InSAR data adaptive automatic processing method and system fusing human-computer interaction feedback

Abstract

The application relates to the technical field of InSAR data processing, and discloses an InSAR data adaptive automatic processing method and system fusing human-computer interaction feedback, which comprises the following steps: establishing a processing task and generating an initial parameter set according to a user-selected SAR image set, an analysis area and a processing target; performing preprocessing and differential interference processing to obtain interference timing data and process evaluation information; determining a target quality index according to a preset evaluation threshold, calculating a parameter correction amount based on a normalized quality deviation in a preset parameter range and generating a parameter adjustment suggestion, iteratively reprocessing from an affected processing stage after receiving user confirmation information until a preset evaluation threshold is reached, result acceptance information is received or a maximum iteration number is reached; and performing time sequence processing to generate a deformation result. The application can realize visual selection, automatic processing and parameter adaptive adjustment of InSAR data, and reduces the repeated calculation amount and the operation complexity.

Description

Technical Field

[0001] This invention relates to the field of InSAR data processing technology, and in particular to an adaptive automated processing method and system for InSAR data that integrates human-computer interaction feedback. Background Technology

[0002] Synthetic Aperture Radar Interferometry (SAR) can acquire surface deformation information from multi-temporal SAR images of the same area. Its processing typically involves multiple stages, including data selection, raw data transcription, orbit correction, geocoding, image registration, baseline calculation, image interferometry, phase filtering, phase unwrapping, atmospheric phase processing, and time series analysis. These different processing stages are highly dependent on each other, and different processing parameters are usually required for different study areas, satellite orbits, number of images, surface cover types, and observation periods.

[0003] In existing processing methods, data selection, processing parameter configuration, program execution, task status monitoring, and result export are typically handled by multiple independent programs. Processors need to transfer data between different programs and modify parameters multiple times based on intermediate results. When image registration accuracy is insufficient, regional coherence is low, the effective range of phase unwrapping is insufficient, or atmospheric phase influence is strong, automated processing methods with fixed parameters struggle to provide clear guidelines for parameter modification based on intermediate processing quality. Furthermore, after parameter changes, unrelated preceding processing steps may need to be re-executed, leading to operational complexity and wasted computational resources. Summary of the Invention

[0004] The main objective of this application is to provide an adaptive automated processing method and system for InSAR data that integrates human-computer interaction feedback. This aims to solve the technical problems of traditional InSAR data processing methods, such as dispersed workflows, complex parameter configurations, difficulty in intuitively evaluating intermediate processing quality, and the need to repeatedly execute a large number of irrelevant processing steps after parameter modifications.

[0005] To achieve the above objectives, this invention provides an adaptive automated processing method for InSAR data that integrates human-computer interaction feedback. The method includes the following steps:

[0006] Acquire the user-selected SAR image set, analysis area, and processing target; establish the processing task and generate the initial parameter set.

[0007] Based on the processing task and the initial parameter set, preprocessing and differential interferometry are performed on the SAR image set to obtain interferometric time-series data and process evaluation information;

[0008] Based on the process evaluation information and the corresponding preset evaluation threshold, the target quality index is determined. Within the preset parameter range, the parameter correction amount of the processing parameter to be adjusted is calculated based on the normalized quality deviation of the target quality index. Candidate parameter values ​​and parameter adjustment suggestions are generated. The user's confirmation information on the parameter adjustment suggestions is received. The initial parameter set is updated according to the confirmation information, and reprocessing is performed starting from the processing stage affected by parameter changes. When the key quality index does not reach the preset evaluation threshold and the number of iterations does not reach the preset maximum number of iterations, the parameter correction amount is recalculated based on the updated process evaluation information, and reprocessing is performed to obtain the target interference time series data.

[0009] Time series processing is performed on the target interference time series data to generate deformation results, and the task status and result files associated with the processing task are output.

[0010] To achieve the above objectives, the present invention also provides an InSAR data adaptive automated processing system that integrates human-computer interaction feedback, comprising:

[0011] The task creation module is used to acquire the user-selected SAR image set, analysis area, and processing target, create the processing task, and generate the initial parameter set.

[0012] An automatic processing module is used to perform preprocessing and differential interferometry on the SAR image set based on the processing task and the initial parameter set, to obtain interferometric time-series data and process evaluation information;

[0013] The feedback optimization module is used to determine the target quality index based on the process evaluation information and the corresponding preset evaluation threshold, calculate the parameter correction amount of the processing parameters to be adjusted based on the normalized quality deviation of the target quality index within a preset parameter range, generate candidate parameter values ​​and parameter adjustment suggestions, receive user confirmation information for the parameter adjustment suggestions, update the initial parameter set according to the confirmation information, and perform reprocessing starting from the processing stage affected by parameter changes; when the key quality index does not reach the preset evaluation threshold and the number of iterations does not reach the preset maximum number of iterations, the parameter correction amount is recalculated based on the updated process evaluation information and reprocessing is performed to obtain target interference time series data;

[0014] The output module is used to perform time series processing on the target interference time series data, generate deformation results, and output the task status and output files associated with the processing task.

[0015] One or more technical solutions proposed in this application have at least the following technical effects:

[0016] This application uses visual interaction as the processing entry point and automated processing pipeline as the processing foundation. It organizes visual data selection, data preprocessing, differential interferometry, time series processing, task monitoring, and result export into related processing tasks, enabling users to obtain deformation results from raw SAR data in one stop, thus reducing the operational complexity of InSAR data processing.

[0017] This application statistically analyzes the quality indicators generated by data preprocessing and differential interferometry, normalizes the deviation of the quality indicators from the preset evaluation threshold, and dynamically calculates the parameter correction amount within the preset parameter range. Through trial calculations in the sample area and comprehensive evaluation, parameter adjustment suggestions are generated, allowing processing personnel to view the recommended parameters, sample area processing results, and estimated processing time in a graphical interface, and then confirm or modify the recommended parameters. This forms a parameter optimization process of manual initial value confirmation, system adaptive fine-tuning, and quality feedback iteration.

[0018] The preset parameter range in this application is based on previous manual processing tasks and parameter impact experiments. The initial parameter values ​​are recommended by the processing platform in conjunction with the current data conditions and confirmed by the user. During the processing, a preset evaluation threshold and a maximum number of iterations are used as termination constraints, and the current optimal result with the highest comprehensive evaluation score is retained when the maximum number of iterations is reached. At the same time, the affected processing stages are determined according to the parameters confirmed by the user, and intermediate processing results that are not affected by parameter changes are retained, which can reduce the amount of redundant calculations caused by parameter adjustments.

[0019] This application can display task statuses such as calculation in progress, calculation completed, and calculation exception in the task list, and associate the logs, submission time, executors, coverage area, and result files of each processing stage with the processing task, which facilitates quick location of exceptions and viewing and exporting processing results. Attached Figure Description

[0020] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0021] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the accompanying drawings used in the description of the embodiments are briefly introduced below. For those skilled in the art, other drawings can be obtained based on these drawings without any creative effort.

