Geological survey project whole-process management system based on workflow engine

Abstract

The application discloses a geological survey project whole-process management system based on a workflow engine, relates to the field of automatic control of geological and mineral exploration, and comprises a central management control platform, a central workflow scheduling center, a cross-agency collaboration gateway and a distributed business database which are internally arranged, a field operation sensing subsystem, a portable geological survey terminal which is integrated with a double-frequency multi-constellation GNSS receiver board card, an intelligent checking calculation base which comprises a geological semantic checking engine and a multi-source spatial data fusion module, a resource and monitoring feedback subsystem which comprises a laboratory resource scheduler, and asynchronous communication between the central management control platform and the intelligent checking calculation base through a message queue bus.The application realizes real-time strong constraints of a process on quality through an evaluation system, solves the problem of inconsistent field positioning accuracy by using a BursaWolf model, and realizes optimal allocation of resources in the whole life cycle of a geological survey project through an incentive type scheduling index.

Description

Technical Field

[0001] This application relates to the field of automated control in geological and mineral exploration, and in particular to a workflow engine-based full-process management system for geological survey projects. Background Technology

[0002] Currently, my country's geological surveys have fully entered the information age. However, in the entire lifecycle management of "project initiation - field data collection - laboratory testing - results submission," traditional workflow engines (such as systems based on the BPMN 2.0 specification) only focus on the flow of task status (i.e., "whoever completes it sends it to whom"), failing to perceive the authenticity and compliance of the business data itself. This leads to the following technical shortcomings:

[0003] 1. Traditional systems cannot identify logical conflicts in the entered text (such as describing Quaternary lithology in the Sinian System), resulting in a large amount of low-quality data entering the expensive post-detection stage.

[0004] 2. The multipath effect in the field causes GPS coordinate shifts, and the lack of a real-time coordinate adjustment and correction mechanism leads to topological disorder of sample points on standard scale maps.

[0005] 3. The allocation of external laboratory resources mainly relies on manual contact and lacks a priority linkage mechanism based on data quality scoring, which results in high-quality data not being processed first and seriously affects the exploration cycle. Summary of the Invention

[0006] This application provides a full-process management system for geological survey projects based on a workflow engine, which aims to achieve full-volume, transparent supervision of geological survey data by constructing a closed loop that deeply couples "quality monitoring" and "process-driven" approaches.

[0007] The geological survey project end-to-end management system based on a workflow engine provided in this application adopts the following technical solution:

[0008] A workflow engine-based end-to-end management system for geological survey projects, including:

[0009] The central management and control platform has a built-in central workflow scheduling center, cross-agency collaborative gateway, and distributed business database; the field operation perception subsystem includes a portable geological survey terminal, which integrates a dual-frequency multi-constellation GNSS receiver board.

[0010] The intelligent verification calculation base includes a geological semantic verification engine and a multi-source spatial data fusion module;

[0011] The resource and monitoring feedback subsystem includes a laboratory resource scheduler;

[0012] Its connection and assembly relationship is as follows: the portable geological survey terminal is bidirectionally connected to the cross-institutional collaborative gateway through an encrypted 4G / 5G private network;

[0013] The central management and control platform communicates asynchronously with the intelligent verification computing base via a message queue bus;

[0014] The cross-institutional collaboration gateway connects to the collaboration laboratory server based on the OAuth2.0 protocol; the central workflow scheduling center integrates a dynamic flow operator, which is used to calculate the quality evaluation function fed back by the geological semantic verification engine. and according to The calculation results generate state machine transition instructions that drive the laboratory resource scheduler in real time.

[0015] Preferably, the evaluation function ) is defined as:

[0016]

[0017] in: This is the overall quality assessment score for the current node. For spatial accuracy score, For the geological semantic consistency score, To evaluate the spatial sampling information entropy score of the uniformity of sampling point distribution, The weighting coefficients are satisfied. foot.

[0018] Preferably, Spatial accuracy score , where RMSE is the root mean square error of the sampling points relative to the standard scale base map. This is the maximum deviation threshold required by the standard.

[0019] Preferably, the geological semantic verification engine calculates through the following steps. :

[0020] The pre-trained GeoBERT model is used to map the “stratigraphic-structural-lithological” description text uploaded by the terminal into d-dimensional feature vectors. Where d takes the value of 768; the GeoBERT model adopts the BERTBase architecture, including 12 Transformer encoder layers, 768 hidden layer dimensions, and 12 self-attention heads. Furthermore, the model is pre-trained on a geoscience corpus using a masked language model (MLM) with no fewer than 1 million geoscience sentences, and retrieves standard feature vectors for corresponding layers from the built-in geological knowledge graph. ;calculate and cosine similarity as :

[0021] .

