source and location.

[0151] Specifically, in another preferred embodiment provided by the present invention, the structured mapping filling module 105 specifically includes:

[0152] A key information extraction unit is used to extract multiple key information from the collected text of the multiple objects;

[0153] The field value conversion unit is used to convert multiple key information items into multiple standardized attribute field values ​​according to a pre-built field mapping rule library;

[0154] A template matching unit is used to match object acquisition templates corresponding to multiple geological survey objects;

[0155] The structured mapping filling unit is used to perform structured mapping filling on the values ​​of the standardized attribute fields according to the multiple object collection templates, thereby generating multiple structured collection data.

[0156] The tracing association establishment unit is used to establish tracing associations for multiple sets of structured collected data, multiple sets of object voice data, and multiple sets of positioning data.

[0157] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A method for structured speech acquisition of geological survey data based on a large language model, characterized in that, The method specifically includes the following steps: When multiple personnel conduct field geological surveys simultaneously, the system acquires the location data of multiple geological survey personnel and the equipment signals of multiple acquisition devices, selects relay equipment, and constructs an acquisition and transmission network. The voice data of multiple geological surveyors is collected and aggregated through the acquisition and transmission network and transmitted over the network via the relay equipment. Based on multiple positioning data, the distance between multiple geological survey personnel is calculated, and the distance between multiple personnel and the voice data of multiple personnel are compared and verified to identify multiple geological survey objects; Based on multiple geological survey objects, the voice data of multiple objects are divided, and preprocessed and recognized based on a preset geological language model to obtain the collected text of multiple objects. Extract multiple key pieces of information from the collected text of multiple objects, and fill them in according to the preset multiple object collection templates to generate multiple structured collection data.

2. The method for structured speech acquisition of geological survey data based on a large language model according to claim 1, characterized in that, When multiple personnel conduct simultaneous field geological surveys, acquiring location data from multiple geological surveyors and device signals from multiple acquisition devices, selecting relay devices, and constructing an acquisition and transmission network specifically includes the following steps: When multiple personnel conduct field geological surveys simultaneously, multiple geological survey personnel should be identified; Real-time location tracking of multiple geological survey personnel was performed to obtain multiple location data. Acquire device signals from the data acquisition equipment of multiple geological surveyors; The signals from multiple devices are compared, and a relay device is selected from the multiple acquisition devices; Based on the multiple positioning data, a data acquisition and transmission network is constructed by converging multiple other acquisition devices adjacent to each other and converging to the relay device.

3. The method for structured speech acquisition of geological survey data based on a large language model according to claim 1, characterized in that, The process of collecting and aggregating voice data from multiple geological surveyors through the acquisition and transmission network to the relay device, and then transmitting it over the network through the relay device, specifically includes the following steps: Voice data of multiple geological survey personnel were collected using multiple acquisition devices. Based on the acquisition and transmission network, the voice data of multiple individuals are collected and transmitted to the relay device; The relay device is used to transmit the voice data of multiple individuals over a network.

4. The method for structured speech acquisition of geological survey data based on a large language model according to claim 1, characterized in that, The step of calculating the distance between multiple geological survey personnel based on multiple location data, comparing and verifying the multiple distances and the voice data of the multiple personnel, and determining the multiple geological survey targets specifically includes the following steps: Based on the aforementioned location data, the distance between multiple geological survey personnel is calculated; The multiple intervals are compared with a preset distance threshold; When the distance between two geological surveyors is less than a distance threshold, it is determined that the two geological surveyors are surveying the same geological object. When the distance between two geological surveyors is not less than the distance threshold, two voice sample segments are extracted from the voice data of the two geological surveyors. Multiple speech sample segments are subjected to feature recognition and comparison to determine whether the speech descriptions refer to the same geological survey object; Record the comparison and judgment results to identify multiple different geological survey objects.

5. The method for structured speech acquisition of geological survey data based on a large language model according to claim 1, characterized in that, The process of dividing the speech data of multiple geological survey objects into multiple object segments, and preprocessing and recognizing them based on a preset geological language model to obtain the collected text of multiple objects specifically includes the following steps: Based on multiple different geological survey objects, the multiple personnel voice data are divided into multiple object voice data; Preprocessing of multiple object speech data with speech noise reduction, echo cancellation and silence removal is performed to obtain multiple preprocessed speech data; Perform basic recognition on multiple preprocessed speech data to obtain multiple basic recognition texts; Based on a pre-defined geological language model, semantic error correction and standardization are performed on multiple basic identification texts to obtain multiple object collection texts.

