Territorial resource survey result checking system based on artificial intelligence

Abstract

The invention relates to the technical field of resource investigation, and discloses a territorial resource investigation achievement checking system based on artificial intelligence, comprising a multi-source data acquisition module used for acquiring a screening achievement from an optical remote sensing interpretation module, 3D scanning model data of an aerial photography module, and territorial change investigation data; and the data consistency comparison module is used for comparing the screening result of the optical remote sensing interpretation module and the 3D scanning model data of the aerial photographing module with the land change survey data so as to obtain the remote sensing interpretation difference degree and the three-dimensional model difference degree. By integrating multi-source data and performing consistency comparison, the accuracy and comprehensiveness of territorial resource survey result verification can be remarkably improved, and the difference degree is calculated by comparing optical remote sensing interpretation and aerial photography 3D scanning data with territorial change survey data, so that a change area is identified more accurately, personal errors and omission are reduced, and the accuracy and comprehensiveness of territorial resource survey result verification are improved. And the verification result is ensured to be more reliable.

Description

Technical Field

[0001] This invention relates to the field of resource survey technology, and in particular to an artificial intelligence-based land and resources survey results verification system. Background Technology

[0002] Currently, the verification of land and resources survey results mainly relies on manual comparison and analysis of single data sources, which is inefficient and highly subjective. Traditional methods struggle to integrate multi-source heterogeneous data (such as optical remote sensing, 3D scanning, and change survey data), resulting in narrow verification coverage and limited accuracy. Furthermore, existing systems lack adaptability to different network environments, failing to maintain the continuity and consistency of verification tasks under conditions of network isolation or offline operation, thus hindering the real-time and large-scale application of field surveys. In addition, the verification process lacks a reliable automated verification mechanism, making it difficult to guarantee the robustness and credibility of results under different conditions.

[0003] An existing patent discloses a method and system for verifying land and resources survey results based on artificial intelligence (publication number CN120196663A). The technology disclosed in this patent mainly relies on batch comparison of historical and current land datasets and distributed server processing, lacking the fusion analysis of multi-source spatial data such as optical remote sensing and 3D scanning, resulting in insufficient ability to identify actual changes in the land surface. Furthermore, it fails to adapt to multiple network environments such as operational intranets and government extranets, limiting the flexibility of field verification; and the verification process lacks a cross-model cross-validation mechanism, making it difficult to guarantee the reliability and robustness of the results under different conditions. Summary of the Invention

[0004] This invention provides an artificial intelligence-based land and resources survey results verification system to address existing technical problems, solving issues such as limited verification dimensions, poor environmental adaptability, and weak verification mechanisms.

[0005] To address the aforementioned technical problems, according to one aspect of the present invention, more specifically, an artificial intelligence-based land and resources survey results verification system, comprising:

[0006] The multi-source data acquisition module is used to acquire screening results from the optical remote sensing interpretation module, 3D scanning model data from the aerial photography module, and land change survey data.

[0007] The data consistency comparison module is used to compare the screening results of the optical remote sensing interpretation module and the 3D scanning model data of the aerial photography module with the land change survey data, so as to obtain the remote sensing interpretation difference degree and the 3D model difference degree respectively.

[0008] The multi-network environment adaptation module is used to adapt the remote sensing interpretation difference degree and the 3D model difference degree to the first network environment, the second network environment, or the Nth network environment, respectively, and output the remote sensing interpretation difference degree parameters and the 3D model difference degree parameters that fit the first network environment and the second network environment, respectively.

[0009] The verification model construction module generates feedback verification models for the first network environment, the second network environment, or the Nth network environment, respectively, based on the parameters of the first network environment, the second network environment, or the Nth network environment.

[0010] The cross-validation module is used to input the parameters adapted by the multi-network environment adaptation module into the feedback verification model generated by different network environments. Therefore:

[0011] 1) When the coefficients output by the feedback verification model meet expectations, it means that the current cross path has passed verification;

[0012] 2) When the coefficients output by the feedback verification model do not meet expectations, it means that the current cross path has not passed the verification.

[0013] 3) Calculate the total number of cross paths and the cross paths that have passed verification, and determine whether the land and resources survey results have passed verification based on whether the proportion of the cross paths that have passed verification in the total number of cross paths exceeds the preset value.