[0022] Preferably, the spatial sampling information entropy score The following normalization formula is used for calculation:

[0023]

[0024] In the formula: The percentage of sample point density within the j-th grid cell, where n is the total number of grid cells; this is achieved by dividing by a factor. make sure The values ​​range from 0 to 1 for different grid sizes.

[0025] Preferably, the multi-source spatial data fusion module is equipped with a CGCS2000 adjustment correction unit. This unit reads the reference layer data, uses the BursaWolf seven-parameter model to solve the translation, rotation and scale change parameters, and performs a national standard geodetic coordinate system transformation on the original GPS coordinates.

[0026] Preferably, the laboratory resource scheduler calculates the comprehensive scheduling index value. The formula is:

[0027]

[0028] in, This is the priority weight coefficient. The overall quality assessment score for the current node is given, where L is the total measured weight of the currently submitted samples, and J is the current equipment load rate of the laboratory. The average test cycle time per unit weight of sample. This is a preset time compensation constant.

[0029] Preferably, the determination logic of the dynamic flow operator is as follows: when At that time, the system automatically sends a test delivery handshake command to the external laboratory; when When this happens, the system suspends the current task and pushes it to the supervisor's manual review queue; when At that time, the system immediately locks the terminal input interface and issues a rectification warning.

[0030] Preferably, the central workflow scheduling center has a dynamic backtracking function. When the subsequent detection results show that there is a systematic error in the original data, a traceability index is established through the unique identifier of the data stream (UUID). This supports reverse freezing of batch data with systematic errors. The specific steps of the reverse freezing are as follows: the system retrieves the distributed business database through the unique identifier of the data stream (UUID), changes the compliance status bit of the corresponding record from "normal" to "suspended", and synchronously triggers the workflow engine to send an interception command to the collaborative laboratory server to stop the testing task of the batch of samples.

[0031] Preferably, the central management and control platform also integrates a blockchain-based rights confirmation module to verify the ownership of each workflow transfer. The results and corresponding responsible person information are calculated into the MerkleTree root hash and stored on the blockchain. The leaf nodes of the MerkleTree are composed of UUID and quality score. The concatenated data of operator identification and timestamp is generated by SHA256 hash operation.

[0032] In summary, this application includes the following beneficial technical effects:

[0033] 1. Achieved comprehensive quality audit of geological survey data: This application constructs a comprehensive quality evaluation function. This system overcomes the limitations of traditional workflow engines that only focus on task flow status (state-based), enabling real-time quality sensing of field-collected data (spatial, semantic, and distribution). Utilizing the GeoBERT deep learning model and spatial information entropy operator, the system can identify logical conflicts such as "Quaternary lithology appears in the Sinian strata" during the geological description stage, intercepting over 85% of low-quality data at the source and significantly reducing the costly rework rate in later testing.

[0034] 2. By integrating the CGCS2000 adjustment correction unit (321) into the intelligent verification calculation base (300), this application utilizes the BursaWolf seven-parameter model to solve the coordinate offset problem caused by multipath effect interference in field operations. By performing a national standard geodetic coordinate system transformation on the original GPS coordinates, the topological logic rigor of the sample points on the standard scale base map is ensured, and the accuracy is improved to the centimeter level.

[0035] 3. An incentive-based resource scheduling index has been introduced. Quality score The algorithm is directly coupled into the laboratory queuing system. This incentive mechanism, which rewards higher quality samples with faster testing, allows high-quality, high-confidence exploration samples to be prioritized for testing. Without adding extra equipment, the overall testing and feedback cycle of geological survey projects is shortened by approximately 45%.

[0036] 4. By coupling the UUID traceability index with the blockchain rights confirmation module, this application associates the flow result of each workflow node with the MerkleTree root hash on the blockchain. This not only solves the problem of determining the responsibilities of "who collects, who evaluates, and who is responsible" in the geological survey process, but also achieves rapid interception of systematic error data through the "reverse freezing" function, ensuring the legal validity of the submitted results. Attached Figure Description

[0037] Figure 1 This is a schematic diagram illustrating the overall principle and flow of the geological survey full-process management method of this application;

[0038] Figure 2 This is a modular connection and assembly structure diagram of the management system of this application, which shows the hierarchical relationship between the central management and control platform, the field operation perception subsystem, the intelligent verification and calculation base and the resource and monitoring feedback subsystem.

