CN115267776A - Multi-baseline phase unwrapping method considering robust Chinese remainder theorem of branch tangent lines - Google Patents

Abstract

The invention discloses a multi-baseline phase unwrapping method considering the robust Chinese remainder theorem of branch and tangent lines. The method includes the following steps: generating a multi-baseline InSAR phase interferogram for the same region of interest; The interferogram sets the branch tangent, and adopts the Chinese remainder theorem to avoid the branch tangent and takes an appropriate approach to obtain the large-scale unwrapped phase; for the branch tangent distribution area, the branch tangent in a local range is generated, and the Chinese remainder theorem is used again for unwrapping. ; Combine the phase unwrapping information of a large area with a local area to obtain the final phase unwrapping information. This scheme uses branch tangents and adaptive rectangular frames to extract areas with strong noise, uses related filtering algorithms to reduce the interference of noise on interference phase information, and uses two Chinese remainder theorem algorithms to disentangle large areas with weak noise information and noise information respectively. Strong local area, splicing its unwrapped phase information to achieve anti-noise multi-baseline phase unwrapping algorithm.

Description

A Multi-Baseline Phase Unwrapping Method Considering the Robust Chinese Remainder Theorem of Branch Tangents

technical field

The invention relates to a multi-baseline phase unwrapping method based on the robust Chinese remainder theorem considering branch tangents, which belongs to the field of microwave remote sensing measurement technology, and is mainly applied in the field of multi-baseline, high-precision phase information unwrapping and digital elevation model reconstruction .

Background technique

In order to obtain the elevation information of the steep and undulating areas that are difficult to measure manually, the Synthetic Aperture Radar Interferometry (InSAR) technology combines signal processing and image processing to achieve digital elevation models for terrain surveying and mapping without on-site measurement ( DEM), which is one of the modern new measurement techniques of radar microwave remote sensing measurement, which is developing rapidly. At the same time, with its all-weather, all-day, penetrating working characteristics, it provides technical support and technical means for digital image processing. At present, this measurement method provides data support and process management for mountain surveying and mapping, environmental monitoring, geological exploration and climate analysis, and is of great value for further research.

Phase unwrapping technology has always been the key in synthetic aperture radar interferometry technology. For the phase unwrapping technology in the inversion of elevation information, the performance of the unwrapping algorithm is directly related to the accuracy and accuracy of the elevation information. Phase unwrapping has developed into a unique and multidisciplinary technology during the research process in recent decades. Especially with the development of deep learning algorithms, the improvement of traditional algorithms and the proposal of new algorithms will inevitably become phase unwrapping technology. New trends in technology.

At present, the methods of phase unwrapping mainly include single-baseline phase unwrapping algorithm and shifted multi-baseline phase unwrapping algorithm based on the number of baselines. The single-baseline phase unwrapping algorithm is limited by the assumption of phase continuity, such as the branch cut method. Although the branch tangent line is set, the area containing residual points is effectively avoided, but the unwrapping effect on the phase undersampled area is not good. , the multi-baseline phase unwrapping algorithm solves the fuzzy numbers of phase differentials by introducing multiple phase interferograms, and is widely used in complex terrains with discontinuous phases, such as the Chinese remainder theorem, which can accurately and simultaneously solve multiple The fuzzy number corresponding to a single point of the interferogram has high accuracy, but it is greatly disturbed by noise, and its noise robustness is poor, so it cannot be used on a large scale in the measured data.

Based on the characteristics of these two methods, it is possible to use the branch-cut method to set up an effective way to unwrap the phase ambiguous numbers with a small degree of noise in a large area in the phase unwrapping of the path and its extension of the Chinese remainder theorem, and then use the adaptive rectangular Extract the points with large noise information by means of frames, filter them, and use the Chinese remainder theorem unwrapping algorithm to process the phase fuzzy numbers in this area, and finally splicing the unwrapped phases of large-scale areas and each rectangular frame area .