[0014] Furthermore, the multi-source data acquisition module specifically utilizes high-resolution optical remote sensing satellites for image interpretation, screens deformed areas and fixes suspected points; and uses UAV-borne lidar to perform 3D scanning of suspected points to obtain hazard parameters and terrain models.

[0015] Furthermore, the multi-network environment adaptation module supports at least two of the following environments: intranet for work, extranet for government affairs, Internet, and offline Internet environment; the multi-network environment adaptation module can allocate verification tasks in different network environments and maintain the consistency of task status and data when switching network environments.

[0016] Furthermore, the remote sensing interpretation difference is used to assess the consistency between the screening results of optical remote sensing and the land survey data, and the calculation formula is as follows:

[0017] ;

[0018] In the above formula, c refers to the remote sensing interpretation difference degree, which is defined as the degree of difference when the screening results are completely consistent with the land survey data. When the screening results are completely inconsistent with the land survey data C represents the set of suspected change areas identified by the optical remote sensing interpretation module (such as deformation points or areas fixed through image interpretation); B represents the set of change areas recorded in the land change survey data. This indicates the size of the overlap between the suspected area detected by optical remote sensing and the area changed in the land use change survey. This indicates the combined size of the suspected area from optical remote sensing and the area changed in the land use change survey.

[0019] Furthermore, the 3D model difference degree is used to compare the consistency of topographic changes between the 3D model data and the land survey data, and the calculation formula is as follows:

[0020] ;

[0021] In the above formula, d refers to the degree of difference between the three-dimensional model and the land survey data. When the 3D model data is completely inconsistent with the land survey data D represents the set of terrain change areas detected in the 3D scanning model data of the aerial photography module (obtained through lidar point cloud analysis, such as areas where the elevation change is greater than a certain threshold); B represents the set of change areas recorded in the land change survey data. Size of the overlap between the 3D scanned area of ​​change and the area of ​​change from the land use change survey; This indicates the combined size of the 3D scan change area and the land change survey change area.

[0022] Furthermore, the specific steps of the verification model construction module in generating feedback verification models for different network environments are as follows:

[0023] 1) Obtain remote sensing interpretation difference parameters and 3D model difference parameters from land and resources samples in the first network environment, the second network environment, or the Nth network environment;

[0024] 2) A third-party inspection team will conduct an on-site inspection of the land and resources samples and provide a pass / fail coefficient for the land and resources survey results for each sample;

[0025] 3) In the first network environment, the second network environment, or the Nth network environment, the correlation between the remote sensing interpretation difference parameter and the pass coefficient, and the correlation between the 3D model difference parameter and the pass coefficient are determined by the control variable method;

[0026] 4) Again, the correlation between the remote sensing interpretation difference parameter and the 3D model difference parameter is determined by the control variable method, and then the feedback verification model in the first network environment, the second network environment or the Nth network environment is fitted and generated.

[0027] Furthermore, the feedback verification model is a mathematical function model generated by fitting remote sensing interpretation difference parameters and 3D model difference parameters under different network environments through the control variable method. This mathematical function model outputs a coefficient that can be used to represent the qualification level of the land and resources survey results.

[0028] Furthermore, in the cross-validation module, the cross-path refers to the process of inputting the parameters output by the multi-network environment adaptation module into the feedback verification model in a single network environment for verification; the total cross-path refers to the total number of verification paths in all network environments, that is, the product of the number of network environments and the set of parameters to be verified.

[0029] Furthermore, the specific steps in the cross-validation module to determine whether the coefficients output by the feedback verification model meet expectations are as follows:

[0030] 1) A preset acceptable threshold range is determined based on historical data of the acceptable coefficients provided by a third-party monitoring group for the sample data;

[0031] 2) Compare the coefficients output by the feedback verification model with the acceptable threshold range;

[0032] 3) If the output coefficient is within the acceptable threshold range, it is determined to meet expectations; if the output coefficient is outside the acceptable threshold range, it is determined to not meet expectations.

[0033] This invention provides an artificial intelligence-based land and resources survey results verification system. Compared with existing technologies, the advantages of this method are as follows:

[0034] 1. This invention significantly improves the accuracy and comprehensiveness of land and resources survey results verification by integrating multi-source data and performing consistency comparison. Furthermore, it calculates the degree of difference by comparing optical remote sensing interpretation and aerial 3D scanning data with land change survey data, thereby more accurately identifying changed areas, reducing human error and omissions, and ensuring that the verification results are more reliable.

