A method for denoising ground-based SAR slope measurement data

Abstract

The present invention relates to the field of ground-based SAR technology, and discloses a ground-based SAR slope measurement data denoising method, wherein a ground-based SAR image sequence and a monitoring parameter sequence of a slope to be observed are input into a neural network model for identifying a risk area of ​​the slope, and the risk area in the corresponding ground-based SAR image at each monitoring time point during the monitoring period and subsequent periods is output; the risk area corresponding to each monitoring time point is marked in the ground-based SAR image corresponding to each monitoring time point to determine the key observation area in the ground-based SAR image corresponding to each monitoring time point; according to the key observation area in the ground-based SAR image corresponding to each monitoring time point, a key denoising range is calibrated in the ground-based SAR image; and the ground-based SAR image acquired at each monitoring time point is denoised according to the key denoising range. The present invention is conducive to dividing the key denoising range of the ground-based SAR image.

Description

Technical Field

[0001] The invention relates to the technical field of ground-based SAR, and in particular to a ground-based SAR slope measurement data denoising method. Background Art

[0002] SAR (Synthetic Aperture Radar) is an active earth observation system that can be installed on aircraft, satellites, spacecraft and other flying platforms to observe the earth around the clock and in all weather conditions, and has a certain surface penetration capability. Therefore, the SAR system has unique advantages in disaster monitoring, environmental monitoring, ocean monitoring, resource exploration, crop yield estimation, mapping and military applications, and can play a role that other remote sensing methods cannot play.

[0003] Ground-based SAR remote sensing technology is a new ground-based active microwave remote sensing technology. Compared with traditional monitoring methods, ground-based SAR remote sensing technology has many advantages, such as: all-weather and all-day reconnaissance and surveillance capabilities, large-scale and long-distance reconnaissance and surveillance capabilities, flexible adjustment of observation angles, and strong real-time monitoring capabilities.

[0004] Ground-based SAR remote sensing technology has broad application prospects in the field of deformation monitoring and can be applied to multiple fields, often used for dam monitoring, bridge monitoring, building monitoring, glacier monitoring, and slope monitoring. Ground-based SAR remote sensing technology can achieve long-term, real-time monitoring of targets, and can detect the deformation of targets in a timely manner, providing strong technical support for safety monitoring.

[0005] Most traditional slope monitoring technologies are conducted in a point-to-point manner, which cannot fully analyze the overall slope displacement changes and cannot meet the needs of slope monitoring.

[0006] Ground-based SAR slope monitoring technology can conduct slope monitoring around the clock and in all weather conditions. Especially in areas with higher slope risks or in seasons with higher slope risks, the monitoring frequency will be more intensive, which results in a large amount of image data being generated when using SAR technology for slope measurement.

[0007] Although SAR technology itself has the advantages of high resolution and all-weather and all-day operation, noise is inevitably introduced during its imaging process. This noise mainly comes from the following aspects:

[0008] Coherent speckle noise: This is the most common type of noise in SAR images, caused by the coherent nature of radar waves. It appears as a granular random interference that reduces the quality of the image and affects feature extraction and interpretation.

[0009] Thermal Noise: Thermal disturbances from the sensor electronics can also create noise in the image.

[0010] Multipath Effect: When radar signals are reflected multiple times and arrive at the receiving antenna, they cause additional interference fringes and other types of noise.

[0011] Image denoising of SAR images is relatively complex because: First, the amount of data in SAR images is usually large, especially for images with high resolution or large area coverage. This is not only because the original acquired data is large in itself, but also because in order to improve accuracy, multi-polarization or multi-band data may be used, which increases the data dimension. Second, the computational complexity of SAR images is high: effective SAR image denoising algorithms often involve complex mathematical operations, such as filtering, transform domain processing, etc., which require high computing resources. Third, the time cost is high: Due to the above two factors, the denoising process of SAR images takes a long time, especially when using non-local means (NLM), wavelet transform or other learning-based methods.

[0012] Therefore, a large amount of noise is generated in the image data of SAR technology, which brings difficulties to the subsequent measurement and analysis, and the data processing volume of the denoising process is large and the denoising time is long. Therefore, it is necessary to propose a denoising method for ground-based SAR slope measurement data to divide the key denoising range of ground-based SAR images. Summary of the invention

[0013] The main purpose of the present invention is to provide a ground-based SAR slope measurement data denoising method, aiming to solve the problems in the existing ground-based SAR slope measurement technology that a large amount of noise is generated in the image data, which brings difficulties to the subsequent measurement and analysis, and the data processing amount of the denoising process is large and the denoising time is long.

