CN117310706A - Discontinuous deformation monitoring method and system for foundation radar - Google Patents

Abstract

The invention provides a discontinuous deformation monitoring method and a discontinuous deformation monitoring system for a foundation radar, which provide relocation error compensation for a two-dimensional azimuth nonlinear model according to scene assumption and model error analysis, and can compensate phase errors caused by radar position deviation. The scheme is suitable for long-term slow deformation monitoring of the foundation radar, can improve the flexibility of radar layout, reduces the complexity of data processing, has higher error compensation precision, improves deformation measurement precision, and is applied to discontinuous monitoring of buildings, and the deformation measurement precision after compensation is superior to millimeter level.

Description

A ground-based radar intermittent deformation monitoring method and system

Technical field

The invention relates to the fields of computer data processing and radar data processing, and is applied to radar positioning error compensation, and in particular, to a ground-based radar discontinuous deformation monitoring method and system.

Background technique

Ground-based SAR has been successfully used for high-precision deformation monitoring in areas affected by natural disasters, dams, mountain glaciers, docks and bridge piers, etc., with its short spatio-temporal baseline, high resolution, short measurement period, and convenient operation, breaking through the space-based SAR Limitations include temporal decoherence, long spatial baselines, and long revisit periods. GBSAR achieves high resolution in range by emitting high-frequency, large-bandwidth electromagnetic waves in the range direction, and uses motion in a fixed orbit to synthesize a virtual large-aperture antenna to achieve high resolution in azimuth.

According to the way of obtaining data, GBSAR can be divided into two monitoring modes: continuous monitoring (C-GBSAR) and discontinuous monitoring (D-GBSAR). However, in the continuous monitoring model, long-term monitoring of GBSAR has data redundancy and data processing complexity; for slow deformation monitoring scenarios, a large amount of manpower and material resources are consumed, and the monitoring efficiency is low. In the intermittent monitoring mode, the monitoring radar has a spatial position movement, which is similar to the spatial baseline of spaceborne SAR.

Currently widely used in intermittent monitoring is GB-SAR intermittent deformation monitoring, which generally includes: image registration, interference pattern generation, permanent scatterers (PS) selection, phase unwrapping, repositioning error compensation, and deformation solution Calculate. The discontinuous measurement mode of ground-based synthetic aperture radar is suitable for monitoring slowly deforming landslides. However, the radar needs to be repeatedly installed and disassembled, which will inevitably cause repositioning errors and seriously affect the accuracy of deformation measurement. During the multi-scene monitoring process of radar, the spatial position will shift, resulting in errors other than deformation in the interference phase. Since the wavelength of the radar is short and the measurement accuracy reaches the sub-millimeter level, millimeter-level deviations in the radar position will produce serious phase errors for the monitored targets. Therefore, the relocation error is very sensitive to the position offset and will seriously reduce the credibility of the deformation inversion results if not accurately corrected.

In spaceborne InSAR, the most widely used method to correct the baseline error is to use a first-order polynomial function or a second-order model to estimate the best-fitting plane. However, the first-order model still has residual phases after error correction, and these errors are compensated The solution is directly based on data analysis and error plane fitting methods to correct the impact of errors, without starting from the error generation mechanism and fundamentally performing error compensation.

Contents of the invention

In view of this, embodiments of the present invention provide a new repositioning error compensation scheme, which can compensate for the phase error introduced due to radar position deviation. This scheme is suitable for long-term slow deformation monitoring of ground-based radars and can improve the flexibility of radar layout. characteristics, reduce the complexity of data processing, and improve the accuracy of deformation measurement.

Specifically, the present invention provides the following technical solutions:

On the one hand, the present invention provides a ground-based radar intermittent deformation monitoring method, which method includes:

S1. Based on the interferograms obtained at different time points during intermittent deformation monitoring, calculate the deformation amount of the monitoring target caused by the deviation of the radar spatial baseline along the x-axis, y-axis, and z-axis;

S2. Based on the deformation of the monitoring target, as well as the relationship between the radar echo phase change and the change in the distance between the monitoring target and the radar, establish a repositioning error model in the x-axis, y-axis, and z-axis directions;

S3. Based on the repositioning error model and combined with the target pitch angle span, calculate the interference phase model of the PS point in the monitoring scene; the interference phase model is:

in, ,/> and/> is the parameter to be estimated,/> is the azimuth angle,/> ,/> ,/> Respectively represent the phase changes caused by the radar's offset along the x-axis, y-axis, and z-axis;

S4. Estimate parameters in the interference phase model to obtain a repositioning error surface.