[0022] Figure 1 This is a flowchart illustrating an embodiment of the InSAR data adaptive automated processing method of this application;

[0023] Figure 2 This is a schematic diagram illustrating the process of data selection and processing for this application.

[0024] Figure 3 This is a schematic diagram of the preprocessing and differential interferometry processing of this application;

[0025] Figure 4 This is a schematic diagram illustrating the generation and reprocessing of parameter adjustment suggestions for this application;

[0026] Figure 5 This is a schematic diagram of the time series processing and output process of this application;

[0027] Figure 6 This is a schematic diagram of the module structure of the InSAR data adaptive automated processing system of this application.

[0028] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0029] It should be understood that the specific embodiments described herein are used to explain the technical solutions of this application and are not intended to limit this application. For those skilled in the art, equivalent substitutions or adaptive adjustments made based on the technical content disclosed in this application can all be used as executable implementation methods of this application.

[0030] To facilitate understanding of the technical solution of this application, a detailed description of this application is provided below with reference to the accompanying drawings. The executing entity in this embodiment can be a server, workstation, computing node, or a data processing platform composed of multiple computing nodes, which has image processing, data storage, task scheduling, network communication, and graphical interface display functions. The following description uses an InSAR data adaptive automated processing platform as the executing entity.

[0031] like Figure 1 As shown, the InSAR data adaptive automated processing method with integrated human-computer interaction feedback provided in this application embodiment includes steps S100 to S400.

[0032] Step S100: Obtain the SAR image set, analysis area and processing target selected by the user, establish the processing task and generate the initial parameter set.

[0033] It should be noted that the SAR imagery in this embodiment can be Sentinel series satellite SAR imagery, or other SAR imagery capable of providing repeating orbit observation data. For Sentinel SAR imagery, data sheets can be represented by orbit path numbers and sheet numbers, and the orbit direction can include ascending and descending orbits. The analysis area can be directly selected by the user in the map interface, or it can be formed by uploading compressed vector area files or map marker files.

[0034] like Figure 2As shown in this embodiment of the invention, the processing platform first loads a base map and displays an administrative region layer, satellite orbital data sheets, satellite orbital data sheets, and an analysis area layer on the map interface. Each image sheet is represented by a bounding box on the map. Users can control the display status of different layers through the layer tree on the left and can directly select the target image sheet, orbital direction, and time range. Through this map-based selection method, users can intuitively view the spatial relationship between the analysis area and the coverage area of ​​the satellite imagery.

[0035] When users need to define an analysis region using an existing region file, they can upload a compressed vector region file or a map marker file to the processing platform. The processing platform checks the uploaded file type, file size, internal file composition, coordinate reference information, and region boundaries, and then parses the analysis region from the uploaded file. For files containing multiple areal regions, the multiple areal regions can be merged into one analysis region, or multiple processing tasks can be created separately according to the user's selection.

[0036] Specifically, the processing platform queries the image catalog based on the analysis area to obtain candidate images that spatially overlap with the analysis area, and calculates the coverage ratio between each candidate image and the analysis area. When a candidate image cannot cover the analysis area within a preset ratio, the candidate image is removed from the candidate results. When the analysis area spans multiple map sheets, multiple map sheets with the same or similar imaging time can be grouped together so that the combined map sheet covers the analysis area.

[0037] The processing platform further filters candidate images based on the user-selected orbit direction, orbit path number, map sheet number, and imaging time range, and checks the completeness of the candidate image data files and metadata. For sentinel SAR images, corresponding map sheets can be filtered by path and frame, and ascending or descending orbit data can be filtered by direction. The processing platform sorts the filtered candidate images according to imaging time to form a SAR image set.

[0038] In practical applications, the processing platform also checks the number of images and the intervals between adjacent images in the SAR image set. When the number of images is lower than a preset minimum, or when there are missing images between adjacent images exceeding a preset time interval, the platform displays the time period of insufficient images on the interface and prompts the user to expand the date range, adjust the map sheet, or change the orbital direction. After the user adjusts the filtering criteria, the processing platform re-queries the image catalog and updates the SAR image set.

[0039] The processing objective represents the type of output and spatial resolution the user desires. Output type can include deformation rate output, cumulative deformation output, or both. Output spatial resolution determines the spatial scale used for subsequent multi-view processing and output raster generation. For example, when analyzing settlement over a large area, a relatively large number of multi-views can be used to improve phase stability; when analyzing local slopes or engineering areas, a relatively small number of multi-views can be used to preserve local spatial details.

[0040] The processing platform writes the SAR image set, analysis area, and processing target into the task record, assigns a task number to the task record, and creates a processing task based on the task record. The task record may also include the task name, creator, creation time, orbit direction, orbit path number, map sheet number, imaging time range, and analysis area range. After the user clicks "Create Task," the processing platform sets the task status to "Pending Calculation" and adds the processing task to the task queue.

[0041] The processing platform generates an initial parameter set according to a preset parameter template. The preset parameter template stores the parameter range, initial recommended value, associated quality index, adjustment direction, and adjustment coefficient for each adjustable processing parameter. The parameter range is determined based on effective parameters used in previous manual processing tasks and parameter impact experiments conducted for different terrain, vegetation, and climate conditions, and is limited to the parameter value range allowed by the corresponding processing procedure.

[0042] For parameters that can be directly determined by the processing task, such as the number of images, orbital data catalog, and analysis area files, the processing platform automatically writes them according to the task records. For adjustable processing parameters such as the number of multi-look images, temporal baseline range, spatial baseline range, filter intensity, effective coherence threshold, unwrapping reference coordinates, and pre-unwrapping processing step size, the processing platform provides initial recommended values ​​from the parameter range based on the SAR image bands, revisit period, and the terrain, vegetation, and climate conditions of the analysis area. Users can confirm or modify the initial recommended values ​​based on their processing experience to form the initial parameter values ​​for the corresponding adjustable processing parameters.

[0043] The initial parameter set includes data preprocessing parameters, differential interferometry parameters, time series parameters, and result export parameters. Data preprocessing parameters include the number of images, master image identifier, orbital data directory, analysis area file, number of multiviews in the azimuth direction, number of multiviews in the range direction, initial search range, number of matching regions, and size of the matching regions. Differential interferometry parameters include the temporal baseline range, spatial baseline range, filter intensity, effective coherence threshold, unwrapping reference coordinates, pre-unwrapping processing step size, number of multiviews, and number of interferograms used for overlay processing. Time series parameters include orbital number or result name, time series processing scale, result multiview parameters, and result display range.