6. The method for structured speech acquisition of geological survey data based on a large language model according to claim 1, characterized in that, The process of extracting multiple key pieces of information from multiple object-collected texts, and then mapping and filling these key pieces of information into a structured manner according to preset object-collected templates to generate multiple structured collected data specifically includes the following steps: Extract multiple key pieces of information from the collected text of multiple objects; According to a pre-built field mapping rule library, multiple key pieces of information are converted into multiple standardized attribute field values; Match object acquisition templates corresponding to multiple geological survey objects; According to the multiple object collection templates, the values ​​of multiple standardized attribute fields are structured and filled in to generate multiple structured collection data. Establish a traceability association among multiple sets of structured data, multiple sets of object voice data, and multiple sets of location data.

7. A geological survey data speech structured acquisition system based on a large language model, characterized in that, The system includes a transmission network construction module, a voice relay transmission module, a survey subject identification module, a voice preprocessing and recognition module, and a structured mapping and input module, wherein: The transmission network construction module is used to acquire the location data of multiple geological surveyors and the equipment signals of multiple acquisition devices when multiple personnel conduct field geological surveys simultaneously, select relay devices, and construct an acquisition and transmission network. The voice relay transmission module is used to collect and aggregate the voice data of multiple geological survey personnel to the relay device through the acquisition and transmission network, and then transmit the data over the network through the relay device. The survey object determination module is used to calculate the distance between multiple geological survey personnel based on multiple positioning data, compare and verify the multiple distances and the multiple voice data of the personnel, and determine multiple geological survey objects; The speech preprocessing and recognition module is used to divide the speech data of multiple geological survey objects into multiple object speech data, and to preprocess and recognize the speech data of multiple objects based on a preset geological big language model to obtain the collected text of multiple objects. The structured mapping and filling module is used to extract multiple key information from the collected text of multiple objects, and to perform structured mapping and filling of multiple key information according to the preset multiple object collection templates to generate multiple structured collection data.

8. The geological survey data speech structure acquisition system based on a large language model according to claim 7, characterized in that, The transmission network construction module specifically includes: Personnel identification unit, used to identify multiple geological survey personnel when multiple personnel are conducting field geological surveys simultaneously; The real-time positioning unit is used to locate multiple geological survey personnel in real time and acquire multiple positioning data. The signal acquisition unit is used to acquire equipment signals from the acquisition devices of multiple geological surveyors; A relay device selection unit is used to compare signals from multiple devices and select a relay device from the multiple acquisition devices. The transmission network construction unit is used to construct a data acquisition and transmission network that connects multiple other data acquisition devices adjacent to each other and converges to the relay device based on multiple positioning data.

9. The geological survey data speech structure acquisition system based on a large language model according to claim 7, characterized in that, The survey subject identification module specifically includes: The distance calculation unit is used to calculate the distance between multiple geological survey personnel based on the multiple positioning data. A distance comparison unit is used to compare multiple distances between each other with a preset distance threshold. The object same determination unit is used to determine that the geological survey objects of two geological surveyors are the same when the distance between them is less than a distance threshold. The segment extraction unit is used to extract two speech sample segments from the speech data of two geological surveyors when the distance between them is not less than a distance threshold. The feature recognition and comparison unit is used to perform feature recognition and comparison on multiple speech sample segments to determine whether the speech description is the same geological survey object. The object determination unit is used to record comparison and judgment results and determine multiple different geological survey objects.

10. The geological survey data speech structure acquisition system based on a large language model according to claim 7, characterized in that, The structured mapping input module specifically includes: A key information extraction unit is used to extract multiple key information from the collected text of the multiple objects; The field value conversion unit is used to convert multiple key information items into multiple standardized attribute field values ​​according to a pre-built field mapping rule library; A template matching unit is used to match object acquisition templates corresponding to multiple geological survey objects; The structured mapping filling unit is used to perform structured mapping filling on the values ​​of the standardized attribute fields according to the multiple object collection templates, thereby generating multiple structured collection data. The tracing association establishment unit is used to establish tracing associations for multiple sets of structured collected data, multiple sets of object voice data, and multiple sets of positioning data.

Images