Contents of the invention

The present invention aims to overcome the shortcomings of the phase unwrapping method in the prior art, the single-baseline branch-cut method algorithm has an island phenomenon for phase mutation unwrapping, and the multi-baseline Chinese remainder theorem algorithm has weak anti-noise, and provides a method that takes into account the branch-cut line The multi-baseline phase unwrapping method based on the robust Chinese remainder theorem adopts the single-baseline phase unwrapping algorithm combined with the multi-baseline phase unwrapping method, which not only solves the problem of difficult unwrapping in the island area, but also enhances the noise robustness of the algorithm sex.

To achieve the above object, the present invention adopts the following technical solutions:

A kind of multi-baseline phase unwrapping method of the robust Chinese remainder theorem considering branch tangents of the present invention, said method comprises the following steps:

Generate multi-baseline InSAR phase interferograms for the same region of interest;

Using the branch cutting method to set the branch tangent line on the interferogram, and adopting the Chinese remainder theorem to avoid the branch tangent line and adopting a suitable way to obtain the large-scale unwrapped phase;

For the branch tangent distribution area in the interferogram, the branch tangent method is used to generate the branch tangent in the local range, and the Chinese remainder theorem is used to unwrap to obtain the unwrapping information of the local area;

The phase unwrapping information of the large-scale area and the unwrapping information of the local area are combined to obtain the phase unwrapping information of all points in the interferogram.

As preferably, said generating multi-baseline InSAR phase interferogram for the same region of interest further includes:

Step 1, obtain multiple multi-baseline SAR images for the same area, select a main image and multiple auxiliary images;

In step 2, the main image and the auxiliary image are subjected to differential interference processing one by one, and arranged according to the length of the baseline to generate multiple multi-baseline InSAR phase interferograms to be unwrapped.

As a preference, the described adopting the branch cutting method to set the branch tangent line on the interferogram, and adopting the Chinese remainder theorem to avoid the branch tangent line and adopting a suitable way to obtain the large-scale unwrapped phase further includes:

Step 3, using the branch cut method to preprocess the interferogram to generate branch cut lines;

Step 4, establish a multi-baseline InSAR geometric model according to the multiple interferograms, and build a Chinese remainder theorem mathematical model according to the multi-baseline InSAR geometric model;

Step 5, according to the preprocessed distribution of branch tangents, avoid branch tangents and adopt a reasonable approach, use the Chinese remainder theorem to unwrap the interferogram, and obtain phase unwrapping information in a large area.

As a preference, for the branch tangent distribution area in the interferogram, the branch tangent method is used to generate the branch tangent in the local range, and the Chinese remainder theorem is used to unwrap to obtain the unwrapping information of the local area, further comprising:

Step 6: Extract one by one with an adaptive rectangular frame for the branch tangent distribution area in the interferogram;

Step 7: Use the branch cutting method to reselect the expansion point for the interference phase in the frame, and generate the branch tangent line in the local range;

Step 8: Filtering the phase information of the points on the branch tangent line in the local range;

Step 9: Unwrap the interferometric phase after filtering in the rectangular frame using the Chinese remainder theorem again to obtain the unwrapped information of the local area.

As preferably, said step 4 further includes:

The parameter expressions of the multi-baseline InSAR system geometric model are as follows:

In the formula, h represents the height difference, λ represents the wavelength, θ represents the side viewing angle, α represents the horizontal angle of the baseline, B _i represents different baseline lengths, R represents the oblique distance between the phase center and the target point, and _ki represents the interference of different points A fuzzy number for the phase.

If you set B ₀ =[B ₁ ,B ₂ ,…,B _n ], where [] means to find the least common multiple, and set the modulus length m _i =B ₀ /B _i , the interferometric phase can be expressed by height difference:

则不同基线下干涉图中对应于各个点位干涉相位的所要求解的模糊数满足以下等式：If set

Then the fuzzy numbers to be solved corresponding to the interferometric phase of each point in the interferogram under different baselines satisfy the following equation:

T＝w _i +k _i m _i

Corresponding to N interferograms formed by N sets of baselines, each corresponding point is composed of N sets of equations, and in each set of equations, m _i is determined, w _i is different, and T is the same. According to mathematical theory , that is, the congruence equations are constructed, and the final fuzzy number _ki to be obtained is the process of solving the mathematical model of the Chinese remainder theorem.

Preferably, the step 5 further includes: according to the pre-processed distribution of branch tangents, for areas with abrupt changes in terrain derived from many branch tangents, unwrapping the interferogram by using the Chinese remainder theorem to obtain phase unwrapping information for a wide range of areas.