[0014] To achieve the above object, the present invention provides a method for denoising ground-based SAR slope measurement data, comprising the following steps:

[0015] Acquire a ground-based SAR image sequence and a monitoring parameter sequence of the slope to be observed within a monitoring period of a preset duration, wherein multiple position points of the slope to be observed are respectively provided with identification elements to form identification points;

[0016] The ground-based SAR image sequence and monitoring parameter sequence of the slope to be observed are input into the neural network model for identifying the risk area of ​​the slope, and the risk area in the corresponding ground-based SAR image at each monitoring time point during the monitoring period and subsequent periods is output, wherein the risk area is calculated based on the identification point;

[0017] The risk areas corresponding to each monitoring time point are marked in the ground-based SAR images corresponding to each monitoring time point, so as to determine the key observation areas in the ground-based SAR images corresponding to each monitoring time point;

[0018] According to the key observation areas in the ground-based SAR images corresponding to each monitoring time point, the key denoising range is calibrated in the ground-based SAR images;

[0019] The ground-based SAR images acquired at each monitoring time point are denoised according to the key denoising range.

[0020] Optionally, the step of acquiring a ground-based SAR image sequence and a monitoring parameter sequence of the slope to be observed within a monitoring period of a preset duration includes:

[0021] Obtaining a set number of ground-based SAR images and monitoring parameters collected at set time intervals for the slope to be observed during a monitoring period, wherein the monitoring parameters include meteorological parameters, terrain parameters, vegetation parameters, and human activity impact parameters;

[0022] Arrange a set number of ground-based SAR images within the same monitoring period into a ground-based SAR image sequence in time sequence;

[0023] A set number of monitoring parameters collected within the same monitoring period are arranged in time sequence to form a monitoring parameter sequence of the slope to be observed.

[0024] Optionally, the step of inputting the ground-based SAR image sequence and the monitoring parameter sequence of the slope to be observed into a neural network model for identifying the risk area of ​​the slope, and outputting the risk area in the corresponding ground-based SAR image at each monitoring time point during the monitoring period and subsequent periods includes:

[0025] The ground-based SAR image sequence and monitoring parameter sequence of the slope to be observed are input into a neural network model for identifying the risk area of ​​the slope;

[0026] In the neural network model, the first risk area of ​​the slope to be observed is determined according to the monitoring parameter sequence of the slope to be observed;

[0027] In the neural network model, the influencing factor of the deformation rate of the slope to be observed is determined according to the monitoring parameter sequence of the slope to be observed;

[0028] Calculate the second risk area in each ground-based SAR image in the monitoring period according to the position of the identification point of the slope to be observed in each ground-based SAR image in the ground-based SAR image sequence;

[0029] Tracking the position of the identification point of the slope to be observed in each ground-based SAR image in the ground-based SAR image sequence, and determining the deformation rate of the identification point of the slope to be observed according to the tracking result;

[0030] According to the deformation rate and the influencing factor, the second risk area in each ground-based SAR image after the monitoring period is obtained;

[0031] According to the first risk area and the second risk area, the risk areas in the corresponding ground-based SAR images at each monitoring time point in the monitoring period and subsequent periods are output.

[0032] Optionally, the step of obtaining the second risk area in each ground-based SAR image after the monitoring period according to the deformation rate and the influencing factor includes:

[0033] According to the deformation rate of each identification point, the migration position of each identification point in each future monitoring time point after the monitoring period is predicted;

[0034] Predicting the regional sequence of the second risk area changing over time based on the migration position of each identification point at each future monitoring time point after the monitoring period;

[0035] According to the region sequence of the second risk region changing with time, the second risk region in each ground-based SAR image corresponding to each future monitoring time point after the monitoring period is determined.

[0036] Optionally, the step of outputting the risk areas in the corresponding ground-based SAR images at each monitoring time point in the monitoring period and subsequent periods according to the first risk area and the second risk area includes:

[0037] Delineating risk areas of different levels in each ground-based SAR image according to the position distribution of the first risk area and the second risk area in each ground-based SAR image;

[0038] Output different levels of risk areas for each ground-based SAR image.

[0039] Optionally, the step of marking the risk areas corresponding to each monitoring time point in the ground-based SAR images corresponding to each monitoring time point to determine the key observation areas in the ground-based SAR images corresponding to each monitoring time point includes:

[0040] Establish a unified coordinate system in the ground-based SAR images corresponding to each monitoring time point;

[0041] In the coordinate system of the ground-based SAR image corresponding to each monitoring time point, the coordinate position of each identification point is marked, and the coordinate position of the risk area is marked.