Preferably, in S1, the deformation amount of the monitoring target caused by the offset along the x-axis direction is:

in, Indicates the initial position of the radar,/> Indicates the position after the radar position is offset,/> is the target location,/> Represents the three-dimensional slant distance between the target and the radar,/> Represents the offset of the radar along the x-axis, /> Represents the pitch angle of PS point.

Preferably, in S1, the monitoring target deformation caused by the offset along the y-axis direction is:

in, Indicates the initial position of the radar,/> Indicates the position after the radar position is offset,/> is the target location,/> Represents the three-dimensional slant distance between the target and the radar,/> Represents the offset of the radar along the y-axis, /> Represents the pitch angle of PS point.

Preferably, in S1, the monitoring target deformation caused by the offset along the z-axis direction is:

in, Indicates the initial position of the radar,/> Indicates the position after the radar position is offset,/> is the target location,/> Represents the three-dimensional slant distance between the target and the radar,/> Represents the offset of the radar along the z-axis, /> Represents the pitch angle of PS point.

Preferably, in S2, the relocation error model is:

in, Represents the offset of the radar along the x-axis, /> Represents the offset of the radar along the y-axis, /> Represents the offset along the z-axis, /> Represents the pitch angle of PS point,/> Represents the radar signal wavelength.

Preferably, in S3, for the interference phase after unwrapping of PS points in the monitoring scene When in matrix form, that is , then the matrix form of the interference phase model is converted to:

in, Represents the error of the model,/> Represents the interference phase in matrix form, the azimuth angle of the PS point , X represents a matrix composed of sine and cosine function values of the azimuth angle, /> Represents the parameters to be estimated, they are expressed as follows:

Preferably, in the establishment of the interference phase model, the target pitch angle Take its value/> Instead, that is:

in, , the target pitch angle span is/> ,/> Represents the offset of the radar along the x-axis, /> Represents the offset of the radar along the y-axis, /> Represents the offset of the radar along the z-axis, /> Represents the radar signal wavelength.

Preferably, in S3, for the screening of PS points, dual thresholds of coherence coefficient and amplitude information are used for screening;

In the phase unwrapping stage after PS point screening, phase unwrapping is performed along the azimuth direction.

Preferably, the estimated relocation error model parameters are solved using the least squares method:

in, represents the estimated relocation error model parameters;

Then, the estimated relocation error phase is derived:

Finally, using the unwrapped PS point phase Subtract the estimated relocation error phase/> , you can solve the deformation variable and obtain the deformation phase/> .

On the other hand, the present invention also provides a ground-based radar intermittent deformation monitoring system, which system includes:

The monitoring target deformation amount calculation module is used to calculate the monitoring target deformation amount caused by the deviation of the radar space baseline along the x-axis, y-axis, and z-axis based on the interferograms obtained at different time points during intermittent deformation monitoring;

The repositioning error model module is used to establish a repositioning error model in the x-axis, y-axis, and z-axis directions based on the deformation of the monitoring target, as well as the relationship between the phase change of the radar echo and the change in the distance between the monitoring target and the radar;

The interference phase model module is used to calculate the interference phase model of the PS point in the monitoring scene based on the repositioning error model and the target pitch angle span; the interference phase model is:

in, ,/> and/> is the parameter to be estimated,/> is the azimuth angle,/> ,/> ,/> Represent respectively the phase changes caused by the radar deflection along the x-axis, y-axis, and z-axis;

A parameter estimation module is used to estimate parameters in the interference phase model to obtain a repositioning error surface.

Preferably, the system further includes:

The image registration module is used to perform image registration based on the amplitude and phase coherence of different SLC images;

Interference pattern generation module, used to generate interference patterns based on registered images;

PS point screening module uses dual thresholds of coherence coefficient and amplitude information to screen PS points;

The phase unwrapping module performs phase unwrapping on the interferogram after PS point screening, and the image data after phase unwrapping is input into the monitoring target deformation amount calculation module;

The deformation solution module calculates the deformation amount based on the repositioning error surface obtained by the parameter estimation module.

In a third aspect, the present invention also provides a ground-based radar intermittent deformation monitoring equipment. The equipment includes a processor and a memory. The processor can call computer instructions stored in the memory to execute the ground-based radar intermittent deformation as described above. Monitoring methods.