[0035] 2. This invention possesses strong network environment adaptability, enabling stable operation in various network environments such as intranets for work, extranets for government affairs, and the Internet. The multi-network environment adaptation module ensures task allocation and data consistency under different network conditions, allowing verification work to proceed without network limitations, improving the system's practicality and coverage, and making it suitable for various field operation scenarios.

[0036] 3. This invention employs artificial intelligence technology to construct a feedback verification model, achieving automation and intelligence in the verification process. By fitting a mathematical model using the controlled variable method, the system can automatically output the pass / fail coefficient, significantly reducing the workload of manual verification, improving processing speed, and minimizing biases caused by subjective factors.

[0037] 4. This invention introduces a cross-validation mechanism, which enhances the credibility and robustness of the verification results through multi-path verification. Parameters are input into feedback verification models in different network environments, and the pass rate of verification paths is statistically analyzed to ensure the reliability of verification results under different conditions. This reduces the risk of deviation in a single environment and improves the overall verification quality.

[0038] 5. This invention can dynamically adjust model parameters based on historical data and real-time samples, maintaining the timeliness and adaptability of verification. Through sample data from third-party monitoring groups, the system continuously optimizes and feeds back the verification model, ensuring that it maintains high accuracy over time and with environmental changes, thereby continuously meeting the needs of land and resources management. Attached Figure Description

[0039] Figure 1 This is a flowchart of the present invention;

[0040] Figure 2 This is a graph showing the relationship between the remote sensing interpretation difference parameter and the pass / fail coefficient of this invention.

[0041] Figure 3 This is a graph showing the relationship between the difference parameters and the pass / fail coefficient of the three-dimensional model of this invention.

[0042] Figure 4 This is a diagram showing the relationship between the remote sensing interpretation difference parameters and the 3D model difference parameters of this invention. Detailed Implementation

[0043] To make the technical solution of the present invention clearer, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0044] like Figure 1-4 As shown, according to one aspect of the present invention, an artificial intelligence-based land and resources survey results verification system is provided, comprising:

[0045] Step S1: Obtain screening results from the optical remote sensing interpretation module, 3D scanning model data from the aerial photography module, and land change survey data; The multi-source data acquisition module is specifically used for: interpreting images using high-resolution optical remote sensing satellites, screening deformed areas and fixing suspected points; and performing 3D scanning of suspected points using UAV-borne lidar to obtain hazard parameters and terrain models.

[0046] Step S2: Compare the screening results of the optical remote sensing interpretation module and the 3D scanning model data of the aerial photography module with the land change survey data to obtain the remote sensing interpretation difference and the 3D model difference respectively.

[0047] Step S3: Perform multi-environment adaptation on the remote sensing interpretation difference and the 3D model difference in the first network environment, the second network environment, or the Nth network environment, and output the remote sensing interpretation difference parameters and the 3D model difference parameters that fit the first network environment and the second network environment, respectively.

[0048] The multi-network environment adaptation module supports at least two of the following environments: intranet for work operations, extranet for government affairs, the internet, and offline internet. This module can allocate verification tasks across different network environments and maintain consistency between task status and data when switching network environments. Therefore:

[0049] Remote sensing interpretation discrepancy is used to assess the consistency between the screening results of optical remote sensing and land survey data. The calculation formula is as follows:

[0050] ;

[0051] In the above formula, c refers to the remote sensing interpretation difference degree, which is defined as the degree of difference when the screening results are completely consistent with the land survey data. When the screening results are completely inconsistent with the land survey data C represents the set of suspected change areas identified by the optical remote sensing interpretation module (such as deformation points or areas fixed through image interpretation); B represents the set of change areas recorded in the land change survey data. This indicates the size of the overlap between the suspected area detected by optical remote sensing and the area changed in the land use change survey. This indicates the combined size of the suspected area from optical remote sensing and the area changed in the land use change survey.