Preferably, in the step 6, the size of the interferogram is M×N, the size of the rectangular frame is u×v, and the size of the adaptive rectangular frame satisfies the formula:

u＝Floor[0.1×M+Ω]

v=Floor[0.1×N+Ω]

where Ω is the signal-to-noise ratio of the interferogram.

Preferably, the filtering process in step 8 adopts a median filtering process.

A single-baseline phase unwrapping algorithm combined with a multi-baseline phase unwrapping method provided by the present invention takes into account the limitations of the islanding phenomenon of the branch-cut method and the weak anti-noise ability of the Chinese remainder theorem algorithm, and uses the branch-cut line and self-adaptive rectangular frame Extract the area with strong noise, use the relevant filtering algorithm to reduce the interference of noise on the interferometric phase information, use the Chinese remainder theorem algorithm twice to unwrap the large area area with weak noise information and the local area with strong noise information, and splicing the unwrapped results The phase information implements a noise-resistant multi-baseline phase unwrapping algorithm.

Description of drawings

Fig. 1 is the differential processing of the interferogram to be unwrapped in the third embodiment of the present application.

Fig. 2 is the multi-baseline InSAR geometric model of the third embodiment of the present application.

Fig. 3 is a three-dimensional and two-dimensional reference map corresponding to the real DEM data of the Isolation Peak National Park in the United States in the third embodiment of the present application.

Fig. 4 is the differential interferogram corresponding to the third and third baselines in the embodiment of the present application, in which the baselines are sorted according to the length of the baselines.

FIG. 5 is a flow chart of Embodiment 3 of the present application.

FIG. 6 shows the searched residual points for the three interferograms in the third embodiment of the present application using the branch cutting method to process the global scope.

Fig. 7 is the branch cut line formed by processing the global range with the branch cut method for the three interferograms according to the third embodiment of the present application.

Fig. 8 is the phase of the large-scale unwrapping obtained by using the Chinese remainder theorem to avoid branch tangents and adopting a suitable method in the third embodiment of the present application.

FIG. 9 is a partial branch tangent line formed in a rectangular frame by using the branch cut method to process any undersampled region in Embodiment 3 of the present application.

FIG. 10 shows the unwrapped phase integrated by using the Chinese remainder theorem again after median filtering of all the phase information in the rectangular frame according to Embodiment 3 of the present application.

FIG. 11 is the unwrapped phase of the whole image stitched together with the large-scale unwrapped phase and the local area in the third embodiment of the present application.

FIG. 12 is an error distribution diagram of the difference between the unwrapped phase and the reference phase obtained in Embodiment 3 of the present application.

FIG. 13 is an error histogram of the difference between the unwrapped phase and the reference phase obtained in Embodiment 3 of the present application.

Detailed ways

The present invention will be further described below in conjunction with the accompanying drawings and specific embodiments.

Embodiment one

As shown in Figure 1, a kind of multi-baseline phase unwrapping method of the robust Chinese remainder theorem of branch tangents of the present invention, said method comprises the following steps:

Step S1, generate multi-baseline InSAR phase interferograms for the same region of interest.

Step S2, using the branch cutting method to set branch tangents for the interferogram, and using the Chinese remainder theorem to avoid the branch tangents and obtain the large-scale unwrapped phases in a suitable way.

Step S3, for the distribution area of the branch tangents in the interferogram, use the branch cutting method to generate the branch tangents in the local range, and use the Chinese remainder theorem to unwrap again to obtain the unwrapping information of the local area.

Step S4, combining the phase unwrapping information of the large-scale area and the unwrapping information of the local area to obtain the phase unwrapping information of all points in the interferogram.

This scheme considers the limitations of the island phenomenon of the branch-cut method and the weak anti-noise ability of the Chinese remainder theorem algorithm, uses the branch tangent line and the adaptive rectangular frame to extract the area with strong noise, and uses the relevant filtering algorithm to weaken the influence of noise on the interferometric phase information Interference, use the Chinese remainder theorem algorithm twice to unwrap the large-area area with weak noise information and the local area with strong noise information, and splice the unwrapped phase information to realize the anti-noise multi-baseline phase unwrapping algorithm.