Compared with the existing technology, the technical solution of the present invention starts from the geometric relationship of the radar three-dimensional baseline, derives the geometric relationship between the repositioning error and system parameters such as the radar baseline, target pitch angle and azimuth angle, and fundamentally reveals the repositioning error phase mathematical principles and physical meanings. Then, based on the scenario assumptions and model error analysis, it was concluded that the pitch angle of the target has little impact on the repositioning error, and a two-dimensional azimuth nonlinear model repositioning error compensation method was proposed. This method does not require additional Data Elevation Model (DEM) data to compensate for relocation errors with high accuracy. Compared with the traditional polynomial model, the method proposed in this scheme has nonlinear error correction capabilities and has higher error compensation accuracy. This solution was applied to intermittent monitoring of buildings, and the deformation measurement accuracy after compensation was better than millimeter level, which verified the practicability and superiority of this method.

Description of the drawings

In order to explain the embodiments of the present invention or the technical solutions in the prior art more clearly, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings in the following description are only These are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without exerting creative efforts.

Figure 1 is a schematic diagram of the three-dimensional coordinate relationship between the target and the radar according to the embodiment of the present invention;

Figure 2 is a schematic diagram of the x-axis component of the intermittent monitoring radar spatial baseline according to the embodiment of the present invention;

Figure 3 is a schematic diagram of the y-axis component of the intermittent monitoring radar spatial baseline according to the embodiment of the present invention;

Figure 4 is a schematic diagram of the z-axis component of the intermittent monitoring radar spatial baseline according to the embodiment of the present invention;

Figure 5 is a comparison chart of two terrain repositioning error compensation results according to the embodiment of the present invention; where: (a) (c) (e) are flat ground compensation results; (b) (d) (f) are linear slope compensation results. ; (a) is the relocation error of the flat terrain model; (b) is the relocation error of the flat terrain model; (c) is the compensation result of the proposed nonlinear model; (d) is the compensation result of the proposed nonlinear model; (e) is the two The compensation result of the first-order polynomial model; (f) is the compensation result of the second-order polynomial model;

Figure 6 is a schematic diagram of the experimental scene and results of the embodiment of the present invention; wherein: (a) is a real view of a teaching building on a certain campus; (b) is a PS point interference pattern before compensation; (c) is a PS point interference pattern after compensation;

Figure 7 is a method flow chart according to an embodiment of the present invention;

Figure 8 is a schematic diagram of the general steps of ground-based radar interferometry;

Figure 9 is a schematic diagram of the repositioning error compensation processing flow.

Detailed ways

The embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be clear that the described embodiments are only some, but not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of the present invention.

Those skilled in the art should know that the following specific examples or specific implementation modes are a series of optimized arrangements enumerated by the present invention to further explain the specific content of the invention, and these arrangements can be combined with each other or used in association with each other, unless the present invention explicitly states that some or a specific embodiment or implementation cannot be associated with or used in conjunction with other embodiments or implementations. At the same time, the following specific examples or implementations are only used as optimized arrangements and are not to be understood as limiting the scope of the present invention.

In order to make up for the deficiencies in the existing technology, this solution starts from the spatial baseline geometric model and proposes a two-dimensional azimuth nonlinear model repositioning error compensation method based on scenario assumptions and model error analysis. Compared with the traditional polynomial model, this method has higher error compensation accuracy. Applying this solution to intermittent monitoring of buildings, the deformation measurement accuracy after compensation is better than millimeter level. Below, this solution will be described in detail with specific embodiments in conjunction with Figure 7 and Figures 1-4.

1. Radar spatial baseline geometry analysis

As shown in Figure 1, this embodiment considers the three-dimensional coordinate relationship between the target and the radar, where Represents the initial position of the radar, is the target location,/> Indicates the offset position of the radar,/> Indicates the pitch angle of the target;/> Represents the azimuth angle,/> Represents the three-dimensional slant distance between the target and the radar,/> Express/> The slant distance projected onto the horizontal plane. For each target of the monitoring scene, we consider the spatial baseline in three directions of the radar.

(1) Spatial baseline model along the x-axis direction

For the case where the radar offset direction is along the x-axis, it is shown in Figure 2 (top view of Figure 1). Assume that the radar is deflected along the x-axis in the positive direction, that is, the x-axis azimuth angle . Since the target distance is on the order of kilometers, the radar offset can be controlled on the order of millimeters. So for the same goal/> , the angle caused by the radar position offset/> The change is relatively small, and it can be considered that the azimuth angle of the target remains unchanged/> . /> Represents the pitch angle of PS point.

First, in order to obtain the deformation phase introduced by the radar position offset , need to solve the target deformation variable/> . According to the cosine theorem, the target deformation variable expression is:

(1)

according to The Taylor expansion formula of is approximated by:

(2)

to the formula Expand to get:

(3)

Therefore, in this embodiment, the deformation expression caused by the spatial baseline is:

(4)

From equation (4), it can be seen that the deformation amount introduced by the radar deflection along the x-axis direction is related to the azimuth angle, elevation angle of the target and the slant range of the target. We can establish the functional relationship between the radar spatial baseline and the target phase through a parametric model.