[0042] Optionally, the step of marking the risk areas corresponding to each monitoring time point in the ground-based SAR images corresponding to each monitoring time point to determine the key observation areas in the ground-based SAR images corresponding to each monitoring time point further includes:

[0043] Determine the expanded observation range of the risk area based on the coordinate position of the risk area in the ground-based SAR image and the deformation rate of the identification points in the risk area;

[0044] Based on the expansion of the observation scope, key observation areas will be determined.

[0045] Optionally, the step of denoising the ground-based SAR images acquired at each monitoring time point according to the key denoising range includes:

[0046] For the key observation areas of the ground-based SAR images collected at each monitoring time point, sort and denoise the images according to the monitoring time point in the computing resources;

[0047] For the non-key observation areas of the ground-based SAR images collected at each monitoring time point, denoising is performed in sequence after the computing resources complete the denoising of the key observation areas.

[0048] Optionally, the denoising accuracy of the key observation area is greater than the denoising accuracy of the non-key observation area.

[0049] Optionally, the deformation rate calculation method and the migration position prediction method of each identification point at each future monitoring time point after the monitoring period are as follows:

[0050] Establish the migration position prediction equation of the i-th identification point:

[0051] ;

[0052] in, Represents the deformation rate of the slope position to be observed corresponding to the i-th identification point; is the corrected displacement of the i-th identification point; t is the time value corresponding to the monitoring time point; is the cumulative displacement of the i-th identification point in the ground-based SAR image collected at the monitoring time point t;

[0053] ;

[0054] ;

[0055] Where n is the number of ground-based SAR images in the ground-based SAR image sequence, is the starting time of the monitoring period, It is the monitoring time point corresponding to the last ground-based SAR image in the ground-based SAR image sequence, that is, the end time of the monitoring period.

[0056] In the technical solution of the present invention, identification elements are respectively set at multiple positions of the slope to be observed, so that each position forms an identification point; at the same time, ground-based SAR images of the slope to be observed taken at different monitoring time points within the monitoring period are obtained to form an image sequence, and a monitoring parameter sequence of the slope to be observed at different monitoring time points within the monitoring period is obtained; a neural network model for identifying the risk area of ​​the slope is used to output the risk area in the corresponding ground-based SAR image at each monitoring time point within the monitoring period and subsequent periods; a key observation area is determined according to the risk area in the corresponding ground-based SAR image at each monitoring time point; for the key observation area in the ground-based SAR image collected at each monitoring time point, focused denoising is performed, and for the area outside the key observation area in the ground-based SAR image, denoising can be delayed or denoised with lower precision. Therefore, the present invention proposes a method for dividing the focused denoising range of the ground-based SAR image, which can focus on denoising the key area and reduce the data processing amount of denoising, especially when the monitoring frequency of the slope is intensive, the processing result of the slope measurement image after denoising can be fed back more focusedly and more quickly. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 This is a flow chart of a first embodiment of a method for denoising ground-based SAR slope measurement data according to the present invention.

[0058] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings in conjunction with the embodiments. DETAILED DESCRIPTION

[0059] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0060] In the following description, suffixes such as "unit", "component" or "unit" used to represent elements are only used to facilitate the description of the present invention, and have no specific meanings. Therefore, "unit", "component" or "unit" can be used in a mixed manner.

[0061] See also Figure 1 In a first embodiment of the present invention, a method for denoising ground-based SAR slope measurement data is provided, comprising the following steps:

[0062] Step S10, obtaining a ground-based SAR image sequence and a monitoring parameter sequence of the slope to be observed within a monitoring period of a preset duration, wherein a plurality of position points of the slope to be observed are respectively provided with identification elements to form identification points;

[0063] Step S20, inputting the ground-based SAR image sequence and the monitoring parameter sequence of the slope to be observed into the neural network model for identifying the risk area of ​​the slope, and outputting the risk area in the corresponding ground-based SAR image at each monitoring time point during the monitoring period and subsequent periods, wherein the risk area is calculated based on the identification point;

[0064] Step S30, marking the risk areas corresponding to each monitoring time point in the ground-based SAR image corresponding to each monitoring time point, so as to determine the key observation areas in the ground-based SAR image corresponding to each monitoring time point;

[0065] Step S40, calibrating a key denoising range in the ground-based SAR image according to the key observation area in the ground-based SAR image corresponding to each monitoring time point;

[0066] Step S50, denoising the ground-based SAR images acquired at each monitoring time point according to the key denoising range.

