Error elimination method and system for bridge deformation monitoring result based on PS-InSAR technology

Abstract

The embodiment of the application discloses a kind of based on PS-InSAR technology bridge deformation monitoring result error elimination method and system, comprising: to SAR image set is selected, then based on PS-InSAR technology obtains the deformation monitoring result of the target bridge in area, finally based on optical satellite image and SAR satellite image confirms belonging to bridge on PS deformation point;Influencing factor data acquisition and the original influencing factor related data obtained are preprocessed to meet model processing requirement;Training model design is carried out based on data characteristics;Further processing is carried out to the data of pre-processing completion, and model is trained multiple times;The fitting error of each group is removed from original deformation monitoring result, and the cumulative deformation monitoring result of actual occurrence of target PS point is obtained.The application is researched by itself and bridge deformation change characteristics to PS-InSAR technology monitoring result, to realize the elimination of error in deformation monitoring, relative to the great advantage and convenience of error elimination in traditional method.

Description

Technical Field

[0001] This invention relates to the field of bridge deformation monitoring technology, specifically to a method and system for error elimination of bridge deformation monitoring results based on PS-InSAR technology. Background Technology

[0002] As a component of transportation infrastructure, the health of bridges directly impacts the smooth flow of traffic on related routes, and bridge collapses can easily lead to traffic accidents. Therefore, monitoring their operational status is essential. Currently, monitoring is mainly conducted through manual fixed-point observation or sensor-based methods. The former is labor-intensive, while the latter is expensive and lacks comprehensive coverage. There is an urgent need for a comprehensive and cost-effective monitoring method. Bridge deformation monitoring technology based on PS-InSAR technology perfectly meets this need, and this technology has received widespread attention in recent years as a novel bridge deformation monitoring technique.

[0003] PS-InSAR technology offers advantages such as all-weather, all-day operation, low cost, and high accuracy in bridge monitoring. The invention of corner reflectors provides a new strategy for monitoring bridge deformation with weak scattering, offering new assurance for its application in this field.

[0004] The deformation of a bridge over a period of time consists of its own permanent deformation and non-permanent deformation caused by variable forces. Non-permanent deformation is mainly caused by factors such as vehicle loads, temperature effects, and wind loads. Non-permanent deformation can affect the accurate assessment of the bridge's health condition and therefore needs to be removed.

[0005] Currently, the removal of non-permanent deformations in bridge deformation monitoring mainly relies on manual monitoring and sensors, which greatly limits the widespread adoption of bridge deformation monitoring technology. Therefore, proposing a convenient and efficient improvement method is of great significance for the promotion of bridge deformation monitoring. Summary of the Invention

[0006] To address this, this invention provides a method for error removal in bridge deformation monitoring results based on PS-InSAR technology, thereby solving the problems of high detection costs, low efficiency, and difficulty in widespread application caused by the manual removal of non-permanent deformations in existing bridge deformation monitoring technologies.

[0007] To achieve the above objectives, the embodiments of the present invention provide the following technical solutions:

[0008] In an embodiment of the present invention, an error elimination method for bridge deformation monitoring results based on PS-InSAR technology is provided, the method comprising the following steps:

[0009] Step S1: Filter the SAR image set, then obtain the deformation monitoring results of the area where the target bridge is located based on PS-InSAR technology, and finally confirm the PS deformation points on the bridge based on optical satellite imagery and SAR satellite imagery.

[0010] Step S2: Combine the deformation monitoring characteristics of PS-InSAR technology with the deformation characteristics of the bridge to determine the applicability of the feature points of the target bridge model;

[0011] Step S3: Obtain influencing factor data and preprocess the obtained raw influencing factor data to meet the model processing requirements;

[0012] Step S4: Design a training model based on data features;

[0013] Step S5: Further process the preprocessed data, train the model multiple times, and finally merge the multiple sets of fitted parameters to generate the final parameters;

[0014] Step S6: Remove the fitting error of each group from the original deformation monitoring results to obtain the actual cumulative deformation monitoring results of the target PS point.