(2) Spatial baseline model along the y-axis direction

For the case where the radar offset direction is along the y-axis, as shown in Figure 3, the same analysis can be obtained based on the geometric relationship:

(5)

Formulas (4) and (5) have the same structure, but the difference is that the radar spatial baseline is projected along different directions, with only the difference in azimuth angle parameters.

(3) Spatial baseline model along the z-axis direction

For the radar spatial baseline along the z-axis direction, as shown in Figure 4, it is obtained through the spatial geometric relationship:

(6)

According to formula (6), it is relatively simple to solve the repositioning error in the vertical direction, and only requires the pitch angle of the target.

So far, we have established a spatial baseline error model in three directions based on the physical geometric mechanism of GBSAR intermittent monitoring. Combined (4) (5) (6), and the proportional relationship between the radar echo phase changes caused by changes in the distance between the object surface and the radar, that is:

in, Represents the radar signal wavelength,/> Represents the phase of the first single-view complex image,/> Represents the phase of the second single-view complex image. After they interfere, they get/> ,/> That is, the PS point phase after unwrapping. /> Represents the first observed distance between the object surface and the radar, that is, the distance before deformation, /> Represents the second observation distance between the object surface and the radar, that is, the distance after deformation. Therefore, within a phase period, the deformation amount corresponds to the value of the interference phase one-to-one. Based on this, the relationship between phase and deformation can be obtained.

Models of repositioning errors and radar spatial baselines in three directions can be obtained:

(7)

in, ,/> ,/> Represent respectively the phase changes caused by the radar's deflection along the x-axis, y-axis, and z-axis. For formula (7) containing/> An item of general radar position offset/> From sub-millimeter level to centimeter level, such as/> , and the distance between the radar and the observation target is kilometers, so in this embodiment, we make the following assumptions:

(8)

The interferogram obtained by GBSAR through intermittent monitoring at two time points contains the radar spatial baseline error, that is, , their modeling can be simplified to the following formula:

(9)

2. Repositioning error compensation

Based on model (9), the relationship between the intermittent monitoring radar repositioning error and the target parameters can be established, and the deformation variable can be solved with high precision. However, generally GBSAR imaging only has two-dimensional information, that is, the measured azimuth angle of the target ( ), slope distance (/> ) and interference phase (/> ), lack of accurate target pitch angle (/> ) or elevation information (/> ). Therefore, directly applying model (9) for error correction is insufficient. Three-dimensional imaging or combination with DEM data is required.

Considering that most of the actual monitoring scenes are far-field monitoring, which is characterized by a long distance to the target, a large azimuth angle, and a relatively small change in the pitch angle of the target. Therefore, the main factor affecting the repositioning error is the azimuth angle. It is assumed here that the target pitch angle , the target pitch angle span is/> . When the maximum value of the difference between the pitch angles of the target is small, it can be assumed that/> , that is, take/> Median/> to replace. So we get the simplified function expression:

(10)

in ,/> and/> are the parameters of the model to be estimated.

For the interference phase of the unwrapped PS points in the monitoring scene , azimuth angle and pitch angle/> . We obtain the matrix form expression of (10):

(11)

in, Represents the error of the model, X represents the matrix composed of sine and cosine functions of the azimuth angle of the PS point, /> Indicates the parameters to be estimated, and they are expressed as follows:

Finally, the relocation error surface can be estimated by applying optimization algorithms such as least squares method and iteration method. Here, we use the least squares method as the preferred solution. We first based on Compute the estimated relocation error model parameters:

(12)

The estimated relocation error phase is then derived:

(13)

Finally, using the unwrapped PS point phase Subtract the estimated relocation error phase/> , you can solve the deformation variable and obtain the deformation phase/> .

Therefore, for far-field monitoring and the target pitch angle span is small, the azimuth angle model can be used to compensate for the repositioning error, without requiring additional information such as the target's elevation, reducing the complexity of data processing and meeting the accuracy requirements of deformation monitoring. .

The repositioning error compensation processing flow is summarized below in conjunction with the above-mentioned embodiments, as shown in Figure 9. We first perform two-dimensional phase unwrapping based on the interference phase of the PS point. Based on the phase unwrapping results, we establish a trigonometric function nonlinear model. This nonlinear model can be obtained based on the existing conventional unwrapping algorithm and phase relationship. Again. After the nonlinear model is established, we can use methods such as the least squares method to estimate the parameters of the model to determine the model parameters. Then, the repositioning error compensation method of the present invention is used to subtract the calculated repositioning error phase, and then subsequent deformation calculation is performed, thereby realizing deformation monitoring. Subsequently, further description will be given with reference to another embodiment.