[0067] In the technical solution of the present invention, identification elements are respectively set at multiple positions of the slope to be observed, so that each position forms an identification point; at the same time, ground-based SAR images of the slope to be observed taken at different monitoring time points within the monitoring period are obtained to form an image sequence, and a monitoring parameter sequence of the slope to be observed at different monitoring time points within the monitoring period is obtained; a neural network model for identifying the risk area of ​​the slope is used to output the risk area in the corresponding ground-based SAR image at each monitoring time point within the monitoring period and subsequent periods; a key observation area is determined according to the risk area in the corresponding ground-based SAR image at each monitoring time point; for the key observation area in the ground-based SAR image collected at each monitoring time point, focused denoising is performed, and for the area outside the key observation area in the ground-based SAR image, denoising can be delayed or denoised with lower precision. Therefore, the present invention proposes a method for dividing the focused denoising range of the ground-based SAR image, which can focus on denoising the key area and reduce the data processing amount of denoising, especially when the monitoring frequency of the slope is intensive, the processing result of the slope measurement image after denoising can be fed back more focusedly and more quickly.

[0068] Specifically, identification elements are respectively provided at multiple locations of the slope to be observed. The identification elements are used to enhance the sensitivity of the ground-based SAR to specific areas or objects. The identification elements may be corner reflectors, radar cross-section enhancers, signs made of highly reflective materials, and the like.

[0069] The positions of the set identification points of the slope to be observed may include: boundary areas in various directions of the slope to be observed, and set key position points of the slope to be observed (for example, position points that are easily deformed, or position points that are set to be focused on).

[0070] The monitoring parameter sequence of the slope to be observed may be set parameters that have an impact on the deformation of the slope to be observed.

[0071] The key observation area is an area that is prone to deformation and other risks determined based on the risk area. The key observation area can be subjected to more accurate denoising in the subsequent image denoising process, while the denoising processing accuracy of the non-key area can be reduced to reduce the data processing amount of the denoising process, thereby providing faster feedback on the processing results of the slope measurement image after denoising.

[0072] Among them, the positions corresponding to the marking points can be more easily identified by ground-based SAR, thereby providing reference data for dividing risk areas and non-risk areas. This can not only avoid the problem of being unable to divide the risk areas of the image before the ground-based SAR image is denoised, but also provide reference data for the data analysis process of the ground-based SAR image after denoising.

[0073] Based on the first embodiment of the ground-based SAR slope measurement data denoising method of the present invention, in the second embodiment of the ground-based SAR slope measurement data denoising method of the present invention, the step S10 includes:

[0074] Step S11, obtaining a set number of ground-based SAR images and monitoring parameters collected at set time intervals during a monitoring period of the slope to be observed, wherein the monitoring parameters include meteorological parameters, terrain parameters, vegetation parameters and human activity impact parameters;

[0075] Step S12, arranging a set number of ground-based SAR images within the same monitoring period into a ground-based SAR image sequence in time sequence;

[0076] Step S13, arranging a set number of monitoring parameters collected within the same monitoring period in time sequence to form a monitoring parameter sequence of the slope to be observed.

[0077] The purpose of acquiring a ground-based SAR image sequence is to calculate the deformation of each identification point in each ground-based SAR image over time.

[0078] Meteorological parameters of the slopes to be observed include: rainfall, snowmelt, freezing, storms and seismic activity.

[0079] The topographic parameters of the slope to be observed include: slope gradient, slope aspect, elevation, slope length, slope surface morphology, geological structure, rock and soil properties and groundwater level.

[0080] The vegetation parameters of the slope to be observed include: vegetation coverage, vegetation type, root characteristics and vegetation height.

[0081] The human activity impact parameters of the slope to be observed include: engineering construction, deforestation, agricultural activities, mining activities and transportation activities.

[0082] Among them, each monitoring time point of the slope to be observed corresponds to a set of monitoring parameters, and each set of monitoring parameters includes: meteorological parameters, terrain parameters, vegetation parameters and human activity impact parameters. The monitoring parameters of different monitoring time points are arranged according to time, and the monitoring parameter sequence of the slope to be observed is obtained.

[0083] Among them, the ground-based SAR image at each monitoring time point corresponds to a set of monitoring parameters at the same monitoring time point.