[0039] Figure 3 The quality evaluation function for this application Multi-dimensional computational logic flowchart;

[0040] Figure 4 This application provides an index for the dynamic scheduling of laboratory resources. A schematic diagram of the calculation flow.

[0041] Attached label: 100, Central Management and Control Platform; 110, Central Workflow Scheduling Center; 120, Cross-Agency Collaboration Gateway; 130, Distributed Business Database;

[0042] 200. On-site operation sensing subsystem; 210. Portable geological survey terminal;

[0043] 300. Intelligent verification calculation base; 310. Geological semantic verification engine; 311. Geological knowledge graph library; 320. Multi-source spatial data fusion module; 321. CGCS2000 adjustment correction unit;

[0044] 400. Resource and Monitoring Feedback Subsystem; 410. Laboratory Resource Scheduler. Detailed Implementation

[0045] The following is in conjunction with the appendix Figure 1 - Figure 4 This application will be described in further detail.

[0046] This application discloses a full-process management system for geological survey projects based on a workflow engine.

[0047] Reference Figure 1The geological survey project end-to-end management system based on a workflow engine includes: a central management and control platform 100, which has a built-in central workflow scheduling center 110, a cross-agency collaboration gateway 120, and a distributed business database 130.

[0048] The central management and control platform 100 is the logical core of the system, integrating a central workflow scheduling center 110, a cross-agency collaboration gateway 120, and a distributed business database 130. The central workflow scheduling center 110 is responsible for the global maintenance of business status and the distribution of dynamic workflow instructions. The cross-agency collaboration gateway 120 has heterogeneous message parsing capabilities, using a pre-built JSONSchema template to convert front-end GeoJSON format data into test application forms recognizable by the laboratory system LIMS. The distributed business database 130 is used to store business logs and blockchain hash values.

[0049] The field operation perception subsystem 200 includes a portable geological survey terminal 210. The portable geological survey terminal 210 integrates a GNSS receiver board with multi-satellite, multi-frequency centimeter-level positioning accuracy. It also integrates an offline GIS engine and a BeiDou short message module. This type of multi-satellite, multi-frequency GNSS receiver can achieve instantaneous centimeter-level positioning accuracy, adapting to handheld geological survey scenarios in complex environments. In geological operations, its high-precision positioning capabilities provide reliable support for data acquisition. The GNSS receiver board outputs raw NMEA format coordinate data. The portable geological survey terminal 210 has a built-in board based on a fully integrated SoC chip, supporting L1 / L2 dual-frequency signal reception and simultaneously processing BDS, GPS, and GLONASS constellation data. Connection to the collaborative laboratory server is achieved through an OAuth 2.0 standard authorization process. Specifically, the cross-institutional collaborative gateway 120 requests an authorization code from the laboratory server, and only after obtaining an AccessToken can asynchronous data transmission based on HTTPS be performed. The portable geological survey terminal 210 accesses the dedicated power / geological access point (APN) provided by the operator through its built-in 5G communication module. At the transport layer, it employs the national standard SM4 symmetric encryption algorithm for end-to-end encryption of service packets and constructs a logically isolated dedicated transmission tunnel based on 5G network slicing technology to ensure the interception of unauthorized access to field-collected data during the backhaul process.

[0050] The connection between the cross-institutional collaboration gateway 120 and the collaborative laboratory server follows the OAuth 2.0 authorization code model. The gateway first requests an authorization code from the laboratory authentication server. After obtaining a trusted AccessToken, it encapsulates it in the request header of the RESTful interface to achieve controlled access to the laboratory's LIMS system resources.

[0051] The on-site operation perception subsystem 200 serves as the front-end data source.

[0052] The intelligent verification calculation base 300 includes a geological semantic verification engine 310 and a multi-source spatial data fusion module 320. The multi-source spatial data fusion module 320 contains a CGCS2000 adjustment correction unit 321. The adjustment correction unit 321 uses a BursaWolf seven-parameter model to correct the original coordinates, eliminating drift caused by terrain occlusion (typically around 515 meters), and ensuring the topological consistency of the sample points within the national standard CGCS2000 coordinate system.