In addition, in another embodiment, the solution of the present invention can also be implemented through a ground-based radar intermittent deformation monitoring system. The core of the system lies in repositioning error compensation. The system specifically includes:

The monitoring target deformation amount calculation module is used to calculate the monitoring target deformation amount caused by the deviation of the radar space baseline along the x-axis, y-axis, and z-axis based on the interferograms obtained at different time points during intermittent deformation monitoring;

The repositioning error model module is used to establish a repositioning error model in the x-axis, y-axis, and z-axis directions based on the deformation of the monitoring target, as well as the relationship between the phase change of the radar echo and the change in the distance between the monitoring target and the radar;

The interference phase model module is used to calculate the interference phase model of the PS point in the monitoring scene based on the repositioning error model and the target pitch angle span; the interference phase model is:

in, ,/> and/> is the parameter to be estimated,/> is the azimuth angle,/> ,/> ,/> Represent respectively the phase changes caused by the radar deflection along the x-axis, y-axis, and z-axis;

A parameter estimation module is used to estimate parameters in the interference phase model to obtain a repositioning error surface.

Preferably, the system further includes:

The image registration module is used to perform image registration based on the amplitude and phase coherence of different SLC images;

Interference pattern generation module, used to generate interference patterns based on registered images;

PS point screening module uses dual thresholds of coherence coefficient and amplitude information to screen PS points;

The phase unwrapping module performs phase unwrapping on the interferogram after PS point screening, and the image data after phase unwrapping is input into the monitoring target deformation amount calculation module;

The deformation solution module calculates the deformation amount based on the repositioning error surface obtained by the parameter estimation module.

During system execution, as shown in Figures 8 and 9, the ground-based radar interferometry process generally includes the following key processing links:

(1) Image registration

During intermittent monitoring, due to the offset of the radar, the generated SLC image will have pixel offsets. If interference processing is performed directly, mismatch noise errors will be introduced. Therefore, image registration is required. Starting from the discontinuous monitoring mechanism, the radar offset is generally in the millimeter to decimeter level, while the target distance is in the kilometer level. Therefore, the pixel coordinate shift corresponding to the SLC image obtained after imaging is small, generally a shift of several pixels. Therefore, directly performing rough registration on the two SLC images can meet the requirements. According to the amplitude and phase coherence of the two SLC images, the image registration can be completed by applying the correlation coefficient method to meet the coherence requirements.

(2) Interference pattern generation

The image obtained by GB-SAR monitoring is a Single-Look-Complex (SLC) image, whose values are distributed in . For timing/> A radar SLC image is used to perform differential interference processing on each target point at different times:

in Is the first/> The interference phase of a point,/> For the first/> Zhang radar echo image,/> For the first/> Zhang radar echo image,/> The symbol indicates taking conjugation,/> Indicates taking the phase.

Generally, the obtained interference phase cannot be directly used to calculate the deformation using the formula of the proportional relationship between the radar echo phase changes caused by changes in the distance between the object surface and the radar. In practice, the interference phase contains various noise interference sources, and the deformation solution must be solved after noise removal. Calculate. The differential interference pattern The interference phase of a target point can be modeled as:

in is the deformation amount of the target point’s sight direction (Line-Of-Sight);/> It is the atmospheric delay phase introduced by changes in atmospheric refractive index at different times;/> is the incoherent noise phase, which can be filtered out by a low-pass phase filter; is the blur phase,/> It is an integer, indicating the degree of ambiguity. In this scheme, we assume that the atmospheric phase in the scene can be ignored and only focus on repositioning error compensation.

(3) PS point selection

The quality of the interferogram directly affects the accuracy of atmospheric phase estimation and deformation solution. Since the interferogram contains unstable interference points such as water and vegetation, different targets are affected differently by atmospheric disturbance and incoherent noise, making it difficult to directly estimate the atmospheric phase. Permanent scatterer interferometry technology can effectively extract stable reference points of the measurement scene and improve the accuracy of deformation inversion by overcoming the influence of non-ideal factors such as time decoherence.