[0084] Furthermore, in the present invention, the acquisition period of the ground-based SAR images and monitoring parameters of the next monitoring period of the slope to be observed can be determined according to the monitoring parameter sequence of the slope to be observed, as well as the magnification ratio of converting the risk area of ​​the slope to be observed into the key observation area.

[0085] Specifically, the specific parameter conditions in each group of monitoring parameters are assigned according to preset conditions (when assigning values, the greater the impact on slope deformation, the greater the assigned value), and the same group of monitoring parameters obtains a weighted summation result according to the assignment to obtain the simulated value of the group of monitoring parameters. According to the value interval corresponding to the summation result of the simulated values ​​of each group of monitoring parameters in the monitoring period, the image acquisition period in the subsequent round of monitoring period of the slope to be observed can be adjusted (the larger the value interval corresponding to the summation result of the simulation value, the smaller the acquisition period and the more intensive the acquisition; conversely, the smaller the value interval corresponding to the summation result of the simulation value, the larger the acquisition period and the longer the acquisition interval), and the magnification ratio of the risk area of ​​the slope to be observed to the key observation area can be adjusted (the larger the value interval corresponding to the summation result of the simulation value, the larger the magnification ratio of the risk area to the key observation area; conversely, the smaller the value interval corresponding to the summation result of the simulation value, the smaller the magnification ratio of the risk area to the key observation area).

[0086] Based on the second embodiment of the ground-based SAR slope measurement data denoising method of the present invention, in a third embodiment of the ground-based SAR slope measurement data denoising method of the present invention, step S20 includes:

[0087] Step S21, inputting the ground-based SAR image sequence and the monitoring parameter sequence of the slope to be observed into a neural network model for identifying the risk area of ​​the slope;

[0088] Step S22, determining the first risk area of ​​the slope to be observed according to the monitoring parameter sequence of the slope to be observed in the neural network model;

[0089] Step S23, in the neural network model, determining the influencing factor of the deformation rate of the slope to be observed according to the monitoring parameter sequence of the slope to be observed;

[0090] Step S24, calculating the second risk area in each ground-based SAR image in the monitoring period according to the position of the identification point of the slope to be observed in each ground-based SAR image in the ground-based SAR image sequence;

[0091] Step S25, tracking the position of the identification point of the slope to be observed in each ground-based SAR image in the ground-based SAR image sequence, and determining the deformation rate of the identification point of the slope to be observed according to the tracking result;

[0092] Step S26, obtaining a second risk area in each ground-based SAR image after the monitoring period according to the deformation rate and the influencing factor;

[0093] Step S27, outputting the risk areas in the corresponding ground-based SAR images at each monitoring time point during the monitoring period and subsequent periods according to the first risk area and the second risk area.

[0094] Specifically, the monitoring parameters of the historically observed slope are used as input data, and the risk area of ​​the historically observed slope is used as output data, which are brought into the neural network model for training to obtain a neural network model for predicting the first risk area of ​​the slope to be observed.

[0095] Furthermore, the monitoring parameter sequence of the slope to be observed records the changes in meteorological parameters, terrain parameters, vegetation parameters and human activity impact parameters at the time of taking the base-based SAR images during the monitoring period, and these conditions are important parameters for determining the deformation acceleration of the slope to be observed. Therefore, according to the monitoring parameter sequence of the slope to be observed, the influencing factors of the deformation rate of the slope to be observed can be determined.

[0096] According to the value interval corresponding to the summation result of the simulation values ​​of each group of monitoring parameters, the influence factor of the deformation rate corresponding to the value interval is determined.

[0097] In the ground-based SAR images corresponding to each monitoring time point outside the monitoring period, the second risk area is determined by the deformation rate of each marker point calculated based on the change of the measured position of the marker point in each ground-based SAR image in the ground-based SAR image sequence collected during the monitoring period. Thus, according to the neural network prediction, the first risk area of ​​the slope to be observed can be obtained, and according to the measured position of each marker point at different monitoring time points, the deformation rate of each marker point and the second risk area in each ground-based SAR image during the monitoring period can be obtained. According to the deformation rate and the influencing factor, the corresponding second risk area of ​​the slope to be observed in each future ground-based SAR image can be predicted.

[0098] Specifically, in the next monitoring period, each ground-based SAR image acquired in the next monitoring period is processed into a ground-based SAR image sequence, and each monitoring parameter acquired in the next monitoring period is processed into a monitoring parameter sequence, and then input into the neural network model.