[0044] The number of images in the initial parameter set is automatically counted by the processing platform based on the SAR image set. The main image identifier can be formed using the imaging date of the main image, and its format can be a consecutive arrangement of year, month, and day. The processing platform can automatically recommend main images or receive main images specified by the user. When automatically recommending main images, the processing platform counts the time interval, spatial baseline, effective coverage ratio, and data integrity status between each image and the other images, prioritizing images located in the middle of the observation period and with relatively small overall differences from the other images.

[0045] The orbit data catalog records the location of precise orbit files. The processing platform can match the precise orbit files corresponding to each SAR image from a pre-configured orbit file library, or it can obtain precise orbit files by calling a preset download program. When the user chooses not to use precise orbit files, the processing platform uses the orbit information attached to the image to complete the initial geometric processing, and further corrects the baseline based on the phase residual characteristics in subsequent differential interferometry processing.

[0046] The analysis area file parameter records the area file corresponding to the analysis area. This area file can be a map marker file uploaded by the user, or an area file generated by the processing platform based on map rendering results. The number of multiviews in the azimuth and distance directions can be configured according to the spatial scale of the observed object. For example, a multiview combination of 4:1, 5:1, 8:2, or 10:2 in the azimuth and distance directions can be used. The above values ​​are executable examples and do not constitute a limitation on the number of multiviews.

[0047] Step S200: Based on the processing task and the initial parameter set, perform preprocessing and differential interferometry on the SAR image set to obtain interferometric time-series data and process evaluation information.

[0048] like Figure 3 As shown in this embodiment of the invention, automated data processing first performs data preprocessing. Data preprocessing includes raw data transcription, orbit correction, geocoding, image registration, analysis area cropping, multi-view processing, and baseline calculation. These multiple processing steps are sequentially invoked by the processing platform based on an initial parameter set, eliminating the need for the user to initiate each processing step individually.

[0049] Raw data transcription is used to read raw SAR data and its metadata, and organizes the image data, imaging time, orbit information, polarization information, and imaging modes from different data packets into an internal data structure that can be read by subsequent processing programs. For SAR products containing multiple sub-bands or multiple data blocks, the processing platform merges the corresponding data according to the imaging order and performs preliminary cropping of the data range according to the analysis area to reduce the amount of data for subsequent processing.

[0050] Orbit correction is used to improve the accuracy of satellite position and velocity information. When the initial parameter set is configured with a precise orbit data catalog, the processing platform matches a precise orbit file covering the corresponding imaging time period from the orbit data catalog according to the imaging time of each SAR image, and uses the precise orbit file to update the satellite position and velocity; when no corresponding precise orbit file is matched, the processing platform records the missing orbit file in the task log, and suspends the task according to the user-preset execution rules or continues processing using orbit information attached to the image.

[0051] Geocoding is used to establish the conversion relationship between radar image coordinates and ground geographic coordinates. The processing platform reads digital elevation data, orbit information, and imaging parameters, calculates and analyzes the distance and azimuth positions of ground points within the area in each SAR image, and generates geometric auxiliary data. This geometric auxiliary data is used for subsequent image registration, terrain phase calculation, and geographic coordinate transformation of deformation results.

[0052] Image registration includes initial registration and fine registration. Initial registration calculates the initial positional offset of the source image relative to the master image based on orbital information, imaging parameters, and digital elevation data, ensuring that corresponding ground features fall within a preset search range. Fine registration selects multiple matching regions with significant scattering characteristics within the analysis area, calculates the offset of each matching region in the distance and azimuth directions, removes matching regions whose offsets significantly deviate from the overall distribution, and then establishes positional correction relationships based on the retained matching regions, resampling the source image to the master image coordinates.

[0053] The processing platform statistically analyzes the positional residuals, the number of effective matching regions, and the effective coverage ratio of each matched region after fine registration. When the positional residuals are not greater than the preset registration threshold and the number of effective matching regions reaches the preset number, the corresponding image is added to the registered image sequence. When the positional residuals exceed the preset registration threshold, insufficient registration quality is recorded in the process evaluation information, and the corresponding image, matching region, and current registration parameters are saved for subsequent parameter adjustment suggestions.

[0054] Multi-view processing is used to reduce the influence of random phase and adjust the spatial resolution of the results. The processing platform performs neighborhood combination on the registered complex images based on the number of azimuth and range multi-views in the initial parameter set to generate multi-view images. To ensure parameter matching in the differential interferometry stage, the number of multi-views used in the data preprocessing stage is written into the parameter record of the intermediate processing results, and the same number of multi-views is read in the differential interferometry stage.

[0055] Baseline calculation is used to obtain the time intervals and spatial baselines between images. The processing platform uses any two images from the SAR image set as candidate combinations and calculates their imaging time interval, vertical baseline, effective overlap area, and data integrity status. When a candidate combination meets the preset time interval, preset baseline range, and preset coverage ratio, the candidate combination is added to the interferometric pair set. The processing platform also checks the number of connections at each imaging time in the interferometric pair set to avoid images that do not form interferometric pairs with other imaging times.

[0056] After data preprocessing, the processing platform performs differential interferometry (DI). DI includes image interferometry, adaptive filtering, phase unwrapping, atmospheric phase processing, baseline residual phase processing, and overlay processing. The processing platform reads the unwrapped reference coordinates, pre-unwrapping step size, number of multi-look views, and number of interferograms from the initial parameter set, and continuously performs differential interferometry according to the corresponding parameters.

[0057] During image interferometry, the processing platform performs conjugate multiplication of the complex pixels of two registered images within the same interferometric pair to obtain the original interferometric phase. Based on geometric auxiliary data and digital elevation data, the topographic phase is calculated and subtracted from the original interferometric phase to obtain an initial set of interferograms. Simultaneously, the processing platform calculates the coherence level map for each initial interferogram; the coherence level is used to represent the correlation level between adjacent observation signals.

[0058] During adaptive filtering, the processing platform sets the filtering intensity based on the local coherence of each pixel and its neighborhood. Weaker smoothing is applied to regions with high coherence to preserve local phase details; in regions with low coherence but still containing effective phase information, the participation of stable phase components is increased to reduce the influence of random phase; regions with coherence below a preset effective threshold are marked so that these regions have lower participation in subsequent phase unwrapping and time-series processing.

[0059] Phase unwrapping is used to convert periodic phases into continuous phases. The unwrapping reference point can be specified by the user on a map or in the result image, or it can be automatically selected by the processing platform from a stable region. The processing platform selects pixels with high average coherence, gentle phase changes, and a large number of effective observations within a preset range around the specified coordinates to form the unwrapping reference region, and performs phase unwrapping on the filtered interferogram starting from this region.