[0015] Further, in step S1, the image selection method includes:

[0016] Acquire a cropped set of SAR images covering the target bridge;

[0017] Acquire image intensity data;

[0018] Calculate the average intensity value for each image;

[0019] The threshold is set based on the intensity value distribution statistics. In the data distribution statistics, we mainly look at the fluctuation range of the average intensity value of most images and set the lower limit of this range as the threshold.

[0020] Delete current SAR images that are significantly below the set threshold.

[0021] Further, in step S2, the method for determining applicability includes:

[0022] Identify the influencing factors on the bridge structure and determine the correlation between the deformation caused by the influencing factors at the feature points and the deformation monitored based on PS-InSAR technology.

[0023] Perform correlation analysis between each feature point and the influencing factor data;

[0024] The formula for calculating the correlation analysis is as follows:

[0025]

[0026] After calculating the correlation between each feature point and image factors, if the proportion of feature points with a correlation coefficient greater than 0.3 is greater than 80%, then it meets the requirements for subsequent model construction.

[0027] Furthermore, the influencing factors mainly include vehicle load, temperature effect, and wind load.

[0028] Further, step S3 specifically includes:

[0029] By analyzing assumptions and errors, influencing factors are identified, and data on these factors are collected.

[0030] One-hot encoding is performed on the influencing factor data to form digital tags;

[0031] The digitally labeled data of influencing factors are spliced ​​together to form the input data for model training.

[0032] Further, step S4 specifically includes:

[0033] The design of the training model can be calculated using a multiple linear regression equation;

[0034] The multiple linear regression equation is:

[0035]

[0036] After calculating and confirming the loss compensation, the relevant loss function is:

[0037]

[0038] Where y is the actual long-term deformation value, and x is the input impact factor data.

[0039] Further, step S4 specifically includes:

[0040] The design of the training model can also be calculated using the ridge regression equation;

[0041] The ridge regression equation is:

[0042]

[0043] After calculating and confirming the loss compensation, the relevant loss function is:

[0044]

[0045] Where y is the actual long-term deformation value, and x is the input impact factor data.

[0046] Further, step S5 specifically includes:

[0047] Acquire multiple images covering the target bridge, ensuring a quantity of more than 30 images;

[0048] The temporal deformation and influencing factors of each feature point data are coherently calculated and arranged in descending order of absolute value. The top 50% are selected as the fitted feature points. Feature points with large deformation and noise points are effectively removed to obtain the feature points that fit perfectly.

[0049] Each selected feature point is fitted individually, and the relevant data for each feature point are grouped sequentially, such as image number 1 to number 24 as group 1, image number 1 to number 25 as group 2, image number 1 to number 26 as group 3, and so on, forming n groups of training data. The earlier the data is acquired, the greater its influence on the result, i.e., it is assigned a higher weight.

[0050] The difference between each data set and the previous item is used to form the corresponding difference data, which is the input item for each training session.

[0051] Based on the PS-InSAR deformation monitoring results, the deformation difference value generated by each feature point in the two SAR images is obtained, and the deformation fitting result data corresponding to each set of input data is obtained;

[0052] Based on multiple sets of data for each feature point, the fitting training is performed on multiple sets of data for each feature point in sequence;

[0053] The average of the coefficients obtained from fitting each set of data is used as the final correction coefficient for the corresponding feature point.

[0054] Furthermore, step 6 specifically includes:

[0055] The deformation feature points on the bridge are divided into two categories: one category belongs to those that have undergone fitting calculation, denoted as category A points, and the other category belongs to those that have not undergone fitting calculation, denoted as category B points.