Embodiment two

This embodiment is an optimization scheme based on the first embodiment. In this embodiment, the steps S1, S2, S3, and S4 of the first embodiment are further expanded.

Step S1 specifically further includes the following steps:

Step 1. Acquire multiple multi-baseline SAR images for the same area, and select a main image and multiple auxiliary images.

In step 2, the main image and the auxiliary image are subjected to differential interference processing one by one, and arranged according to the length of the baseline to generate multiple multi-baseline InSAR phase interferograms to be unwrapped.

Step S2 specifically further includes the following steps:

Step 3, using the branch cutting method to preprocess the interferogram to generate branch cutting lines.

Step 4, establish a multi-baseline InSAR geometric model according to the multiple interferograms, and build a Chinese remainder theorem mathematical model according to the multi-baseline InSAR geometric model;

Step 5, according to the preprocessed distribution of branch tangents, avoid branch tangents and adopt a reasonable approach, use the Chinese remainder theorem to unwrap the interferogram, and obtain phase unwrapping information in a large area.

Specifically, the step 3 using the branch cutting method to generate branch tangent lines involves two-dimensional phase unwrapping theory and residual points.

存在以下关系：First of all, the two-dimensional phase unwrapping is for the rectangular space of M×N size, where a point in a space is selected as the starting point to expand the phase in different directions, then the two-dimensional phase unwrapping path is divided into horizontal and Vertical direction, where the interferometric phase φ and the unwrapping phase of the starting point are set

The following relationship exists:

Among them, W[φ(i,j)] is defined as the phase winding function, and in the path mentioned above, the phase gradients in the horizontal and vertical directions can be defined as:

△ _i φ(i,j)=φ(i,j)-φ(i-1,j) (2)

△ _j φ(i,j)=φ(i,j)-φ(i,j-1) (3)

According to the different paths, use the phase wrapping function to wrap the phase gradients in the vertical and horizontal directions of adjacent pixels:

△ _i W[φ(i,j)]＝△ _i φ(i,j)+2π△ _i k(i,j) (5)

△ _j W[φ(i,j)]＝△ _j φ(i,j)+2π△ _j k(i,j) (6)

Using the phase wrapping function again for the above wrapping phase gradients in different orientations, the vertical and horizontal wrapping phases of adjacent pixels can be obtained:

W ₂ [△ _i W ₁ [φ(i,j)]＝△ _i φ(i,j)+2π[△ _i k ₁ (i,j)+△ _i k ₂ (i,j)] (7)

W ₃ [△ _j W ₁ [φ(i,j)]＝△ _j φ(i,j)+2π[△ _j k ₃ (i,j)+△ _j k ₄ (i,j)] (8)

To ensure that the winding phase is between -π and π, the corresponding parameters in the formula must satisfy the following relationship:

π≤△ _i φ(i,j)<π (9)

π≤△ _j φ(i,j)<π (10)

△ _i k ₁ (i, j) + △ _i k ₂ (i, j) = 0 (11)

△ _j k ₃ (i, j) + △ _j k ₄ (i, j) = 0 (12)

Next, reliable path integration will be adopted to achieve unwrapping, first vertical integration and then horizontal integration, first horizontal integration and then vertical integration.

Integrate vertically and then horizontally:

First, the first column in the vertical direction is unwrapped:

After the pixels in the first column are unwrapped, it is used as a new starting column. At this time, each column performs horizontal gradient integration processing on the previous column in turn. After j columns are processed, the unwrapping process is completed.

Integrate first horizontally and then vertically:

Firstly, the first column in the horizontal direction is processed for pixel unwrapping:

After the pixels in the first row are unwrapped, it is used as a new starting row. At this time, each row performs horizontal gradient integration processing on the previous row in turn, and the unwrapping process is completed after the i row is processed.

Ideally, if the phase is defined as continuous, according to any one of the above two path integration methods, the winding phase at one point can be extended to the real phase of the whole image, but in reality, if there is noise, the path The selection will produce deviations, and even accumulate point errors and expand them into larger interval errors.