[0053] The BursaWolf rotation matrix R used in the CGCS2000 adjustment correction unit 321 is:

[0054]

[0055] The coordinate transformation model is as follows: Where T is the translation vector. The position vector rotation convention is used here, with the rotation angle... Defined as follows: when viewed from the origin in the positive direction of each axis, clockwise rotation is considered positive. The rotation matrix R uses an antisymmetric structure for small-angle approximation simplification to improve the real-time correction speed of the computational base.

[0056] Among them, rotation angle Using radians as the unit, a clockwise rotation is defined as positive when viewed from the origin towards the positive direction of the coordinate axis. This deterministic matrix definition eliminates the sign mirroring issue that may occur during coordinate transformations in different software environments.

[0057] Reference Figure 3 The system performs automated auditing through the intelligent verification calculation base 300.

[0058] Spatial accuracy score The calculations rely on the CGCS2000 adjustment correction unit 321. This unit reads the baseline layer data and uses the BursaWolf seven-parameter model.

[0059]

[0060] In the formula: These represent the three translation parameters of the BursaWolf model. The three rotation angle parameters represent the Bursa-Wolf model, and k represents the scale factor of the Bursa-Wolf model. Using this formula, the system can correct the original drift points to the national standard coordinate system and calculate the RMSE value relative to the base map. The position vector rotation convention used in this embodiment conforms to the ISO 19111 standard, and its rotation angle... Defined as: when viewed from the origin in the positive directions of each axis, clockwise rotation is considered positive. In its linearized adjustment matrix, the element in the first row and second column is... The element in the first row and third column is .

[0061] Geological semantic score The model was derived using the GeoBERT model from the geological semantic verification engine 310. The GeoBERT model employed pre-trained parameters based on SciBERT during the initialization phase. The training corpus consisted of approximately 120 million characters from national regional geological survey reports and mining area prospectus reports from the past 10 years. In the input layer, strata, structure, and lithology were used as triples, with a maximum sequence length of 128. During the fine-tuning phase, entity link enhancement was performed using the geological knowledge graph 311, enabling the model to recognize the spatiotemporal exclusivity between the "Sinian System" and the "Cambrian System." This model maps text to feature vectors. And compared with the standard triplet vector in the Geological Knowledge Graph 311 Calculate the cosine similarity.

[0062] The GeoBERT model employs a masked language modeling (MLM) task during the pretraining phase, randomly masking 15% of the tokens in the geoscientific text. During the fine-tuning phase, it uses entity relationships from the geological knowledge graph 311 to refine the hidden layer feature vectors. Apply semantic constraints to ensure the feature space Similar strata nodes cluster together, while dissimilar nodes separate.

[0063] Spatial sampling information entropy The uniformity of the sample point distribution is then quantified using the Shannon entropy formula.

[0064] The resource and monitoring feedback subsystem 400 includes a laboratory resource scheduler 410. The laboratory resource scheduler 410 connects to external laboratories via the cross-institutional collaboration gateway 120 using the OAuth2.0 protocol to capture laboratory load information in real time.

[0065] During system operation, the central workflow scheduling center 110 integrates a dynamic flow operator. This operator calls the intelligent verification calculation base 300 to calculate the comprehensive quality evaluation function. This determines the direction of the business process. The evaluation function is defined as: .

[0066] in, This represents the overall quality assessment score of the current node, where the index i is a natural number and indicates the unique time-series index of the business node in the current workflow. These are the core parameters that drive the workflow state machine transitions. Spatial accuracy score, used to quantify the accuracy after adjustment correction of sample point coordinates. Denotes the weight coefficients of each indicator in the evaluation function and satisfies .

[0067] Through formula The root mean square error (RMSE) represents the average deviation of the measured points relative to the standard base map. It is the maximum error range allowed by industry standards, such as 15 meters.

[0068] The geological semantic consistency score is used to quantify the matching degree between the field-entered descriptions and the standard geological knowledge graph. The geological semantic verification engine 310 uses a pre-trained GeoBERT model to map the "strata-lithology" description text entered by the investigators into high-dimensional feature vectors. And compared with the standard triplet vector in the Geological Knowledge Graph 311 Perform cosine similarity calculation. This method can automatically identify logical contradictions, such as misrecording Cenozoic features in the Cambrian system description.