Commonly used PS selection methods include: coherence coefficient threshold method, amplitude dispersion threshold method, amplitude information method, phase variance method, phase noise threshold method, etc. We can choose, for example, the coherence coefficient threshold method and the amplitude threshold method for joint PS point screening:

1) Coherence coefficient threshold method: Theoretically, a high coherence coefficient can reflect a high signal-to-noise ratio of the interference pattern. The selection principle is to estimate the coherence coefficient based on the adjacent pixel values around the target pixel. The expression is:

in Represents the coherence coefficient of the pixel,/> and/> The master-slave single-view complex images respectively constitute the kth interference pair,/> Represents the conjugate of a complex number, m and n are the sliding window sizes, which can generally be set to/> as needed. wait. in/> In the case of an interference image, each pixel corresponds to/> coherence coefficient/> . Therefore, the average of the coherence coefficients of the pixels can be taken, that is

To measure the coherence coefficient of the target point, set the threshold , extract satisfies/> target point, get PS points.

2) Amplitude information method: The greater the amplitude of the echo at a scattering point, the stronger the backscattering of the radar at that point, and the greater the possibility that the target point is a hard and stable scatterer. Therefore, based on the amplitude of the echo information, strong scattering points can be initially screened, providing an important reference for the selection of PS points. By setting the amplitude threshold , you can filter out items greater than the threshold/> The scattering points are used as candidate points of PS points. According to the experimental results, the larger the amplitude threshold, the fewer points with large phase errors are selected; as the amplitude value decreases, the number of points with large phase errors gradually increases.

(4) Phase unwrapping

The radar SLC image is subjected to differential interference processing to obtain an interference pattern, and the interference phase may be entangled in time and space. Phase unwrapping is an essential step to restore the true phase of the PS point and perform high-precision deformation calculations. The phase unwrapping process includes two-dimensional space unwrapping and one-dimensional time unwrapping. In the spatial domain, two-dimensional unwrapping is generally performed based on road-jet integral, minimum norm or network flow optimization methods. In the time domain, Kalman filtering or Euler expansion can be used. Time unwrapping can correctly unwrap the phase of discontinuous regions, but it requires more time-series interferograms, and the phase must satisfy the established model; spatial unwrapping can use fewer interferograms for unwrapping (such as single-frame interference Figure), and the processing after phase unwrapping is more flexible.

How to restore the phase difference in the real deformation information and calculate the ambiguity value has always been a difficulty in the field of DInSAR. Since its development, there are currently various phase unwrapping methods: path tracking method, minimum norm method, optimization estimation method, etc. However, many methods are based on the phase continuity assumption, that is, the interference phase gradient of adjacent points is less than half a wavelength.

For phase unwrapping in the field of GB-SAR deformation monitoring, (1) in the spatial dimension: due to the discrete and irregular distribution of PS points, the unwrapping method is usually based on irregular grids, such as the minimum cost flow algorithm based on triangulation etc.; (2) In the time dimension: Since the ground-based SAR monitoring time baseline is short, it can be assumed that the phase fluctuation is small, so the time dimension can be unwrapped based on the phase continuity assumption and integration along the time series.

(5) Repositioning error compensation

After the phase is unwrapped, the repositioning error compensation method given in the above embodiment of the present invention can be used to perform error compensation. No further details will be given here.

(6) Deformation calculation

After reasonable error compensation is carried out, the repositioning error surface is obtained, and subsequent specific deformation calculations can be performed to obtain specific data.

In the following, this solution will be further elaborated based on specific comparative test examples and actual scenario cases.

In this embodiment, this solution is further explained in conjunction with a simulation comparison test. In actual intermittent monitoring, radar generally only moves within a horizontal two-dimensional plane. In this embodiment, we only consider the repositioning error within the radar horizontal plane and do not consider .

Considering that most scenarios can be approximated as flat ground or linear slopes, two experiments with different monitoring scenarios are simulated and designed here: flat ground and linear slopes to verify the superiority of the proposed method. The parameters of the scenario are shown in Table 1. In this embodiment, the radar offset direction and size are set to mm, mainly considers the radar baseline in the horizontal direction and does not consider the radar height baseline. In order to evaluate the accuracy of the model, the Root Mean Square Error (RMSE) and Maximum Absolute Error (MAE) indicators are introduced. The experimental results are shown in Figure 5 below.

Table 1 Scenario simulation parameters

Table 2 Comparison of compensation results root mean square error (RMSE) and maximum absolute error (MAE)

It can be seen from Figure 5 and Table 2 that the repositioning error compensation result of the solution proposed in this application is better than that of the second-order polynomial model, and the root mean square error of its residual phase is smaller. However, for the two-dimensional model, when the pitch angle span of the monitoring scene increases from 0 to 10 degrees, its root mean square error increases. Therefore, the proposed method will be affected by the pitch angle. When the pitch angle span is large, the repositioning error compensation accuracy of the proposed method decreases. It should be pointed out here that the advantage of the proposed model is that it does not require additional DEM data. This simplifies the complexity of D-GBSAR data processing while meeting sub-millimeter deformation monitoring needs.