[0056] For points of type A, the corresponding fitting equation is applied to calculate the error caused by the variable action in each time period, and then this part of the error is removed in each time period;

[0057] For B-type points, search for the three closest A-type points to the target point, determine their respective weights using the inverse distance weighting method, calculate the coefficients of the fitting equation corresponding to the B-type points, and then remove the error using a method similar to that used for A-type points.

[0058] In another embodiment of the present invention, an error elimination system for bridge deformation monitoring results based on PS-InSAR technology is provided, the system comprising:

[0059] Deformation information acquisition module: Based on PS-InSAR technology, the deformation information of the monitored object is extracted from radar images. This deformation information includes the PS points of the bridge.

[0060] Model training module: Design the corresponding training model according to the established model training formula, and fit the parameters based on the factor data information;

[0061] Processing module: Processes the acquired deformation information data and digitizes the information of the influencing factors, compares the fitting results with the original detection information and removes the incorrect ones.

[0062] According to embodiments of the present invention, the method has the following advantages:

[0063] 1. By adaptively calculating and classifying the monitoring results of PS-InSAR technology with the characteristics of bridge deformation changes, it can eliminate errors in deformation monitoring. Compared with traditional methods that require a large amount of instrument monitoring data for error elimination, it has great advantages and convenience. Moreover, it reduces costs and improves monitoring efficiency and accuracy compared with existing technologies.

[0064] 2. Based on the actual scenario of bridge deformation monitoring, this method combines the advantages of PS-InSAR technology with the accurate extraction of permanent bridge deformation. At the same time, a targeted strategy is adopted for the removal of non-permanent deformation, so that the obtained permanent deformation results of the bridge can meet the needs of bridge health assessment to the greatest extent. This method is of great significance for the promotion of bridge deformation monitoring. Attached Figure Description

[0065] To more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings in the following description are merely exemplary, and those skilled in the art can derive other embodiments based on the provided drawings without creative effort.

[0066] The structures, proportions, sizes, etc. illustrated in this specification are only for the purpose of assisting those skilled in the art in understanding and reading the content disclosed herein, and are not intended to limit the conditions under which the present invention can be implemented. Therefore, they have no substantial technical significance. Any modifications to the structure, changes in the proportions, or adjustments to the size, without affecting the effects and objectives that the present invention can produce, should still fall within the scope of the technical content disclosed in the present invention.

[0067] Figure 1 A flowchart of an error elimination method for bridge deformation monitoring results based on PS-InSAR technology provided in an embodiment of the present invention;

[0068] Figure 2 The image quality removal process disclosed in the method provided in the embodiments of the present invention is illustrated in the diagram.

[0069] Figure 3 The bridge PS point determination method disclosed in the embodiments of the present invention requires simultaneous reference to high-resolution optical images and SAR satellite images;

[0070] Figure 4 A schematic diagram of the distribution of PS points belonging to bridges, selected based on the high-resolution optical images and SAR satellite images used in the embodiments of the present invention;

[0071] Figure 5 A schematic diagram of a preprocessed training data example provided in an embodiment of the present invention;

[0072] Figure 6 A schematic diagram of the time series result of the correction value obtained by fitting a target PS point in the method disclosed in the embodiment of the present invention and the original time series deformation result;

[0073] Figure 7 A schematic diagram of the temporal deformation result obtained after the feature points disclosed in the method provided in the embodiment of the present invention are subjected to non-permanent deformation removal by the method;

[0074] Figure 8 This is a structural diagram of an error elimination system for bridge deformation monitoring results based on PS-InSAR technology, provided in an embodiment of the present invention. Detailed Implementation

[0075] The following specific embodiments illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0076] The terms such as "upper," "lower," "left," "right," and "middle" used in this specification are merely for clarity of description and are not intended to limit the scope of the invention. Any changes or adjustments to their relative relationships, without substantially altering the technical content, should also be considered within the scope of the invention.