[0060] The pre-unwrapping processing step size is used to control the processing scale in the initial unwrapping stage. The processing platform can first establish a low-resolution phase structure with a relatively large processing step size, and then use the low-resolution processing result as a reference for subsequent fine unwrapping, thereby improving the stability of large-scale phase connections. The pre-unwrapping processing step size can be set according to the analysis area range and the distribution of low-coherence regions, and step S300 will propose adjustment suggestions based on unwrapping quality indicators.

[0061] In one optional implementation, for cases where the analysis area simultaneously includes densely built-up areas, vegetated areas, and slope deformation areas, the processing platform divides the analysis area into stable scattering areas, low-coherence transition areas, and deformation-sensitive areas based on the coherence map, local phase change amplitude, effective scattering pixel density, and terrain slope before performing adaptive filtering and phase unwrapping. An overlapping processing band of preset width is set on both sides of the boundary of different areas, and the overlapping processing band is used to connect the processing results obtained from different areas.

[0062] Specifically, for stable scattering regions, the processing platform employs a smaller filtering neighborhood and lower filtering intensity to preserve local phase details near buildings and engineering structures. For low-coherence transition regions, the platform progressively expands the filtering neighborhood according to the local coherence level, and prioritizes neighboring pixels distributed along the phase continuity direction based on the main direction of phase change within the neighborhood. For deformation-sensitive regions, the platform identifies local areas with rapid phase changes based on the preliminary deformation rate obtained from the stacking process, and restricts the interaction of pixels located on either side of these local areas. Compared to applying a uniform filtering scale to the entire analysis area, this regionalized processing improves the phase stability of low-coherence regions and reduces detail loss caused by excessive smoothing at local deformation edges.

[0063] During phase unwrapping, the processing platform selects unwrapping reference regions within each region and prioritizes extending continuous phase from stable scattering regions to adjacent deformation-sensitive regions and low-coherence transition regions. For mutually separated effective unwrapped regions, the processing platform selects pixels within the overlapping processing band whose coherence reaches a preset threshold and whose phase change amplitude is lower than a preset threshold. It then calculates the phase difference between the unwrapping results of different regions at the selected pixels and adjusts the unwrapping result of one region according to the integer period interval with the highest occurrence rate, thereby connecting the mutually separated effective unwrapped regions.

[0064] When the number of pixels meeting the conditions within the overlapping processing band is lower than a preset number, or the phase difference at the boundary after connection is still higher than a preset threshold, the processing platform only expands the pre-unwrapping processing step size for the corresponding low-coherence transition region and re-executes adaptive filtering and phase unwrapping; other regions continue to use the already obtained processing results. By employing partitioning parameters, overlapping connections, and local reprocessing for complex surface cover areas, the integrity of the phase connection between stable scattering regions and low-coherence regions can be improved, while reducing the data range and computation time for reprocessing.

[0065] During atmospheric phase processing, the processing platform analyzes the spatial and temporal variations of the unwrapped phase. Phase components that vary continuously over a large spatial area and change rapidly between adjacent imaging times are treated as atmospheric phase candidate components and reduced. Phase components that persist across multiple imaging times and exhibit relatively stable changes over time are retained as deformation candidate components. The processing platform also analyzes the correlation between the unwrapped phase and terrain height, and reduces terrain-related residual phases in zones based on terrain height.

[0066] For processing tasks that do not utilize precise orbital data or have biased baseline estimations, the processing platform adjusts the baseline parameters based on the variation characteristics of the pre-correction phase in the range and terrain directions, and recalculates the corresponding terrain residual phase. After adaptive filtering, phase unwrapping, atmospheric phase processing, and baseline residual phase processing, a set of corrected interferograms is formed.

[0067] The processing platform selects a preset number of interferograms from the calibrated interferogram set and performs overlay processing to obtain a preliminary estimate of the surface deformation rate. The number of interferograms used for overlay processing is less than the total number of interferograms generated and can be set according to the number of images in the SAR image set. In one executable implementation, the number can be set to be no less than half the number of images, while not exceeding the maximum number of interferograms that can maintain the correlation between the main imaging times.

[0068] The processing platform organizes the set of corrected interferograms according to the imaging time, forming interferometric time-series data, and statistically analyzes the quality indicators generated by data preprocessing and differential interferometry. Process evaluation information may include registration quality indicators, coherence quality indicators, unwrapping quality indicators, phase correction quality indicators, and preliminary rate quality indicators.

[0069] Registration quality indices can be formed by location residuals, the number of effective matching regions, and the effective coverage ratio; coherence quality indices can be formed by average coherence, the proportion of low-coherence regions, and the proportion of effective coherence regions; unwrapping quality indices can be formed by the proportion of effective unwrapped area, the number of separated regions, and the proportion of anomalous phase pixels; phase correction quality indices can be formed by large-scale residual phase intensity, terrain-related residual intensity, and the difference in results between different interferograms; preliminary rate quality indices can be formed by the spatial continuity of the stacking results and the rate difference between different combinations of interferograms.

[0070] The processing platform continuously updates the task status during data preprocessing and differential interferometry processing. The task status can be displayed as "calculating," "calculation completed," or "calculation error," and can further record the current processing stage, submission time, start time, end time, executors, and coverage area. The runtime information output by each processing program is written to the corresponding stage's processing log, and users can view the specific processing stage and results in the task interface.

[0071] Step S300: Determine the target quality index based on the process evaluation information and the corresponding preset evaluation threshold; calculate the parameter correction amount of the processing parameter to be adjusted based on the normalized quality deviation of the target quality index within the preset parameter range; generate candidate parameter values ​​and parameter adjustment suggestions; receive user confirmation information for the parameter adjustment suggestions; update the initial parameter set according to the confirmation information; and perform reprocessing starting from the processing stage affected by parameter changes; if the key quality index does not reach the preset evaluation threshold and the number of iterations does not reach the preset maximum number of iterations, recalculate the parameter correction amount based on the updated process evaluation information and perform reprocessing to obtain target interference time series data.

[0072] like Figure 4 As shown in this embodiment of the invention, the processing platform first compares each quality indicator in the process evaluation information with its corresponding preset evaluation threshold. When the registration position residual is higher than the preset registration threshold, or the number of effective matching regions is lower than the preset number, the registration quality indicator is determined as the target quality indicator; when the average coherence is lower than the preset coherence threshold, or the proportion of low coherence regions is higher than the preset proportion, the coherence quality indicator is determined as the target quality indicator; when the proportion of effective unwrapping area is lower than the preset unwrapping threshold, the unwrapping quality indicator is determined as the target quality indicator. For the coherence quality indicator, the processing platform performs global statistics on the effective pixels of the coherence map within the analysis area to obtain the mean coherence coefficient, the variance of the coherence coefficient, the proportion of effective coherent pixels, and the proportion of low coherence pixels, so that a single local coherence value is not used as the sole basis for parameter adjustment.