In this scheme, for the case where the target pitch angle span is small, we draw the conclusion that the repositioning error phase is mainly affected by the azimuth angle. We can design a radar spatial baseline that meets the conditions based on the pitch angle parameters and deformation measurement accuracy of the monitoring scene without requiring additional DEM data, improving the efficiency of intermittent monitoring data processing.

Below, we verify the superiority of the two-dimensional azimuth model through measured data of actual scene cases. The experimental scene selects a teaching building in the west area of a certain campus. The monitoring distance of the building is 100 meters to 300 meters. The atmospheric delay phase is small, making it easy to analyze the repositioning error compensation effect. The arc SAR wavelength is 12.4 mm, the range resolution is 0.3 m, and the azimuth resolution is 0.114°. The azimuth angle of the target in the monitoring scene is distributed in the range of [50°, 130°], and the pitch angle range of the target is approximately [0°, 11°]. The pitch angle span of the radar target in the monitoring scene is small, and the repositioning error is mainly affected by the azimuth angle. In this experimental example, the method proposed in this application is used for relocation error correction, and no additional DEM data is required.

This solution uses the dual threshold PS point screening method of coherence coefficient and amplitude information to select PS points of the image; the threshold screening of the coherence coefficient method and the amplitude information method are both existing technologies, and existing solutions in the existing technology can be used. , which will not be described again here. Before this, the images need to be registered. In order to emphasize the necessity of registration, the number of PS points extracted by the same method before and after registration is used as an indicator. Before registration, 20776 PS points were selected under the same threshold. After registration, 29852 PS points were selected under the same threshold. Since there is mismatch noise in the image, the coherence of the target is reduced before registration, so the number of PS points selected based on the coherence coefficient method is small. After registration, the number of PS points in the selected area increases and there are more sample points, which improves the accuracy of subsequent error parameter estimation.

Figure 6 is one of the intermittent monitoring results. Among them, (a) in Figure 6 is a real view of a teaching building in a certain campus; (b) in Figure 6 is the PS point interference pattern before compensation; (c) in Figure 6 is the PS point interference pattern after compensation. Before compensation, the repositioning error exists in the form of interference fringes, making it difficult to accurately calculate the deformation amount. In (b) of Figure 6, we find that the interference phase distribution is regularly wrapped along the azimuth direction, so we perform phase unwrapping along the azimuth direction. After two-bit unwrapping of the PS point interference phase, based on the proposed nonlinear model compensation, the root mean square error of the residual phase is approximately 0.197 rad (0.194 mm). After compensation, the deformation of the building target is 0, which is in line with the actual scenario and verifies the feasibility and practicability of the proposed model.

In yet another embodiment, this solution can be implemented in the form of equipment, and the equipment can include corresponding modules that perform each or several steps in each of the above embodiments. Therefore, each step or several steps of each of the above-described embodiments may be performed by a corresponding module, and the electronic device may include one or more of these modules. A module may be one or more hardware modules specifically configured to perform corresponding steps, or implemented by a processor configured to perform corresponding steps, or stored in a computer-readable medium for implementation by a processor, or by Some combination to achieve this. The device can be implemented using a bus architecture.

Any process or method descriptions in flowcharts or otherwise described herein may be understood to represent modules, segments, or portions of code that include one or more executable instructions for implementing the specified logical functions or steps of the process. , and the scope of the preferred embodiments of the present invention includes additional implementations in which functions may be performed out of the order shown or discussed, including in a substantially simultaneous manner or in the reverse order, depending on the functionality involved, which shall It should be understood by those skilled in the technical field to which the implementation of this solution belongs. The processor performs the various methods and processes described above. For example, the method implementation in this solution can be implemented as a software program, which is tangibly included in a machine-readable medium, such as a memory. In some implementations, part or all of a software program may be loaded and/or installed via memory and/or communication interfaces. When a software program is loaded into memory and executed by a processor, one or more steps in the method described above may be performed. Alternatively, in other implementations, the processor may be configured to perform one of the methods described above in any other suitable manner (eg, by means of firmware).

The logic and/or steps represented in the flowcharts or otherwise described herein may be embodied in any readable storage medium for instruction execution system, apparatus or device (such as a computer-based system, including a processor). system or other system that can fetch instructions from and execute instructions from an instruction execution system, device or device), or be used in conjunction with such instruction execution system, device or device.

The above are only specific embodiments of the present invention, but the protection scope of the present invention is not limited thereto. Any person familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present invention. All are covered by the protection scope of the present invention. Therefore, the protection scope of the present invention should be subject to the protection scope of the claims.