[0069] The spatial sampling information entropy score is used to assess the scientific validity of the uniformity of sample point distribution within a geographic grid. Its formula is: This parameter introduces the concept of entropy from information theory to evaluate the scientific validity of the sampling point placement. If the sampling points are too clustered in space, it indicates a high probability distribution. Unevenness will decrease the entropy value, thus lowering the overall score.

[0070] In this embodiment, the system's mathematical model is based on dimensional homogeneity. The output of the comprehensive quality evaluation function F(Ni) is the normalized dimensionless score. The spatial accuracy score Qgeo is based on the formula... The calculation stipulates that when the root mean square error (RMSE) is greater than or equal to the maximum deviation threshold (Dtol), the score is forcibly set to 0. The geological semantic consistency score (Qsem) is calculated using the cosine angle between the feature vector (Vobs) and the standard vector (Vstd), i.e. This value is located in the interval [−1,1], and in this system, its positive value is used in the calculation.

[0071] Laboratory resource scheduler 410 calculates the comprehensive scheduling index value The formula is:

[0072]

[0073] in, This represents the overall priority index value for laboratory resource scheduling; a larger value indicates a higher priority. This represents the prior priority weighting coefficient of the project; J represents the overall quality assessment score of the current business node i; J represents the current equipment load rate of the laboratory (range). The value is determined by the ratio of "number of tasks in progress / maximum throughput of equipment" fed back in real time by the collaborative laboratory server; L represents the measured total weight of the currently submitted samples (unit: Kg). This indicates the average time taken for testing a unit weight of sample (unit: hours / kg). This is the preset time compensation constant (unit: hours).

[0074] This formula implements a "quality incentive mechanism": because Located in the numerator, the higher the quality score of the item, the better. A positive increase results in a higher scheduling weight, allowing for priority access to laboratory testing. Furthermore, the denominator term... With time compensation constant The units of measurement are all standardized to "hours," ensuring the physical rigor of index value calculations. This is achieved by introducing... , It can effectively smooth out queuing jitter in the laboratory during peak geological survey periods. Experimental data shows that this mechanism can shorten the detection feedback cycle of high-quality samples by about 45%.

[0075] The dynamic flow operator executes the following decision logic: when At that time, the system automatically sends a test delivery handshake command to the laboratory through the gateway, realizing fully automated appointment. Located in the numerator, high-quality projects receive higher queuing priority. This "quality incentive" mechanism achieves deep coupling between business processes and laboratory resources, effectively solving the problem of blind resource allocation caused by inconsistent data quality; when When the system suspends the task and transfers it to manual review; when At that time, the system forcibly locked the terminal input interface and required the investigator to retest.

[0076] Spatial sampling information entropy score The following normalization formula is used for calculation:

[0077]

[0078] In the formula: This represents the probability distribution of sample points within the j-th spatial grid cell, where n represents the total number of geographic grids used to calculate information entropy; by dividing by... Factors ensure The values ​​range from 0 to 1 for different grid sizes.

[0079] The pre-trained GeoBERT model is used to map the “stratigraphic-structural-lithological” description text uploaded by the terminal into d-dimensional feature vectors. The GeoBERT model employs the BERTBase architecture, comprising a 12-layer Transformer encoder, 768 hidden layers, and 12 self-attention heads. Furthermore, the model was pre-trained using a masked language model (MLM) on a geoscience corpus containing at least 1 million geoscientific sentences, and retrieves standard feature vectors for corresponding layers from the built-in geological knowledge graph 311. The GeoBERT model, built upon the 12-layer Transformer encoder of BERTBase, enhances semantics through a 'geological entity alignment layer'. During training, negative sampling samples are constructed using entity triples from the geological knowledge graph 311, and fine-tuned using a cosine contrastive loss function, enabling the model to recognize geological time sequences. Finally, the descriptive text is mapped to... dimensional hidden layer feature vector ;calculate and cosine similarity as :

[0080]

[0081] Vobs represents the d-dimensional geological description observation feature vector output by the GeoBERT model, where d=768. This represents the feature vector of a standard stratigraphic triplet retrieved from the geological knowledge graph database 311.