[0099] The ground-based SAR image sequences in adjacent monitoring periods should have overlapping sequences, and the monitoring parameter sequences in adjacent monitoring periods should also have overlapping data.

[0100] Based on the third embodiment of the ground-based SAR slope measurement data denoising method of the present invention, in a fourth embodiment of the ground-based SAR slope measurement data denoising method of the present invention, step S26 includes:

[0101] Step S261, predicting the migration position of each identification point at each future monitoring time point after the monitoring period according to the deformation rate of each identification point;

[0102] Step S262, predicting the region sequence of the second risk region changing over time according to the migration position of each identification point at each future monitoring time point after the monitoring period;

[0103] Step S263, determining the second risk area in each ground-based SAR image corresponding to each future monitoring time point after the monitoring period according to the area sequence of the second risk area changing with time.

[0104] Specifically, in step S261, the migration position of each identification point at each future monitoring time point after the monitoring period is predicted, and the migration position prediction equation of the i-th identification point in the ninth embodiment can be used for calculation.

[0105] In step S262, if the identification points in the second risk area are not migrated in the future, the area range of the second risk area will remain unchanged at the subsequent monitoring time points. If the identification points in the second risk area are migrated in the future, the area range of the second risk area at each future monitoring time point will expand according to the predicted migration positions of each identification point in the second risk area at different time points. The second risk area at each monitoring time point is the area circled by the migration position of the identification point at the corresponding moment.

[0106] Based on this principle, in fact, the ground-based SAR images collected at each different monitoring point in the present invention will cause the corresponding image priority denoising range to change due to the unchanged or expanded second risk area. Therefore, the technical solution of the present invention can adaptively adjust the image priority denoising range according to the slope deformation.

[0107] Furthermore, according to the deformation rate of each identification point, the identification point whose deformation rate exceeds a preset rate is determined as the first risk identification point.

[0108] Based on the fourth embodiment of the ground-based SAR slope measurement data denoising method of the present invention, in a fifth embodiment of the ground-based SAR slope measurement data denoising method of the present invention, step S27 includes:

[0109] Step S271, delineating risk areas of different levels in each ground-based SAR image according to the position distribution of the first risk area and the position distribution of the second risk area in each ground-based SAR image;

[0110] Step S272: output risk areas of different levels in each ground-based SAR image.

[0111] Specifically, the first risk area can be considered as an empirical risk area, and the second risk area is a risk area determined according to deformation and influencing factors.

[0112] The intersection of the first risk area and the second risk area is calculated to obtain a risk area of ​​the first risk level; and the risk area of ​​the second risk level is obtained based on the non-intersecting area of ​​the first risk area and the second risk area.

[0113] Among them, according to the deformation rate of the identification points contained in the risk area of ​​the second risk level, the risk area of ​​the second risk level can be divided into several areas with different risk levels.

[0114] Based on the fifth embodiment of the ground-based SAR slope measurement data denoising method of the present invention, in a sixth embodiment of the ground-based SAR slope measurement data denoising method of the present invention, step S30 includes:

[0115] Step S31, establishing a unified coordinate system in the ground-based SAR images corresponding to each monitoring time point;

[0116] Step S32, marking the coordinate position of each identification point in the coordinate system of the ground-based SAR image corresponding to each monitoring time point, and marking the coordinate position of the risk area.

[0117] Specifically, each identification point and risk area in each ground-based SAR image is marked by the coordinate position in the coordinate system.

[0118] Based on the sixth embodiment of the ground-based SAR slope measurement data denoising method of the present invention, in the seventh embodiment of the ground-based SAR slope measurement data denoising method of the present invention, the step S30 further includes:

[0119] Step S33, determining the expanded observation range of the risk area according to the coordinate position of the risk area in the ground-based SAR image and the deformation rate of the identification point in the risk area;

[0120] Step S34, determining the key observation area according to the expanded observation range.

[0121] Specifically, in this embodiment, the size of the expanded observation range can also be determined based on the maximum deformation rate of each identification point in the risk area. It is easy to understand that the larger the maximum deformation rate, the larger the expanded observation range, and conversely, the smaller the maximum deformation rate, the smaller the expanded observation range.

[0122] Based on the seventh embodiment of the ground-based SAR slope measurement data denoising method of the present invention, in the eighth embodiment of the ground-based SAR slope measurement data denoising method of the present invention, step S50 includes:

[0123] Step S51, for the key observation area of ​​the ground-based SAR image collected at each monitoring time point, sorting and denoising are performed in the computing resources according to the monitoring time point of the image;

[0124] Step S52, for the non-key observation areas of the ground-based SAR images collected at each monitoring time point, denoising is performed in sequence after the computing resources complete the denoising of the key observation areas.