Based on the above situation, it is necessary to introduce the concept of residual point, that is, to determine the phase continuity by solving the gradient of each azimuth of adjacent pixels:

△ ₁ = φ(i+1,j)-φ(i,j) (17)

△ ₂ ＝φ(i+1,j+1)-φ(i+1,j) (18)

△ ₃ ＝φ(i,j+1)-φ(i+1,j+1) (19)

△ ₄ ＝φ(i,j)-φ(i,j+1) (20)

The sum of all azimuth gradients solved is the residual:

The continuity of the phase information is judged by the residual of the solution, that is: T=0, the point is judged as a phase continuous point, T≠0, the point is judged as a residual point, usually defined as: T>0, the point is judged as positive Residual point, T<0, this point is judged as a negative residual point.

The introduction of residual points has a guiding effect on the path tracking algorithm, especially for the single baseline unwrapping algorithm, which guarantees the assumption of phase continuity in a large area.

Specifically, the step 4 further includes:

The parameter expressions of the multi-baseline InSAR system geometric model are as follows:

In the formula, h represents the height difference, λ represents the wavelength, θ represents the side viewing angle, α represents the horizontal angle of the baseline, B _i represents different baseline lengths, R represents the oblique distance between the phase center and the target point, and _ki represents the interference of different points A fuzzy number for the phase.

If B ₀ =[B ₁ ,B ₂ ,…,B _n ] is set, where [] means to find the least common multiple, and the modulus length m _i =B ₀ /B _i is set at the same time, the interferometric phase can be expressed by height difference:

则不同基线下干涉图中对应于各个点位干涉相位的所要求解的模糊数满足以下等式：If set

Then the fuzzy numbers to be solved corresponding to the interferometric phase of each point in the interferogram under different baselines satisfy the following equation:

T=w _i +k _i m _i (24)

Corresponding to N interferograms formed by N groups of baselines, each corresponding point is composed of N groups of equations, and in each group of equations, m _i is determined, w _i is different, and T is the same. According to mathematical theory , that is, the congruence equations are constructed, and the final fuzzy number _ki to be obtained is the process of solving the mathematical model of the Chinese remainder theorem.

Specifically, the step 5 further includes, according to the preprocessed distribution of branch tangents, for areas with abrupt changes in terrain derived from many branch tangents, unwrapping the interferogram by using the Chinese remainder theorem to obtain phase unwrapping information for large-scale areas.

Step S3 specifically further includes the following steps:

Step 6: Extract one by one with an adaptive rectangular frame for the branch tangent distribution area in the interferogram;

Step 7: Use the branch cutting method to reselect the expansion point for the interference phase in the frame, and generate the branch tangent line in the local range;

Step 8: Perform filtering processing on the phase information of the points on the branch tangent line in the local range, and the filtering processing in the step 8 adopts median filtering processing;

Step 9: Unwrap the interferometric phase after filtering in the rectangular frame using the Chinese remainder theorem again to obtain the unwrapped information of the local area.

Specifically, in the step 6, the size of the interferogram is M×N, the size of the rectangular frame is u×v, and the size of the adaptive rectangular frame satisfies the formula:

u＝Floor[0.1×M+Ω] (25)

v=Floor[0.1×N+Ω] (26)

where Ω is the signal-to-noise ratio of the interferogram.

Embodiment Three

This embodiment is to apply the method of Embodiment 2, using the data of the IsolationPeak National Park in the United States as a specific operation example, to describe the solution of this application in detail.

As shown in Figure 1, Figure 2, Figure 3, and Figure 4, the data used in this embodiment are the main image of one scene and the auxiliary images of three scenes, and the corresponding interferograms under three sets of different baselines are generated according to the multi-baseline geometric model and the principle of difference as the data to be unwrapped. As shown in FIG. 5 , the specific implementation process of the method of this embodiment will be further described below.

Step 1: Divide the four scene images into the main image and the auxiliary image, perform differential interference processing, and arrange them according to the length of the baseline to form a multi-baseline InSAR phase interferogram.

Step 2: For the accuracy of data and comparative analysis of experiments, Figure 3 shows the 3D and 2D reference images of the real DEM data of the Isolation Peak National Park in the United States.