[0082] Furthermore, the central workflow scheduling center 110 possesses dynamic backtracking capabilities. When subsequent testing results indicate systematic errors in the original data, a traceability index is established using the unique identifier (UUID) of the data stream. This supports reverse freezing of batches of data with systematic errors. The specific steps for reverse freezing are as follows: the system retrieves the distributed business database 130 using the unique identifier (UUID) of the data stream, changes the compliance status bit of the corresponding record from "normal" to "suspended," and simultaneously triggers the workflow engine to send an interception command to the collaborative laboratory server, suspending the testing task for that batch of samples. The central management and control platform 100 utilizes a blockchain ownership confirmation module to record the ownership of each node transfer. The result is constructed using a MerkleTree root hash on the blockchain, ensuring end-to-end data tamper-proofing. The blockchain ownership confirmation module adopts the Hyperledger Fabric architecture. (MerkleTree leaf nodes...) The construction is as follows: Each block contains 50 workflow trace hashes, and the root hash is synchronized to the distributed business database 130 to ensure that every score from the collection end to the approval end has legal evidence-level traceability.

[0083] The judgment thresholds of 0.85 and 0.65 set in this application are preset based on sensitivity analysis of historical rework data from geological surveys. Among them, 0.85 is the high confidence threshold, corresponding to 3 times the standard deviation. The compliant sample range under the principle; 0.65 is the mandatory interception threshold. A value below this means that the spatial deviation or semantic logic conflict of the sample points has exceeded the limits allowed by industry standards.

[0084] The central workflow scheduling center 110 uses UUID identifiers to establish a data flow mapping table. When a systematic deviation is detected in a batch of test data, such as instrument temperature drift, the system automatically retrieves the corresponding UUID record from the distributed business database 130. By modifying the "compliance status bit" of that batch of data in the database from "Active" to "Frozen," a stop signal is simultaneously sent to the collaborating laboratory server. This process is implemented based on the workflow engine's message listening mechanism, ensuring real-time interception before the physical testing in the laboratory begins.

[0085] In the mineral survey project, investigators entered sampling points using handheld terminals. .

[0086] Spatial positioning: Raw WGS84 coordinates (120.12345, 37.54321) were acquired using GNSS. The translation parameters for this area were then solved in real-time using CGCS2000 adjustment correction element 321. Convert it to CGCS2000.

[0087] Semantic validation: The input text is "Sinian Penglai Group, fine-grained quartzite". The GeoBERT model outputs feature vectors. The spectral library indicates that the standard stratigraphic level for this coordinate should be the "Neoarchean Jiaodong Group".

[0088] Result calculation: Calculated... .

[0089] Weighted score .

[0090] Process Response: Because The system automatically suspends the testing process for that sample point and pushes it to the project chief engineer's mobile office app, where it enters the "manual review" queue, thus avoiding invalid testing of erroneous samples.

[0091] The principle and process of this application are as follows:

[0092] 1. Data perception and adjustment: The portable geological survey terminal 210 collects raw field data, which is instantly converted into national standard coordinates by the CGCS2000 adjustment correction unit 321 of the intelligent verification calculation base 300, eliminating geometric drift.

[0093] 2. Multi-dimensional quality audit: The intelligent verification calculation base simultaneously verifies spatial accuracy. Geological semantics and spatial distribution uniformity Automated numerical calculations are performed across three dimensions.

[0094] 3. Dynamic workflow determination: The central workflow scheduling center 110 determines the dynamic workflow based on the calculated data. The overall score drives the state machine. High-quality data enters the "fully automatic reservation" track, while low-quality data is "forcibly intercepted and locked" at the field end.

[0095] 4. Resource Priority Allocation: High-quality results are approved. By leveraging laboratory resources, we can gain priority access to testing.

[0096] 5. Ownership and Traceability: All key decisions and data characteristics are uploaded to the chain through the MerkleTree root hash and associated with UUID, realizing the entire process from field collection to laboratory testing as "visible, traceable, and controllable".

[0097] The above are merely optional embodiments of this application and are not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A workflow engine-based end-to-end management system for geological survey projects, characterized by: include: The central management and control platform (100) has a built-in central workflow scheduling center (110), cross-agency collaboration gateway (120) and distributed business database (130); the field operation perception subsystem (200) includes a portable geological survey terminal (210), which integrates a dual-frequency multi-constellation GNSS receiver board. The intelligent verification calculation base (300) includes a geological semantic verification engine (310) and a multi-source spatial data fusion module (320). Resource and monitoring feedback subsystem (400), including laboratory resource scheduler (410); The connection and assembly relationship is as follows: the portable geological survey terminal (210) is bidirectionally connected to the cross-institutional collaborative gateway (120) through an encrypted 4G / 5G private network; The central management and control platform (100) communicates asynchronously with the intelligent verification computing base (300) via a message queue bus; The cross-institutional collaboration gateway (120) connects to the collaboration laboratory server based on the OAuth2.0 protocol; the central workflow scheduling center (110) integrates a dynamic flow operator, which is used to calculate the quality evaluation function fed back by the geological semantic verification engine (310). and according to The calculation results generate state machine transition instructions that drive the laboratory resource scheduler (410) in real time.