[0073] Then, the processing platform pre-stores the correlation between quality indicators and processing parameters. Candidate parameters for registration quality indicators may include the main image, initial search range, number of matching regions, and size of the matching regions; candidate parameters for coherence quality indicators may include temporal baseline range, spatial baseline range, number of multi-look operations, and filter intensity; candidate parameters for unwrapping quality indicators may include unwrapping reference coordinates, effective coherence threshold, and pre-unwrapping processing step size; candidate parameters for phase correction quality indicators may include spatial processing scale, temporal processing scale, and terrain zoning scale. For each continuously adjustable processing parameter, the correlation also records the parameter's minimum value, maximum value, adjustment direction, and adjustment coefficient; the adjustment direction indicates whether the parameter is increased or decreased within the current parameter range to improve the corresponding quality indicator, and the adjustment coefficient is used to limit the magnitude of parameter changes in a single round.

[0074] Next, the processing platform calculates the normalized quality deviation for the target quality indicators. For quality indicators where a larger value indicates higher processing quality, the difference between the preset evaluation threshold and the current indicator value is divided by the difference between the preset evaluation threshold and the preset lower limit. For quality indicators where a smaller value indicates higher processing quality, the difference between the current indicator value and the preset evaluation threshold is divided by the difference between the preset upper limit and the preset evaluation threshold. A result less than 0 is taken as 0, and a result greater than 1 is taken as 1, thus unifying the quality deviations of different dimensions to the range of 0 to 1. When a candidate parameter is associated with multiple target quality indicators, the normalized quality deviations and their corresponding adjustment directions are weighted according to preset weights to obtain the comprehensive deviation direction of the candidate parameter.

[0075] For continuous candidate parameters, the processing platform multiplies the parameter range width, adjustment coefficient, and comprehensive deviation direction of the candidate parameter to obtain the basic parameter correction amount. This basic parameter correction amount is then scaled using multiple preset step size coefficients to form multiple corrected candidate parameter values. Each candidate parameter value is restricted to between the corresponding parameter's minimum and maximum values. The preset step size coefficients can include multiple values ​​less than 1, equal to 1, and greater than 1 to simultaneously form conservative adjustment values, standard adjustment values, and accelerated adjustment values. For non-continuous candidate parameters such as the main image and unwrapped reference area, the processing platform forms a finite candidate set based on the aforementioned correlation, preset parameter templates, and historical task parameters of similar areas, without using continuous parameter correction amounts for calculation.

[0076] The processing platform matches the target quality indicators with the aforementioned correlation to determine the processing stage to be adjusted and candidate parameters. It then uses the parameter correction amount, preset parameter templates, and historical task parameters from similar regions to generate multiple candidate parameter values ​​or combinations thereof. For example, when the average coherence is low, candidate parameter values ​​such as shortening the interferometric pair time interval, reducing the interferometric pair spatial baseline, increasing the number of multiple views, or increasing the filtering intensity in low-coherence regions can be generated. When the effective unwrapping area ratio is low, candidate parameter values ​​such as changing the unwrapping reference region or adjusting the pre-unwrapping processing step size can be generated.

[0077] To reduce the computational burden caused by multiple trial calculations across the entire analysis area, the processing platform selects a pre-defined sample region within the analysis area. This sample region includes areas with high average coherence, areas with low average coherence, areas with high initial deformation rates, and the edge regions of the analysis area. By simultaneously selecting different types of sample regions, the parameter calculation results can reflect the main surface conditions.

[0078] The processing platform applies each candidate parameter value or combination of candidate parameters to the sample region for adjustment stages, obtaining the corresponding sample processing results and statistically analyzing the degree of improvement in quality indicators and the processing time. The degree of improvement in quality indicators represents the effectiveness of the candidate parameters relative to the current parameters in improving the target quality indicator, while the processing time represents the computation time required to execute the corresponding processing stage after adopting the candidate parameters. For non-target key quality indicators, the processing platform also records their changes to avoid significantly decreasing other key quality indicators while trying to improve one.

[0079] The processing platform normalizes the improvement degree of the quality indicators and the processing time for each candidate parameter, and calculates a comprehensive evaluation score according to preset quality weights and time weights. For candidate parameters that cause any non-target key quality indicator to decrease by more than a preset allowable range, the processing platform removes them from the recommendation set. For candidate parameters that cause all target quality indicators to reach the corresponding preset evaluation thresholds, they are prioritized and arranged in descending order of comprehensive evaluation score. When no candidate parameter reaches all preset evaluation thresholds, the remaining candidate parameters are arranged in descending order of comprehensive evaluation score. Therefore, the current optimal in this embodiment represents the parameter with the highest comprehensive evaluation score among the candidate parameters that have been tested in this round and satisfy the above constraints, rather than the global optimal obtained by exhaustively searching all possible parameter values.

[0080] The processing platform generates parameter adjustment suggestions based on the ranking results. Each parameter adjustment suggestion includes the processing stage to be adjusted, the current parameter value, candidate parameter values ​​or combinations of candidate parameters, parameter correction amount, quality indicators of the sample area before and after processing, comprehensive evaluation score, estimated processing time, affected intermediate processing results, and recommended order. Users can view the phase map or coherence map of the sample area before and after processing on the interface, and choose to accept the recommended parameters, modify the recommended parameters, or retain the current parameters; when users modify the recommended parameters, the modified parameters are still limited to the corresponding preset parameter range.

[0081] The candidate parameter values ​​selected or modified by the user are written into the initial parameter set as confirmation information. The processing platform determines the affected processing stage based on the correspondence between the parameters and the processing stages. For example, changing the master image will affect image registration, baseline calculation, image interferometry, and subsequent processing, so the processing platform will re-execute from image registration; modifying the interferometric pair time interval will affect interferometric pair construction and subsequent processing, so the processing platform will retain the registered image sequence and re-execute from baseline calculation and interferometric pair construction.

[0082] When modifying the filter intensity, unwrapping reference coordinates, or pre-unwrapping processing step size, the processing platform retains the initial interferogram set and re-executes from the adaptive filtering or phase unwrapping stage; when only modifying the time series parameters, the processing platform retains the target interferometric time series data and executes from the time series processing stage; when only modifying the result display range or result name, the processing platform retains the deformation results and only regenerates the result file.

[0083] The processing platform performs a file availability check on intermediate processing results prior to the affected processing stage. After confirming that the intermediate processing results correspond to the current processing task, it executes reprocessing starting from the affected processing stage to obtain updated interferometric time series data and update process evaluation information. It also associates and saves the parameter set used in this iteration, the update process evaluation information, the comprehensive evaluation score, and the iteration number. When the key quality indicators in the update process evaluation information reach the corresponding preset evaluation threshold, the updated interferometric time series data is determined as the target interferometric time series data.