[0125] Specifically, each ground-based SAR image of the slope to be observed may be uploaded to a cloud server, and the cloud server allocates computing resources to execute steps S10 to S50.

[0126] In order to quickly obtain the denoising result of each ground-based SAR image, the key observation area of ​​each ground-based SAR image can be denoised according to the monitoring time point of the image in the computing resources, and the denoising result of the key observation area can be obtained first.

[0127] Since the key observation area is the area where risks are most likely to occur in the ground-based SAR image, the key observation area is denoised first, so that the denoised image of the key area can be obtained first, which is convenient for subsequent image analysis and processing.

[0128] As for the denoising of the non-key observation area, the denoising is performed after the denoising of the image part in the key observation area is completed.

[0129] Specifically, the denoising accuracy of the key observation area is greater than that of the non-key observation area.

[0130] The accuracy of image denoising has a direct and important impact on the quality of the final image and the application effect. When the denoising accuracy is not enough, there will still be a lot of noise remaining in the image, which may lead to poor visual quality, difficulty in feature recognition or complex subsequent processing: Therefore, a higher denoising accuracy can be used for key observation areas. In order to save denoising processing time, a lower denoising accuracy can be used for non-key observation areas.

[0131] Based on the eighth embodiment of the ground-based SAR slope measurement data denoising method of the present invention, in the ninth embodiment of the ground-based SAR slope measurement data denoising method of the present invention, the calculation method of the deformation rate and the prediction method of the migration position of each identification point at each future monitoring time point after the monitoring period are as follows:

[0132] Establish the migration position prediction equation of the i-th identification point:

[0133] ;

[0134] in, Represents the deformation rate of the slope position to be observed corresponding to the i-th identification point; is the corrected displacement of the i-th identification point; t is the time value corresponding to the monitoring time point; is the cumulative displacement of the i-th identification point in the ground-based SAR image collected at the monitoring time point t;

[0135] ;

[0136] ;

[0137] Where n is the number of ground-based SAR images in the ground-based SAR image sequence, is the starting time of the monitoring period, It is the monitoring time point corresponding to the last ground-based SAR image in the ground-based SAR image sequence, that is, the end time of the monitoring period.

[0138] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a computer-readable storage medium (such as ROM / RAM, magnetic disk, optical disk) as described above, and includes a number of instructions for enabling a terminal device to enter the method described in each embodiment of the present invention.

[0139] In the description of this specification, the description with reference to the terms "an embodiment", "another embodiment", "other embodiments", or "first embodiment to Xth embodiment" etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, method steps or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0140] It should be noted that, in this article, the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "comprises a ..." does not exclude the existence of other identical elements in the process, method, article or system including the element.

[0141] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0142] The above are only preferred embodiments of the present invention, and are not intended to limit the patent scope of the present invention. Any equivalent structure or equivalent process transformation made using the contents of the present invention specification and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.

Claims

1. A method for denoising ground-based SAR slope measurement data, characterized in that: The steps include: Acquiring a ground-based SAR image sequence and a monitoring parameter sequence of a slope to be observed within a monitoring period of a preset duration, comprising: acquiring a set number of ground-based SAR images and monitoring parameters collected at set time intervals within a monitoring period of the slope to be observed; arranging a set number of ground-based SAR images within the same monitoring period into a ground-based SAR image sequence in time sequence; arranging a set number of collected monitoring parameters within the same monitoring period in time sequence to form a monitoring parameter sequence of the slope to be observed; wherein identification elements are respectively provided at a plurality of position points of the slope to be observed to form identification points, and the monitoring parameters include meteorological parameters, terrain parameters, vegetation parameters and parameters of human activity impact; Inputting a ground-based SAR image sequence and a monitoring parameter sequence of a slope to be observed into a neural network model for identifying a risk area of ​​the slope, and outputting risk areas in corresponding ground-based SAR images at each monitoring time point during a monitoring period and a subsequent period, including: inputting a ground-based SAR image sequence and a monitoring parameter sequence of a slope to be observed into a neural network model for identifying a risk area of ​​the slope; determining a first risk area of ​​the slope to be observed in the neural network model according to the monitoring parameter sequence of the slope to be observed; In the neural network model, the influence factor of the deformation rate of the slope to be observed is determined according to the monitoring parameter sequence of the slope to be observed; the second risk area in each ground-based SAR image in the ground-based SAR image sequence is calculated according to the position of the identification point of the slope to be observed in each ground-based SAR image in the ground-based SAR image sequence; the position of the identification point of the slope to be observed in each ground-based SAR image in the ground-based SAR image sequence is tracked, and the deformation rate of the identification point of the slope to be observed is determined according to the tracking result; according to the deformation rate and the influence factor, the second risk area in each ground-based SAR image after the monitoring period is obtained; according to the first risk area and the second risk area, the risk area in the corresponding ground-based SAR image at each monitoring time point in the monitoring period and subsequent periods is output; wherein the risk area is calculated based on the identification point; The risk areas corresponding to each monitoring time point are marked in the ground-based SAR images corresponding to each monitoring time point, so as to determine the key observation areas in the ground-based SAR images corresponding to each monitoring time point; According to the key observation areas in the ground-based SAR images corresponding to each monitoring time point, the key denoising range is calibrated in the ground-based SAR images; The ground-based SAR images acquired at each monitoring time point are denoised according to the key denoising range.