[0077] like Figure 1 As shown, this invention provides a method for error removal in bridge deformation monitoring results based on PS-InSAR technology. The method includes the following steps:

[0078] Step S1: Filter the SAR image set, then obtain the deformation monitoring results of the area where the target bridge is located based on PS-InSAR technology, and finally confirm the PS deformation points on the bridge based on optical satellite image and SAR satellite image.

[0079] Image selection methods specifically include:

[0080] Determine the monitoring period and select an appropriate SAR image set;

[0081] Determine the rectangular clipping vector covering the bridge, i.e., the length of the clipping vector can be more than 1km longer than the bridge body, and the width can be set to 1km-2km;

[0082] Crop the main image and all secondary images;

[0083] Acquire a cropped set of SAR images covering the target bridge;

[0084] Acquire image intensity data;

[0085] Calculate the average intensity value for each image;

[0086] The threshold is set based on the intensity value distribution statistics. In the data distribution statistics, we mainly look at the fluctuation range of the average intensity value of most images and set the lower limit of this range as the threshold.

[0087] Delete current SAR images whose intensity is significantly lower than the set threshold.

[0088] Based on the above steps, the selection and confirmation of the main image in the image set are completed during the application of PS-InSAR technology.

[0089] The main image and the auxiliary image (the part of the image set excluding the main image) are registered, interferometrically processed, PS point selection is performed, phase modeling and parameter estimation are performed, and deformation extraction is performed.

[0090] The PS point was confirmed to be located on the bridge by combining high-resolution optical imagery and SAR imagery.

[0091] like Figure 2 As shown, cropped SAR images of the target area are obtained based on the set rules, the average intensity of each SAR image is calculated, and a threshold is set to delete SAR images with poor quality to improve the monitoring results. In this example, the threshold is set to 0.33, and images below this threshold are deleted.

[0092] like Figure 3As shown, it displays the high-resolution optical imagery (reflecting ground feature type information) and SAR satellite imagery (reflecting ground feature intensity information) on which it is based. The calculation and analysis of bridge deformation monitoring results need to be based on the effective deformation monitoring points belonging to the bridge. Therefore, the judgment can be made by combining high-resolution optical imagery and SAR satellite imagery.

[0093] Using the above-described feature point selection method, to ensure that the selected feature points definitely belong to bridges, one method is used to view ground feature type information, and another is used to view ground feature intensity information. Combining these methods ensures accuracy. Based on the high-resolution optical imagery (reflecting ground feature type information) and SAR satellite imagery (reflecting ground feature intensity information), the distribution of PS points belonging to bridges is selected. The selection results are as follows: Figure 4 As shown, the dual information basis can ensure the accuracy of PS point selection.

[0094] Step S2: Determine the applicability of the target bridge model by combining the deformation monitoring characteristics of PS-InSAR technology (total deformation of bridge due to the superposition of various factors) and the deformation characteristics of the bridge (part of the deformation shows trend and periodic characteristics, while the other part of the deformation caused by the environment shows complexity).

[0095] Specifically, it includes:

[0096] The deformation of bridge structures is mainly affected by permanent and variable forces. Variable forces mainly include vehicle loads, temperature effects, and wind loads. Therefore, the deformation exhibits a certain periodicity and complexity.

[0097] Images acquired by the same SAR satellite covering the same area at very similar times each day. Based on assumptions for practical applications, since vehicle traffic is primarily driven by daily activities and work schedules in a specific region, vehicle loads exhibit significant similarity, which is largely negated when extracting deformation using PS-InSAR technology. Bridge deformation monitored using PS-InSAR technology includes both permanent deformation and deformable deformation influenced by external factors. Based on the above analysis, the influence of vehicle load can be ignored when extracting temporal deformation using PS-InSAR technology. Furthermore, under relatively stable bridge conditions, deformations at feature points caused by factors such as temperature and environmental changes show a certain correlation with deformations monitored using PS-InSAR technology.