[0084] When a key quality indicator fails to reach the preset evaluation threshold, but the user believes the processing result meets the actual analysis requirements after reviewing the processed graph, they can send a result acceptance message. The processing platform then uses this message to update the interferometric time series data and designate it as the target interferometric time series data. If the key quality indicator still fails to reach the preset evaluation threshold and no result acceptance message is received, the processing platform uses the updated process evaluation information as the process evaluation information for the next round, recalculates the normalized quality deviation and parameter correction, and generates parameter adjustment suggestions again. The processing platform presets a maximum number of iterations; when the current iteration count reaches the maximum, it stops generating new parameter corrections and selects the interferometric time series data with the highest comprehensive evaluation score from all completed iterations as the current optimal result, designating it as the target interferometric time series data. Therefore, any one of the following conditions—the key quality indicator reaching the preset evaluation threshold, receiving a user result acceptance message, or reaching the maximum number of iterations—serves as the iteration termination condition.

[0085] In one executable implementation, the processing platform can save processing parameters confirmed by the user under different regions, orbits, and land cover conditions, and use parameters adopted in similar tasks as candidate parameters for subsequent tasks. Parameter templates can be saved by the user and reused in the same map sheet or similar analysis areas, thereby reducing repetitive configuration processes.

[0086] In one optional implementation, for cases where the analysis area spans a large distance or where there are significant differences in land cover between different sub-regions, the processing platform groups the sample areas into multiple sample groups according to spatial location and quality type, and separately calculates the registration quality, coherence quality, unwrapping quality, and preliminary rate quality of each sample group. The processing platform retains only candidate parameter values ​​that improve at least one sample group and do not cause a decrease in key quality indicators of other sample groups exceeding a preset allowable range as valid candidate parameter values. This avoids masking the processing quality changes in local areas by the average result when using the overall regional average quality index.

[0087] For the valid candidate parameter values, the processing platform first performs a rapid trial calculation using sample data with reduced spatial resolution. Then, it performs a verification trial calculation using the original spatial resolution on a preset number of candidate parameter values ​​that rank highly in the rapid trial calculation results. In addition to statistically analyzing the improvement degree of the target quality indicators and processing time, the verification trial calculation also statistically analyzes the effective coverage ratio of the sample area, the continuity of the phase boundary, and the range of changes in the initial deformation rate. All statistical results are displayed side-by-side with the processing results corresponding to the current parameters. Users can confirm candidate parameter values ​​for the entire analysis area or only for one or more sample groups.

[0088] When a user confirms candidate parameter values ​​only for a subset of sample groups, the processing platform generates a local parameter applicability range based on the spatial range of the corresponding sample group, and determines the affected processing stage and affected intermediate processing results based on the changed processing parameters. Intermediate processing results outside the local parameter applicability range are retained by the processing platform; intermediate processing results within the local parameter applicability range and generated by the affected processing stage are recalculated from the corresponding processing stage. A preset width overlap region is set between adjacent local parameter applicability ranges. Within this overlap region, the processing platform selects the output result based on the coherence level and the distance from the pixel to the region boundary, ensuring a continuous connection between the local reprocessing results and the retained results.

[0089] After reprocessing, the processing platform compares the updated process evaluation information with the original process evaluation information for each sample group. When the key quality indicators of the confirmed sample group reach the preset evaluation threshold, and the changes in the key quality indicators of the unconfirmed sample group do not exceed the preset allowable range, the updated interferometric time series data is determined as the target interferometric time series data; otherwise, parameter adjustment suggestions are regenerated only for sample groups that do not meet the conditions. Compared to uniformly modifying parameters and recalculating the entire affected processing stage based on the overall evaluation results of the entire region, this implementation method can retain the spatial areas where the quality requirements have been met and their intermediate processing results, reducing redundant calculations for large-scale processing tasks and minimizing the impact of local parameter adjustments on the processing results of other regions.

[0090] Step S400: Perform time series processing on the target interference time series data to generate deformation results, and output the task status and result file associated with the processing task.

[0091] like Figure 5 As shown, in this embodiment of the invention, time series processing includes effective interferogram selection, time series overlay processing, regression analysis, atmospheric phase correction, spatial filtering, deformation rate calculation, and cumulative deformation calculation. The processing platform first selects effective interferograms based on the average coherence, effective unwrapped area ratio, and anomalous phase ratio corresponding to the target interferometric time series data, and checks whether sufficient observational correlation is maintained between each imaging time.

[0092] When removing low-quality interferograms results in a lack of effective observations at a certain imaging moment, the processing platform selects a relatively high-quality interferogram from the unselected interferograms that can connect to that imaging moment as a supplement. The processing platform establishes a phase observation sequence according to the imaging time and uses the unwrapped reference region as the reference region for time series processing.

[0093] Temporal overlay processing is used to estimate the overall deformation trend based on deformation observations from multiple interferograms, while regression analysis is used to obtain the deformation at each imaging time relative to the reference time and the deformation rate within the observation period. The processing platform sets the participation level of each interferogram according to its coherence quality, effective area, and time interval, so that high-quality interferograms have a greater impact on the calculation results, while low-quality interferograms have a smaller impact.

[0094] Atmospheric phase correction and spatial filtering are used to further reduce the non-deformation phase in the target interferometric time-series data. The processing platform analyzes the phase variation characteristics in spatial range, imaging time, and terrain height, respectively, and reduces the phase components that are continuously distributed in a large spatial range and change rapidly at different imaging times, while retaining the phase components that persist at multiple imaging times and change stably over time.

[0095] The processing platform generates deformation rate and cumulative deformation based on the regression analysis results. Deformation rate represents the average deformation per unit time within the observation period, while cumulative deformation represents the deformation at each imaging moment relative to the reference moment. Pixels with fewer than a preset number of valid observations are marked as invalid pixels; for pixels with sufficient valid observations but relatively low result stability, the corresponding result quality level can be output.

[0096] The processing platform uses the geometric auxiliary data generated in step S200 to transform the deformation rate and cumulative deformation from radar image coordinates to the geographic coordinates corresponding to the analysis area, generating deformation results with geographic location attributes. The pixel size of the resulting raster is determined based on the spatial resolution of the results in the processing target.

[0097] The parameter configuration for time series processing can include track number or custom result name, analysis area file, unwrapping reference coordinates, time series processing scale, number of multiviews, number of interferograms used for time series processing, result multiview parameters, and result display range. In one executable implementation, the result multiview parameter can be fixed at 1, so that the time series results maintain the spatial scale obtained from the previous processing.

[0098] The display range of the deformation rate map and the cumulative deformation map can be configured by the user. For example, the absolute value of the display range can be set between 0.05 and 0.1, and adjusted according to the specific deformed object. The display range is only used to generate the rendered map and does not change the original deformation values ​​saved in the geographic raster file.