[0052] The parameter for remote sensing interpretation dissimilarity in the first network environment, the second network environment, or the Nth network environment is:

[0053] ;

[0054] In the above formula, , , Represents the weighting coefficient, where ; This indicates the feature matching degree between the set of suspected change areas screened by the optical remote sensing interpretation module and the set of change areas recorded in the land change survey data (based on SIFT feature point matching). This indicates the texture similarity (based on the gray-level co-occurrence matrix) between the set of suspected change areas screened by the optical remote sensing interpretation module and the set of change areas recorded in the land change survey data.

[0055] When the first network environment, the second network environment, or the Nth network environment are within the operational intranet, the weighting coefficient is generally taken as... , , When using the government's external network, the weighting coefficient is generally taken as... , , In internet or offline environments, the weighting coefficient is generally taken as... , , .

[0056] In this embodiment, the 3D model difference degree is used to compare the consistency of topographic changes between the 3D model data and the land survey data. The calculation formula is as follows:

[0057] ;

[0058] In the above formula, d refers to the degree of difference between the three-dimensional model and the land survey data. When the 3D model data is completely inconsistent with the land survey data D represents the set of terrain change areas detected in the 3D scanning model data of the aerial photography module (obtained through lidar point cloud analysis, such as areas where the elevation change is greater than a certain threshold); B represents the set of change areas recorded in the land change survey data. Size of the overlap between the 3D scanned area of ​​change and the area of ​​change from the land use change survey; This indicates the combined size of the 3D scan change area and the land change survey change area.

[0059] The parameter for the 3D model difference degree in the first network environment, the second network environment, or the Nth network environment is:

[0060] ;

[0061] In the above formula, , , Represents the weighting coefficient, where ; This indicates the consistency of volume changes between the set of terrain change areas detected in the 3D scanning model data of the aerial photography module and the set of change areas recorded in the land change survey data. This indicates the geometric similarity between the set of terrain change areas detected in the 3D scanning model data of the aerial photography module and the set of change areas recorded in the land change survey data (based on point cloud registration accuracy).

[0062] When the first network environment, the second network environment, or the Nth network environment are within the operational intranet, the weighting coefficient is generally taken as... , , When using the government's external network, the weighting coefficient is generally taken as... , , In internet or offline environments, the weighting coefficient is generally taken as... , , .

[0063] Step S4: Based on the parameters of the first network environment, the second network environment, or the Nth network environment, generate feedback verification models for the first network environment, the second network environment, or the Nth network environment, respectively.

[0064] For example, the specific steps to construct a feedback verification model for known data in the first network environment are as follows:

[0065] 1) Obtain the remote sensing interpretation difference parameters and 3D model difference parameters from the land and resources samples in the first network environment;

[0066] 2) A third-party inspection team will conduct an on-site inspection of the land and resources samples and provide a pass / fail coefficient for the land and resources survey results for each sample;

[0067] The third-party monitoring team, comprised of senior internal staff, evaluates known land and resources samples. For example, they conduct on-site inspections of 1000 known samples and compare the results with the original land and resources survey findings. The samples are then ranked according to the magnitude of the deviation (samples with the same deviation are excluded). If the deviation of a result is less than that of the other 500 samples, the sample's pass rate is considered to be 50.0%.

[0068] 3) In the first network environment, the second network environment, or the Nth network environment, the correlation between the remote sensing interpretation difference parameter and the pass coefficient, and the correlation between the 3D model difference parameter and the pass coefficient are determined by the control variable method;

[0069] 1. Among these 1000 samples, 100 samples with a 3D model difference parameter of 0.4 are selected. The relationship between the remote sensing interpretation difference parameter and the pass coefficient is determined through these 100 samples. The pass coefficient at this time is denoted as pass coefficient 1.

[0070] Therefore, based on the relationship between the pass coefficient 1 and the remote sensing interpretation difference parameter, we can establish a mathematical expression:

[0071] (Formula 1);

[0072] In Formula 1 above, , Used for control A constant that approximates the pass / fail coefficient of 1. And by... Figure 2 The data in the middle can be determined , When, in Formula 1 It approaches the pass coefficient 1 (where, express Figure 2 The curve in the figure shows the pass / fail coefficient (1) indicated by the red dot.

[0073] 2. Among these 1000 samples, 100 samples with a remote sensing interpretation difference parameter of 0.4 are selected. The relationship between the difference parameter of the 3D model and the pass coefficient is determined through these 100 samples. The pass coefficient at this time is denoted as pass coefficient 2.