[0098] Based on the time of acquisition of SAR satellite images, the highest temperature, lowest temperature, average temperature, weather and meteorological data of the day are obtained. Among them, meteorological data and temperature data are potentially related to wind load, temperature-induced deformation, and periodic deformation caused by environmental influences. This invention simplifies the logic while ensuring the accuracy of the results, making it more suitable for wide-ranging promotion and practical application.

[0099] Perform correlation analysis between each feature point and temperature data X and weather data Y;

[0100] The formula for calculating correlation is:

[0101]

[0102] After calculating the correlation between each feature point and image factors, if the proportion of feature points with a correlation coefficient greater than 0.3 is greater than 80%, then it meets the requirements for subsequent model construction.

[0103] Step S3: Acquisition of influencing factor data and preprocessing of the acquired raw influencing factor data to meet the model processing requirements;

[0104] Specifically, it includes:

[0105] Based on the assumptions and error analysis in step 2, the main error influencing factors are determined to be temperature factor and meteorological factor;

[0106] Temperature-related factors include the highest, lowest, and average temperatures of the day the images were acquired;

[0107] Meteorological factors refer to the weather conditions of the day, mainly categorized as light rain, cloudy, sunny, overcast, showers, and heavy rain;

[0108] Meteorological data is digitized by performing one-hot encoding;

[0109] Temperature data and meteorological data are concatenated and used as input data for model training.

[0110] Preprocessing of training data is an important step in model training, such as... Figure 5 As shown, it illustrates an example of preprocessed training data.

[0111] Step S4: Design a training model based on data features;

[0112] Specifically, it includes:

[0113] It can be calculated using a multiple linear regression model or a ridge regression equation. The choice can be made based on the actual situation. Generally, a multiple linear regression model is sufficient. However, a ridge regression model is preferred when the preprocessed data is significantly affected by noise.

[0114] The equation for the multiple linear regression model is:

[0115]

[0116] The ridge regression equation is:

[0117]

[0118] The relevant loss function is:

[0119]

[0120] In the above formula, y represents the actual long-term deformation value, x represents the input influencing factor data, and the others are variables to be trained.

[0121] Step S5: Further process the preprocessed data, train the model multiple times, and finally merge the multiple sets of fitted parameters to generate the final parameters;

[0122] Specifically, it includes:

[0123] During model training, since the shorter the time span in the monitoring period, the greater the proportion of deformation caused by temperature and meteorological changes in the deformation results obtained from the monitoring, earlier temperature data and environmental data are given higher weights during the model training process.

[0124] Acquire multiple images covering the target bridge, ensuring a quantity of more than 30 images;

[0125] The correlation between the temporal deformation and influencing factors of each selected feature point data is calculated and sorted from largest to smallest absolute value. The top 50% are selected as fitting feature points (points with low correlation are caused by two situations: one is that the local long-term deformation value is large, and the other is that the local noise is too large to fit). In this method, some feature points with large deformation and noise points are effectively removed so that the remaining feature points can better fit the deformation features caused by environmental and other factors.

[0126] For each feature point, a separate fit is performed. The relevant data for each feature point are grouped sequentially, such as group 1 for image number 1 to 24, group 2 for image number 1 to 25, group 3 for image number 1 to 26, and so on, to form n groups of training data (the earlier the data is acquired, the greater its influence on the result, i.e., it is given a higher weight).

[0127] The difference between each data set and the previous item is used to form the corresponding difference data, which is the input item for each training session.

[0128] Based on the PS-InSAR deformation monitoring results, the deformation difference value generated by each feature point in the two SAR images is obtained, and the deformation fitting result data corresponding to each set of input data is obtained;

[0129] Based on multiple sets of data for each feature point, the fitting training is performed on multiple sets of data for each feature point in sequence;

[0130] The average of the coefficients obtained from fitting each set of data is used as the final correction coefficient for the corresponding feature point.