[0099] The processing platform automatically generates raster rendering files, geographic raster files, and map browsing files based on the deformation results. Raster rendering files can be in BMP format for quick display of results in the task interface; geographic raster files can be in TIF format to save deformation values ​​and geographic coordinates; and map browsing files can be in KMZ format for overlaying and displaying deformation results in map browsing software.

[0100] The processing task, initial parameter set, intermediate processing results, process evaluation information, confirmation information, target interference time series data, deformation results, task status, and processing log are associated through task number. The processing platform updates the task status to "Calculation in Progress" at the start of each processing stage, updates the processing progress upon completion of the corresponding stage, and updates the task status to "Calculation Completed" upon completion of all stages. If an exception occurs in the processing program, the task status is updated to "Calculation Exception".

[0101] The task list displays the processing stage, task submission time, executors, analysis area coverage, and current status. Users can access the task details page from the task list to view logs for each processing stage. The processing logs record information in chronological order, including program startup, parameter reading, data file reading, intermediate result generation, quality indicator statistics, exception information, and program termination, making the processing process searchable and facilitating anomaly localization.

[0102] When a task status indicates a computational anomaly, the processing platform displays the anomaly handling stage, processing logs, current parameters, and generated intermediate processing results on the task details page. After the user modifies the parameters or confirms to continue processing, the processing platform starts executing subsequent processing from the intermediate processing results saved before the anomaly handling stage, without needing to re-execute processing stages that have already been completed and are unaffected by parameter modifications.

[0103] This application also provides an adaptive automated processing system for InSAR data that integrates human-computer interaction feedback. Please refer to [link / reference]. Figure 6 The InSAR data adaptive automated processing system that integrates human-computer interaction feedback includes:

[0104] The task creation module is used to acquire the user-selected SAR image set, analysis area, and processing target, create the processing task, and generate the initial parameter set.

[0105] An automatic processing module is used to perform preprocessing and differential interferometry on the SAR image set based on the processing task and the initial parameter set, to obtain interferometric time-series data and process evaluation information;

[0106] The feedback optimization module is used to determine the target quality index based on the process evaluation information and the corresponding preset evaluation threshold, calculate the parameter correction amount of the processing parameters to be adjusted based on the normalized quality deviation of the target quality index within a preset parameter range, generate candidate parameter values ​​and parameter adjustment suggestions, receive user confirmation information for the parameter adjustment suggestions, update the initial parameter set according to the confirmation information, and perform reprocessing starting from the processing stage affected by parameter changes; when the key quality index does not reach the preset evaluation threshold and the number of iterations does not reach the preset maximum number of iterations, the parameter correction amount is recalculated based on the updated process evaluation information and reprocessing is performed to obtain target interference time series data;

[0107] The output module is used to perform time series processing on the target interference time series data, generate deformation results, and output the task status and output files associated with the processing task.

[0108] Compared with the prior art, the beneficial effects of the InSAR data adaptive automated processing system with integrated human-computer interaction feedback provided in this application are the same as the beneficial effects of the InSAR data adaptive automated processing method with integrated human-computer interaction feedback provided in the above embodiments. Moreover, the other technical features of the InSAR data adaptive automated processing system with integrated human-computer interaction feedback are the same as the features disclosed in the methods of the above embodiments, and will not be repeated here.

[0109] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.

[0110] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct or indirect applications in other related technical fields, are included in the patent protection scope of this application.

Claims

1. A method for adaptive automated processing of InSAR data integrating human-computer interaction feedback, characterized in that, The method includes the following steps: Acquire the user-selected SAR image set, analysis area, and processing target; establish the processing task and generate the initial parameter set. Based on the processing task and the initial parameter set, preprocessing and differential interferometry are performed on the SAR image set to obtain interferometric time-series data and process evaluation information; Based on the process evaluation information and the corresponding preset evaluation threshold, the target quality index is determined. Within the preset parameter range, the parameter correction amount of the processing parameter to be adjusted is calculated based on the normalized quality deviation of the target quality index. Candidate parameter values ​​and parameter adjustment suggestions are generated. The user's confirmation information on the parameter adjustment suggestions is received. The initial parameter set is updated according to the confirmation information, and reprocessing is performed starting from the processing stage affected by parameter changes. When the key quality index does not reach the preset evaluation threshold and the number of iterations does not reach the preset maximum number of iterations, the parameter correction amount is recalculated based on the updated process evaluation information, and reprocessing is performed to obtain the target interference time series data. Time series processing is performed on the target interference time series data to generate deformation results, and the task status and result files associated with the processing task are output.

2. The InSAR data adaptive automated processing method integrating human-computer interaction feedback as described in claim 1, characterized in that, The steps for acquiring the user-selected SAR image set, analysis area, and processing target, establishing the processing task, and generating the initial parameter set include: It receives the image sheet selected by the user on the map interface, or parses the regional file uploaded by the user to obtain the analysis area; Candidate images are queried based on the analysis area, and the candidate images are filtered according to image size, orbital direction and imaging time to form a SAR image set; Processing targets are formed based on the user-selected output type and spatial resolution; The SAR image set, the analysis area, and the processing target are written into the task record to establish a processing task, and an initial parameter set is generated according to a preset parameter template. The parameters determined by the task information are automatically written into the initial parameter set, and the user confirms the initial parameter values ​​of the adjustable processing parameters within the parameter range defined by the preset parameter template.

3. The InSAR data adaptive automated processing method integrating human-computer interaction feedback as described in claim 2, characterized in that, The steps for generating an initial parameter set according to a preset parameter template include: The number of images in the SAR image set is counted, and the main image is selected from the SAR image set; The system reads the parameter range, initial recommended value, associated quality index, adjustment direction and adjustment coefficient of each adjustable processing parameter in the preset parameter template, generates an initial recommended value based on the SAR image band, revisit period and the terrain, vegetation and climate conditions of the analysis area, and receives confirmation or modification of the initial recommended value by the user within the parameter range. Write the region file corresponding to the analysis region into the region parameters, and set the number of multi-views in the distance direction and the number of multi-views in the azimuth direction according to the spatial resolution of the results; The number of images, the main image identifier, the orbital data directory, the region parameters, the number of multiviews, and the adjustable processing parameters confirmed or modified by the user are written into the initial parameter set.

4. The InSAR data adaptive automated processing method integrating human-computer interaction feedback as described in claim 1, characterized in that, Based on the processing task and the initial parameter set, the steps of performing preprocessing and differential interferometry on the SAR image set to obtain interferometric time-series data and process evaluation information include: The SAR image set is subjected to preprocessing including raw data transcription, orbit correction, geocoding, image registration and multi-view processing in sequence to obtain the registered image sequence and geometric auxiliary data; The baselines between images are calculated based on the registered image sequence, and an interferometric pair set is constructed according to preset baseline conditions; Differential interferometry is performed on the set of interferometric pairs using the aforementioned geometric auxiliary data to obtain interferometric time-series data; The quality indicators generated by the preprocessing and differential interference processing are statistically analyzed to form process evaluation information.