[0074] Therefore, based on the relationship between the pass coefficient 2 and the 3D model difference parameter, we can establish a mathematical expression:

[0075] (Formula 2);

[0076] In formula 2 above, , Used for control A constant that approximates the pass / fail coefficient 2. And by... Figure 3 The data in the middle can be determined , When, in Formula 2 It approaches the pass coefficient 2 (where, express Figure 3 The curve in the figure shows the pass / fail coefficient (2) represented by the blue dot.

[0077] 4) Again, the correlation between the remote sensing interpretation difference parameter and the 3D model difference parameter is determined by the control variable method, and then the feedback verification model in the first network environment, the second network environment or the Nth network environment is fitted and generated.

[0078] From these 1000 samples, 100 consecutive samples are randomly selected, and the remote sensing interpretation difference parameters, 3D model difference parameters and pass coefficients are obtained for these 100 samples. The pass coefficient at this time is denoted as pass coefficient 3.

[0079] Therefore, based on the relationship between the remote sensing interpretation difference parameter and the pass / fail coefficient, and the relationship between the 3D model difference parameter and the pass / fail coefficient, the expressions for the remote sensing interpretation difference parameter, the 3D model difference parameter, and the pass / fail coefficient are as follows:

[0080] H= × (Formula 1) × (Formula 2);

[0081] In the above, A constant used to control the approximation of formula H and the pass coefficient 3. And by... Figure 4 The data in the middle can be determined At this point, formula H approaches the approximate pass coefficient 3. The mathematical expression for the feedback verification model generated by the fit at this time is:

[0082] ;

[0083] In the above formula, H represents the coefficient output by the feedback verification model under the known first network environment. When the coefficient output by the feedback verification model meets expectations, it indicates that the current path under the first network environment has passed verification.

[0084] 5) Based on the feedback verification model in the known first network environment, it is possible to verify unknown land and resources survey results.

[0085] Step S5: Input the parameters adapted by the multi-network environment adaptation module into the feedback verification model generated for different network environments, then:

[0086] 1. When the coefficients output by the feedback verification model meet expectations, it indicates that the current cross-path has passed verification;

[0087] 2. If the coefficients output by the feedback verification model do not meet expectations, it means that the current cross path has not passed the verification.

[0088] Third, calculate the total number of cross paths and the cross paths that have passed verification, and determine whether the land and resources survey results have passed verification based on whether the proportion of verified cross paths in the total number of cross paths exceeds a preset value.

[0089] In the cross-validation module, the cross-path refers to the process of inputting the parameters output by the multi-network environment adaptation module into the feedback verification model in a single network environment for verification. The total cross-path refers to the total number of verification paths across all network environments, i.e., the product of the number of network environments and the set of parameters to be verified. Furthermore, the specific steps in the cross-validation module to determine whether the coefficients output by the feedback verification model meet expectations are as follows:

[0090] 1) A preset acceptable threshold range is determined based on historical data of the acceptable coefficients provided by a third-party monitoring group for the sample data;

[0091] 2) Compare the coefficients output by the feedback verification model with the acceptable threshold range;

[0092] 3) If the output coefficient is within the acceptable threshold range, it is determined to meet expectations; if the output coefficient is outside the acceptable threshold range, it is determined to not meet expectations.

[0093] 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 land and resources survey results verification system based on artificial intelligence, characterized in that, include: The multi-source data acquisition module is used to acquire screening results from the optical remote sensing interpretation module, 3D scanning model data from the aerial photography module, and land change survey data. The data consistency comparison module is used to compare the screening results of the optical remote sensing interpretation module and the 3D scanning model data of the aerial photography module with the land change survey data, so as to obtain the remote sensing interpretation difference degree and the 3D model difference degree respectively. The multi-network environment adaptation module is used to adapt the remote sensing interpretation difference degree and the 3D model difference degree to the first network environment, the second network environment, or the Nth network environment, respectively, and output the remote sensing interpretation difference degree parameters and the 3D model difference degree parameters that fit the first network environment and the second network environment, respectively. The verification model construction module generates feedback verification models for the first network environment, the second network environment, or the Nth network environment, respectively, based on the parameters of the first network environment, the second network environment, or the Nth network environment. The cross-validation module is used to input the parameters adapted by the multi-network environment adaptation module into the feedback verification model generated by different network environments. Therefore: 1) When the coefficients output by the feedback verification model meet expectations, it means that the current cross path has passed verification; 2) When the coefficients output by the feedback verification model do not meet expectations, it means that the current cross path has not passed the verification. 3) Calculate the total number of cross paths and the cross paths that have passed verification, and determine whether the land and resources survey results have passed verification based on whether the proportion of the cross paths that have passed verification in the total number of cross paths exceeds the preset value.