2. The method for denoising ground-based SAR slope measurement data according to claim 1, characterized in that: The step of obtaining the second risk area in each ground-based SAR image after the monitoring period according to the deformation rate and the influencing factor comprises: According to the deformation rate of each identification point, the migration position of each identification point in each future monitoring time point after the monitoring period is predicted; Predicting the regional sequence of the second risk area changing over time based on the migration position of each identification point at each future monitoring time point after the monitoring period; According to the region sequence of the second risk region changing with time, the second risk region in each ground-based SAR image corresponding to each future monitoring time point after the monitoring period is determined.

3. The method for denoising ground-based SAR slope measurement data according to claim 2, characterized in that: The step of outputting the risk areas in the corresponding ground-based SAR images at each monitoring time point in the monitoring period and subsequent periods according to the first risk area and the second risk area comprises: Delineating risk areas of different levels in each ground-based SAR image according to the position distribution of the first risk area and the second risk area in each ground-based SAR image; Output different levels of risk areas for each ground-based SAR image.

4. The method for denoising ground-based SAR slope measurement data according to claim 3, characterized in that: The step of marking the risk areas corresponding to each monitoring time point in the ground-based SAR images corresponding to each monitoring time point to determine the key observation areas in the ground-based SAR images corresponding to each monitoring time point includes: Establish a unified coordinate system in the ground-based SAR images corresponding to each monitoring time point; In the coordinate system of the ground-based SAR image corresponding to each monitoring time point, the coordinate position of each identification point is marked, and the coordinate position of the risk area is marked.

5. The method for denoising ground-based SAR slope measurement data according to claim 4, characterized in that: The step of marking the risk areas corresponding to each monitoring time point in the ground-based SAR images corresponding to each monitoring time point to determine the key observation areas in the ground-based SAR images corresponding to each monitoring time point also includes: Determine the expanded observation range of the risk area based on the coordinate position of the risk area in the ground-based SAR image and the deformation rate of the identification points in the risk area; Based on the expansion of the observation scope, key observation areas will be determined.

6. The method for denoising ground-based SAR slope measurement data according to claim 5, characterized in that: The step of denoising the ground-based SAR images acquired at each monitoring time point according to the key denoising range includes: For the key observation areas of the ground-based SAR images collected at each monitoring time point, sort and denoise the images according to the monitoring time point in the computing resources; For the non-key observation areas of the ground-based SAR images collected at each monitoring time point, denoising is performed in sequence after the computing resources complete the denoising of the key observation areas.

7. The method for denoising ground-based SAR slope measurement data according to claim 6, characterized in that: The denoising accuracy of the key observation area is greater than that of the non-key observation area.

8. The method for denoising ground-based SAR slope measurement data according to any one of claims 1 to 7, characterized in that: The calculation method of the deformation rate and the prediction method of the migration position of each identification point at each future monitoring time point after the monitoring period are as follows: Establish the migration position prediction equation of the i-th identification point: ； in, Represents the deformation rate of the slope position to be observed corresponding to the i-th identification point; is the corrected displacement of the i-th identification point; t is the time value corresponding to the monitoring time point; is the cumulative displacement of the i-th identification point in the ground-based SAR image collected at the monitoring time point t; ； ； Where n is the number of ground-based SAR images in the ground-based SAR image sequence, is the starting time of the monitoring period, It is the monitoring time point corresponding to the last ground-based SAR image in the ground-based SAR image sequence, that is, the end time of the monitoring period.

Images