2. The workflow engine-based full-process management system for geological survey projects according to claim 1, characterized in that: The evaluation function ) is defined as: in: This is the overall quality assessment score for the current node. Scoring for spatial accuracy. For the geological semantic consistency score, To evaluate the spatial sampling information entropy score of the uniformity of sampling point distribution, The weighting coefficients are and satisfy the following conditions: .

3. The workflow engine-based full-process management system for geological survey projects according to claim 1, characterized in that: Spatial accuracy score , where RMSE is the root mean square error of the sampling points relative to the standard scale base map. This is the maximum deviation threshold required by the standard.

4. The workflow engine-based full-process management system for geological survey projects according to claim 1, characterized in that: The geological semantic verification engine (310) calculates through the following steps : The pre-trained GeoBERT model is used to map the "stratigraphic-structural-lithological" descriptive text uploaded by the terminal into d-dimensional feature vectors. Where d takes the value of 768; wherein, the GeoBERT model adopts the BERTBase architecture, including 12 Transformer encoder layers, 768 hidden layer dimensions and 12 self-attention heads, and the model is pre-trained on a geoscience professional corpus with a masked language model MLM of no less than 1 million geoscience sentences, and retrieves the standard feature vector of the corresponding layer from the built-in geological knowledge graph library (311). ;calculate and cosine similarity as : 。 5. The workflow engine-based full-process management system for geological survey projects according to claim 1, characterized in that: The spatial sampling information entropy score The following normalization formula is used for calculation: In the formula: The percentage of sample point density within the j-th grid cell, where n is the total number of grid cells; this is achieved by dividing by a factor. make sure The values ​​range from 0 to 1 for different grid sizes.

6. The workflow engine-based full-process management system for geological survey projects according to claim 1, characterized in that: The multi-source spatial data fusion module (320) is equipped with a CGCS2000 adjustment correction unit (321). This unit reads the reference layer data, uses the BursaWolf seven-parameter model to solve the translation, rotation and scale change parameters, and performs national standard geodetic coordinate system transformation on the original GPS coordinates.

7. The geological survey project end-to-end management system based on a workflow engine as described in claim 1, characterized in that: The laboratory resource scheduler (410) calculates the comprehensive scheduling index value. The formula is: in, This is the priority weight coefficient. The overall quality assessment score for the current node is given by L, where L is the total measured weight of the current submitted samples (in kg), and J is the current equipment load rate of the laboratory (range 0-1). The average testing cycle time per unit weight of sample (unit: hours / kg). This is the preset time compensation constant (unit: hours).

8. The workflow engine-based full-process management system for geological survey projects according to claim 1, characterized in that: The determination logic of the dynamic flow operator is as follows: when At that time, the system automatically sends a test delivery handshake command to the external laboratory; when When this happens, the system suspends the current task and pushes it to the supervisor's manual review queue; when At that time, the system immediately locks the terminal input interface and issues a rectification warning.

9. The workflow engine-based full-process management system for geological survey projects according to claim 1, characterized in that: The central workflow scheduling center (110) has a dynamic backtracking function. When the later detection results show that there is a systematic error in the original data, a traceability index is established through the unique identifier UUID of the data stream. It supports the reverse freezing of batch data with systematic errors. The specific steps of the reverse freezing are as follows: the system retrieves the distributed business database (130) through the unique identifier UUID of the data stream, changes the compliance status bit of the corresponding record from "normal" to "suspended", and synchronously triggers the workflow engine to send an interception command to the collaborative laboratory server to stop the testing task of the batch of samples.

10. The workflow engine-based full-process management system for geological survey projects according to claim 1, characterized in that: The central management and control platform (100) also integrates a blockchain-based rights confirmation module, which records the ownership of each workflow transfer. The results and corresponding responsible person information are calculated into the MerkleTree root hash and stored on the blockchain. The leaf nodes of the MerkleTree are composed of UUID and quality score. The concatenated data of operator identification and timestamp is generated by SHA256 hash operation.

Images