2. The artificial intelligence-based land and resources survey results verification system according to claim 1, characterized in that: The multi-source data acquisition module specifically utilizes high-resolution optical remote sensing satellites to interpret images, screen deformed areas, and fix suspected points; it also uses UAV-borne lidar to perform 3D scanning of suspected points to obtain hazard parameters and terrain models.

3. The artificial intelligence-based land and resources survey results verification system according to claim 1, characterized in that: The multi-network environment adaptation module supports at least two of the following environments: intranet for operations, extranet for government affairs, Internet, and offline Internet. The multi-network environment adaptation module can allocate verification tasks in different network environments and maintain the consistency of task status and data when switching network environments.

4. The land and resources survey results verification system based on artificial intelligence according to claim 1, characterized in that: The remote sensing interpretation discrepancy is used to assess the consistency between the screening results of optical remote sensing and the land survey data. The calculation formula is as follows: ； In the above formula, c refers to the remote sensing interpretation difference degree, which is defined as the degree of difference when the screening results are completely consistent with the land survey data. When the screening results are completely inconsistent with the land survey data C represents the set of suspected change areas screened by the optical remote sensing interpretation module; B represents the set of change areas recorded in the land change survey data.

5. The artificial intelligence-based land and resources survey results verification system according to claim 1, characterized in that: The 3D model difference degree is used to compare the consistency of topographic changes between the 3D model data and the land survey data. The calculation formula is as follows: ； In the above formula, d refers to the degree of difference between the three-dimensional model and the land survey data. When the 3D model data is completely inconsistent with the land survey data D represents the set of terrain change areas detected in the 3D scanning model data of the aerial photography module; B represents the set of change areas recorded in the land change survey data.

6. The artificial intelligence-based land and resources survey results verification system according to claim 1, characterized in that: The specific steps for the verification model construction module to generate feedback verification models for different network environments are as follows: 1) Obtain remote sensing interpretation difference parameters and 3D model difference parameters from known land and resources samples in the first network environment, the second network environment, or the Nth network environment; 2) A third-party inspection team will conduct a second survey of the known land and resources samples and provide a pass / fail coefficient for the land and resources survey results for each sample; 3) In the first network environment, the second network environment, or the Nth network environment, the correlation between the remote sensing interpretation difference parameter and the pass coefficient, and the correlation between the 3D model difference parameter and the pass coefficient are determined by the control variable method; 4) Again, the correlation between the remote sensing interpretation difference parameter and the 3D model difference parameter is determined by the control variable method, and then the feedback verification model in the first network environment, the second network environment or the Nth network environment is fitted and generated.

7. The artificial intelligence-based land and resources survey results verification system according to claim 1, characterized in that: The feedback verification model is a mathematical function model generated by fitting remote sensing interpretation difference parameters and 3D model difference parameters under different network environments through the control variable method. The mathematical function model outputs a coefficient that can be used to represent the qualification level of the land and resources survey results.

8. The artificial intelligence-based land and resources survey results verification system according to claim 1, characterized in that: In the cross-validation module, the cross-path refers to the process of inputting the parameters output by the multi-network environment adaptation module into the feedback verification model in a single network environment for verification; the total cross-path refers to the total number of verification paths in all network environments, that is, the product of the number of network environments and the set of parameters to be verified.

9. The artificial intelligence-based land and resources survey results verification system according to claim 1, characterized in that: The specific steps in the cross-validation module to determine whether the coefficients output by the feedback verification model meet expectations are as follows: 1) A preset acceptable threshold range is determined based on historical data of the acceptable coefficients provided by a third-party monitoring group for the sample data; 2) Compare the coefficients output by the feedback verification model with the acceptable threshold range; 3) If the output coefficient is within the acceptable threshold range, it is determined to meet expectations; if the output coefficient is outside the acceptable threshold range, it is determined to not meet expectations.

Images