Claims (10)

1. A method for monitoring discontinuous deformation of a ground-based radar, the method comprising:

s1, calculating monitoring target deformation caused by the displacement of a radar space baseline along the directions of an x axis, a y axis and a z axis based on interference patterns obtained at different time points in intermittent deformation monitoring;

s2, based on the deformation of the monitoring target and the relation between the radar echo phase change and the radar distance change, an x-axis, y-axis and z-axis repositioning error model is established;

s3, calculating an interference phase model of a PS point in the monitoring scene based on the repositioning error model and combining a target pitch angle span; the interference phase model is as follows:

wherein,、/>and->Is the parameter to be estimated, +.>Is azimuth angle, +>、/>、/>Respectively representing phase changes caused by the displacement of the radar along the directions of the x axis, the y axis and the z axis;

and S4, estimating parameters in the interference phase model to obtain a repositioning error curved surface.

2. The method according to claim 1, wherein in S1, the monitoring target deformation amount caused by the offset along the x-axis direction is:

wherein,indicating the radar initial position +.>Indicating radar position offset post position->For the target position +.>Three-dimensional skew representing target and radar, +.>Represents the offset of the radar along the x-axis, < >>The pitch angle at the PS point is indicated.

3. The method according to claim 1, wherein in S1, the monitoring target deformation amount caused by the offset along the y-axis direction is:

wherein,indicating the radar initial position +.>Indicating radar position offset post position->For the target position +.>Three-dimensional skew representing target and radar, +.>Represents the offset of the radar along the y-axis, < >>The pitch angle at the PS point is indicated.

4. The method according to claim 1, wherein in S1, the monitoring target deformation amount caused by the offset along the z-axis direction is:

wherein,indicating the radar initial position +.>Indicating radar position offset post position->For the target position +.>Three-dimensional skew representing target and radar, +.>Representing the offset of the radar along the z-axis, < >>The pitch angle at the PS point is indicated.

5. The method according to claim 1, wherein in S2, the relocation error model is:

wherein,represents the offset of the radar along the x-axis, < >>Represents the offset of the radar along the y-axis, < >>Representing the offset of the radar along the z-axis, < >>Pitch angle, denoted PS Point, +.>Representing the radar signal wavelength.

6. The method according to claim 1, wherein in S3, when the interference phase for PS points in the monitored scene is in matrix form, the interference phase isThe matrix form of the interferometric phase model is then converted into:

wherein,error of representation model +.>Represents the interference phase in matrix form, azimuth angle of PS point +.>，

。

7. The method according to claim 1, wherein in the interferometric phase model build, a target pitch angleIn which value +.>Instead, namely:

wherein,the target pitch span is +.>，/>Represents the offset of the radar along the x-axis, < >>Represents the offset of the radar along the y-axis, < >>Representing the offset of the radar along the z-axis, < >>Representing the radar signal wavelength.

8. The method according to claim 1, wherein in S3, for screening PS points, a dual threshold of coherence coefficient and amplitude information is used for screening;

and in the phase unwrapping stage after PS point screening, phase unwrapping is carried out along the azimuth direction.

9. A ground-based radar intermittent deformation monitoring system, the system comprising:

the monitoring target deformation amount calculation module is used for calculating the monitoring target deformation amount caused by the displacement of the radar space base line along the directions of the x axis, the y axis and the z axis based on the interference patterns obtained at different time points in the intermittent deformation monitoring;

the repositioning error model module is used for establishing an x-axis, y-axis and z-axis repositioning error model based on the deformation of the monitoring target and the relation between the radar echo phase change and the radar distance change of the monitoring target;

the interference phase model module is used for calculating an interference phase model of a PS point in a monitoring scene based on the repositioning error model and combined with a target pitch angle span; the interference phase model is as follows:

wherein,、/>and->Is the parameter to be estimated, +.>Is azimuth angle, +>、/>、/>Respectively representing phase changes caused by the displacement of the radar along the directions of the x axis, the y axis and the z axis;

and the parameter estimation module is used for estimating parameters in the interference phase model to obtain a repositioning error curved surface.

10. The system of claim 9, wherein the system further comprises:

the image registration module is used for carrying out image registration based on the amplitude phase coherence of different SLC images;

the interference pattern generation module is used for generating an interference pattern based on the registered image;

the PS point screening module adopts a dual threshold value of the coherence coefficient and the amplitude information to carry out PS point screening;

the phase unwrapping module is used for carrying out phase unwrapping on the interference pattern after PS point screening, and the image data after phase unwrapping is input into the monitoring target deformation amount calculating module;

and the deformation calculation module is used for calculating the deformation based on the repositioning error curved surface obtained by the parameter estimation module.