5. The InSAR data adaptive automated processing method integrating human-computer interaction feedback as described in claim 4, characterized in that, The steps of performing differential interferometry on the set of interferometric pairs using the geometric auxiliary data to obtain interferometric time series data include: Perform image interferometry and terrain phase reduction on the set of interferometric pairs to obtain an initial set of interferograms; The filtering intensity is adjusted according to the local coherence of the initial interferogram set to obtain a filtered interferogram set; The unwrapping reference region is determined based on the user-specified coordinates or stable region filtering results. Phase unwrapping is performed on the filtered interferogram set, and atmospheric phase and baseline residual phase are reduced to obtain the corrected interferogram set. A preset number of interferograms are selected from the set of corrected interferograms and overlaid to obtain the preliminary deformation rate. The set of corrected interferograms is then organized according to the imaging time to form interferometric time-series data.

6. The InSAR data adaptive automated processing method integrating human-computer interaction feedback as described in claim 1, characterized in that, The steps of generating parameter adjustment suggestions based on the process evaluation information and receiving user confirmation of the parameter adjustment suggestions include: The quality indicators in the process evaluation information are compared with the corresponding preset evaluation thresholds, and at least one of the following is statistically analyzed: mean coherence coefficient, variance of coherence coefficient, proportion of effective coherent pixels, proportion of effective unwrapped area, and proportion of abnormal phase pixels, to obtain the target quality indicator that has not reached the preset evaluation threshold. The target quality index is matched with the preset quality index and processing parameter correlation to determine the processing stage to be adjusted and the candidate parameters, and the parameter range, adjustment direction and adjustment coefficient of the candidate parameters are read. The normalized quality deviation is calculated based on the current index value of the target quality index, the preset evaluation threshold, and the preset value boundary. The normalized quality deviation, the parameter range width of the candidate parameters, the adjustment direction, and the adjustment coefficient are combined to obtain the parameter correction amount. The corrected parameter value is then restricted to the parameter range to generate multiple candidate parameter values. Within the analysis area, a sample region is selected, and the adjustment process is performed on the sample region using multiple candidate parameter values ​​or combinations of candidate parameters. The improvement degree of quality indicators and processing time corresponding to each candidate parameter value or combination of candidate parameters are statistically analyzed. The improvement degree of the quality indicators and the processing time are normalized and a comprehensive evaluation score is calculated. Candidate parameter values ​​or combinations that cause non-target key quality indicators to decrease by more than a preset allowable range are eliminated. Candidate parameter values ​​or combinations that cause the target quality indicators to reach the preset evaluation threshold are prioritized. If the preset evaluation threshold is not reached, the remaining candidate parameter values ​​or combinations are ranked according to the comprehensive evaluation score. The parameter adjustment suggestions are generated and displayed. The candidate parameter values ​​or combinations selected or modified by the user are received as confirmation information.

7. The InSAR data adaptive automated processing method integrating human-computer interaction feedback as described in claim 6, characterized in that, The steps of updating the initial parameter set according to the confirmation information and performing reprocessing starting from the processing stage affected by parameter changes to obtain target interferometric time series data include: Write the confirmation information into the initial parameter set, record the current iteration number, and match the changed processing parameters with the preset parameter and processing stage correspondence to determine the affected processing stage; Retain intermediate processing results prior to the affected processing stage, and perform reprocessing starting from the affected processing stage to obtain updated interference time series data and updated process evaluation information; The key quality indicators in the evaluation information of the update process are compared with the corresponding preset evaluation thresholds; When the key quality indicator reaches the preset evaluation threshold or the user's result acceptance information is received, the updated interference time series data is determined as the target interference time series data; when the key quality indicator does not reach the preset evaluation threshold, the result acceptance information is not received, and the current iteration number is less than the preset maximum iteration number, the parameter correction amount is recalculated based on the update process evaluation information and parameter adjustment suggestions are generated; when the preset maximum iteration number is reached, the interference time series data with the highest comprehensive evaluation score in each iteration is determined as the current optimal result and used as the target interference time series data.

8. The InSAR data adaptive automated processing method integrating human-computer interaction feedback as described in claim 1, characterized in that, The steps of performing time series processing on the target interferometric time series data to generate deformation results and outputting the task status and result files associated with the processing task include: Valid interferograms are selected based on the quality indicators and interferometric pair temporal relationships corresponding to the target interferometric time series data, and a phase observation sequence is established according to the imaging time. The phase observation sequence is subjected to time-series overlay processing and regression analysis. Based on spatial variation characteristics, temporal variation characteristics and topographic variation characteristics, non-deformation phases are reduced to obtain the deformation and deformation rate at each imaging moment. The deformation amount and the deformation rate are converted to the geographic coordinates corresponding to the analysis area to generate deformation results; The resulting file is generated based on the deformation results, and the task status is generated based on the execution information of each processing stage.

9. The InSAR data adaptive automated processing method integrating human-computer interaction feedback as described in claim 8, characterized in that, The steps of generating a result file based on the deformation result and generating a task status based on the execution information of each processing stage include: The deformation results are converted into raster rendering files, geographic raster files, and map browsing files, and each file is associated with and stored in relation to the processing task. The task status is updated when each processing stage begins, ends, or an anomaly occurs. The task status includes the current processing stage, processing progress, submission time, executor, coverage area, and anomaly information. The processing logs generated at each processing stage are associated with and displayed with the corresponding task status; When an anomaly is indicated in the task status, the system receives a user's instruction to continue processing, starts subsequent processing from the saved intermediate processing results, and updates the task status and the output file.

10. An adaptive automated processing system for InSAR data integrating human-computer interaction feedback, characterized in that, The system includes: The task creation module is used to acquire the user-selected SAR image set, analysis area, and processing target, create the processing task, and generate the initial parameter set. An automatic processing module is used to perform preprocessing and differential interferometry on the SAR image set based on the processing task and the initial parameter set, to obtain interferometric time-series data and process evaluation information; The feedback optimization module determines the target quality index based on the process evaluation information and the corresponding preset evaluation threshold. Within a preset parameter range, it calculates the parameter correction amount of the processing parameters to be adjusted based on the normalized quality deviation of the target quality index, generates candidate parameter values ​​and parameter adjustment suggestions, receives user confirmation information for the parameter adjustment suggestions, updates the initial parameter set according to the confirmation information, and performs reprocessing starting from the processing stage affected by parameter changes. When the key quality index does not reach the preset evaluation threshold and the number of iterations does not reach the preset maximum number of iterations, it recalculates the parameter correction amount based on the updated process evaluation information and performs reprocessing to obtain target interference time series data. The output module is used to perform time series processing on the target interference time series data, generate deformation results, and output the task status and output files associated with the processing task.

Images