[0131] Step S6: Remove the fitting error of each group from the original deformation monitoring results to obtain the actual cumulative deformation monitoring results of the target PS point;

[0132] Specifically, it includes:

[0133] The deformation feature points on the bridge are divided into two categories: one category belongs to those that have undergone fitting calculation, denoted as category A points, and the other category belongs to those that have not undergone fitting calculation, denoted as category B points.

[0134] For points of type A, the corresponding fitting equation is applied to calculate the error caused by the variable action in each time period, and then this part of the error is removed in each time period;

[0135] For B-type points, search for the three closest A-type points to the target point, determine their respective weights using the inverse distance weighting method, calculate the coefficients of the fitting equation corresponding to the B-type points, and then remove the error using a method similar to that used for A-type points.

[0136] like Figure 6 This shows the time series results of the corrected values ​​obtained by fitting a target PS point and the original time series deformation results;

[0137] like Figure 7 The result shows the time-series deformation of the feature point after non-permanent deformation removal using this method. This result is of great significance for judging the health of the bridge.

[0138] The monitoring method involved in this invention is not only applicable to deformation monitoring in landslide scenarios, but also applicable to deformation monitoring in other scenarios.

[0139] Example 2:

[0140] like Figure 8 As shown, this invention provides an embodiment of error removal for bridge deformation monitoring results based on PS-InSAR technology. The system includes:

[0141] Deformation information acquisition module 10: Based on PS-InSAR technology, it extracts deformation information of the monitored object through radar images. The deformation information includes the PS points of the bridge.

[0142] Model training module 20: Design the corresponding training model according to the established model training formula, and fit the parameters according to the factor data information;

[0143] Processing module 30: Processes the acquired deformation information data and digitally processes the information of the influencing factors, compares the fitting results with the original detection information and removes the incorrect ones.

[0144] Although the present invention has been described in detail above with general descriptions and specific embodiments, modifications or improvements can be made to it, which will be obvious to those skilled in the art. Therefore, all such modifications or improvements made without departing from the spirit of the present invention fall within the scope of protection claimed by the present invention.

Claims

1. A method for error elimination in bridge deformation monitoring results based on PS-InSAR technology, characterized in that, The method includes the following steps: Step S1: Filter the SAR image set, then obtain the deformation monitoring results of the area where the target bridge is located based on PS-InSAR technology, and finally confirm the PS deformation points on the bridge based on optical satellite imagery and SAR satellite imagery. Step S2: Combine the deformation monitoring characteristics of PS-InSAR technology with the deformation characteristics of the bridge to determine the applicability of the feature points of the target bridge model; Step S3: Obtain influencing factor data and preprocess the obtained raw influencing factor data to meet the model processing requirements; Step S4: Design a training model based on data features; Step S5: Further process the preprocessed data, train the model multiple times, and finally merge the multiple sets of fitted parameters to generate the final parameters; Step S6: Remove the fitting error of each group from the original deformation monitoring results to obtain the actual cumulative deformation monitoring results of the target PS point; In step S2, the method for determining applicability includes: Identify the influencing factors on the bridge structure and determine the correlation between the deformation caused by the influencing factors at the feature points and the deformation monitored based on PS-InSAR technology. Perform correlation analysis between each feature point and the influencing factor data; Step S5 specifically includes: Acquire multiple images covering the target bridge, ensuring a quantity of more than 30 images; The temporal deformation and influencing factors of each feature point data are coherently calculated and arranged in descending order of absolute value. The top 50% are selected as the fitted feature points. Feature points with large deformation and noise points are effectively removed to obtain the feature points that fit perfectly. Each selected feature point is fitted individually, and the data related to each feature point are grouped sequentially, and so on to form n groups of training data. The earlier the data is obtained, the greater its influence on the result, that is, it is assigned a higher weight. The difference between each data set and the previous item is used to form the corresponding difference data, which is the input item for each training session. Based on the PS-InSAR deformation monitoring results, the deformation difference value generated by each feature point in the two SAR images is obtained, and the deformation fitting result data corresponding to each set of input data is obtained; Based on multiple sets of data for each feature point, the fitting training is performed on multiple sets of data for each feature point in sequence; The average of the coefficients obtained from fitting each set of data is used as the final correction coefficient for the corresponding feature point.

2. The error elimination method for bridge deformation monitoring results based on PS-InSAR technology as described in claim 1, characterized in that, In step S1, the image selection method includes: Acquire a cropped set of SAR images covering the target bridge; Acquire image intensity data; Calculate the average intensity value for each image; The threshold is set based on the intensity value distribution statistics. In the data distribution statistics, we mainly look at the fluctuation range of the average intensity value of most images and set the lower limit of this range as the threshold. Delete current SAR images that are significantly below the set threshold.

3. The error elimination method for bridge deformation monitoring results based on PS-InSAR technology as described in claim 1, characterized in that, In step S2, the method for determining applicability includes: Identify the influencing factors on the bridge structure and determine the correlation between the deformation caused by the influencing factors at the feature points and the deformation monitored based on PS-InSAR technology. Perform correlation analysis between each feature point and the influencing factor data; The formula for calculating the correlation analysis is as follows: After calculating the correlation between each feature point and image factors, if the proportion of feature points with a correlation coefficient greater than 0.3 is greater than 80%, then it meets the requirements for subsequent model construction.

4. The error elimination method for bridge deformation monitoring results based on PS-InSAR technology as described in claim 3, characterized in that, The influencing factors mainly include vehicle load, temperature effect, and wind load.

5. The error elimination method for bridge deformation monitoring results based on PS-InSAR technology as described in claim 1, characterized in that, Step S3 specifically includes: By analyzing assumptions and errors, influencing factors are identified, and data on these factors are collected. One-hot encoding is performed on the influencing factor data to form digital tags; The digitally labeled data of influencing factors are spliced ​​together to form the input data for model training.

6. The error elimination method for bridge deformation monitoring results based on PS-InSAR technology as described in claim 1, characterized in that, Step S4 specifically includes: The design of the training model can be obtained by calculating a multiple linear regression equation; the multiple linear regression equation is: The loss compensation was confirmed through calculation, and the relevant loss function is as follows: Where y is the actual long-term deformation value and x is the input influencing factor data.

7. The error elimination method for bridge deformation monitoring results based on PS-InSAR technology as described in claim 6, characterized in that, Step S4 specifically includes: The design of the training model can also be derived using the ridge regression equation; the ridge regression equation is: Where y is the actual long-term deformation value and x is the input influencing factor data.

8. The error elimination method for bridge deformation monitoring results based on PS-InSAR technology as described in claim 1, characterized in that, Step 6 specifically includes: The deformation feature points on the bridge are divided into two categories: one category belongs to those that have undergone fitting calculation, denoted as category A points, and the other category belongs to those that have not undergone fitting calculation, denoted as category B points. For points of type A, the corresponding fitting equation is applied to calculate the error caused by the variable action in each time period, and then this part of the error is removed in each time period; For B-type points, search for the three closest A-type points to the target point, determine their respective weights using the inverse distance weighting method, calculate the coefficients of the fitting equation corresponding to the B-type points, and then remove the error using a method similar to that used for A-type points.

9. An error elimination system for bridge deformation monitoring results based on PS-InSAR technology, the system being used to execute the method as described in claim 1, characterized in that, The system includes: Deformation information acquisition module: Based on PS-InSAR technology, the deformation information of the monitored object is extracted from radar images. This deformation information includes the PS points of the bridge. Model training module: Design the corresponding training model according to the established model training formula, and fit the parameters based on the factor data information; Processing module: Processes the acquired deformation information data and digitizes the information of the influencing factors, compares the fitting results with the original detection information and removes the incorrect ones.

Images