计算，正残差点用1表示，负残差点用-1表示。Step 3: For the three interferograms, use the branch cutting method to process the positive and negative residual points searched in the global range, and use the formula

For calculation, positive residual points are represented by 1, and negative residual points are represented by -1.

Step 4: Use the residual points extracted in Step 3 to set branch tangents to the global phase, which serves as a path tracking guide.

Step 5: Based on the generated branch tangents in the interferogram, use the Chinese remainder theorem to avoid the branch tangents and obtain the large-scale unwrapped phases in a suitable way.

Fig. 6 is the residual points searched by using the branch cutting method to process the global range of the three interferograms, and Fig. 7 is the branch tangent line formed by using the branch cutting method to process the global range on the three interferograms.

Fig. 8 is the large-scale unwrapped phase obtained by using the Chinese remainder theorem to avoid branch tangents and adopting a suitable approach.

Step 6: Unwrapped phase generation in the global range, it is necessary to consider the area of local undersampling and phase mutation, and use the branch cutting method to process the regions of undersampling and phase mutation one by one to form the local branch tangent line in the rectangular frame , and mark the position, where there are many phase mutation points in an under-sampled area during the experiment, in order to ensure 60% stable phase information, the size of the adaptive rectangular frame is expanded to 98*52, as shown in Figure 9.

Step 7: Perform median filter processing on the phase information of the points on the branch tangent line. If there are only 1-2 phase mutation points in the rectangular frame, it is assumed that the unwrapping effect in the rectangular frame is good, and there is no need for filtering to unwrap again.

Step 8: Unwrap the interferometric phase after filtering in the rectangular frame using the Chinese remainder theorem again, and the result is shown in Figure 10.

Step 9: Combining the overall phase unwrapping information and the small area unwrapping information, the phase unwrapping information of all points in the interferogram is obtained; as shown in Figure 11, it is the whole area spliced by the large-scale unwrapping phase and the local area. unwrapped phase of an image.

Step 10: In order to verify the anti-noise performance of the algorithm and the unwrapping performance at the sudden phase change, make an error distribution diagram and an error histogram of the difference between the unwrapped phase and the reference phase.

It can be seen from Figure 11 that the multi-baseline phase unwrapped map integrated by the algorithm is in good agreement with the original DEM as a whole, and there is no island phenomenon.

It can be seen from Figure 12 that in the error distribution diagram of the difference between the unwrapped phase and the reference phase by this algorithm, the overall phase error value is relatively low, and only a few bright spots are neglected, and only 1-2 in the rectangular frame have not been filtered. The phase mutation point of , the error value is larger.

It can be seen from Figure 13 that the distribution of the error histogram is also around 0rad, and the frequency of mutation points is very low, so the robustness of the algorithm has been improved, and the anti-noise ability has been verified.

Claims (8)

1. a kind of multi-baseline phase unwrapping method of the robust Chinese remainder theorem of branch tangent, it is characterized in that, described method comprises the following steps: Generate multi-baseline InSAR phase interferograms for the same region of interest; Using the branch cutting method to set the branch tangent line on the interferogram, and adopting the Chinese remainder theorem to avoid the branch tangent line and adopting a suitable way to obtain the large-scale unwrapped phase; For the branch tangent distribution area in the interferogram, the branch tangent method is used to generate the branch tangent in the local range, and the Chinese remainder theorem is used to unwrap to obtain the unwrapping information of the local area; The phase unwrapping information of the large-scale area and the unwrapping information of the local area are combined to obtain the phase unwrapping information of all points in the interferogram. 2. The multi-baseline phase unwrapping method of the robust Chinese remainder theorem considering branch tangents according to claim 1, wherein said generation of multi-baseline InSAR phase interferograms for the same region of interest further comprises: Step 1, obtain multiple multi-baseline SAR images for the same area, select a main image and multiple auxiliary images; In step 2, the main image and the auxiliary image are subjected to differential interference processing one by one, and arranged according to the length of the baseline to generate multiple multi-baseline InSAR phase interferograms to be unwrapped. 3. The multi-baseline phase unwrapping method of the robust Chinese remainder theorem considering branch tangents according to claim 1, characterized in that, the branch tangent method is used to set branch tangents to the interferogram, and the Chinese remainder is adopted The theorem avoids the branch tangents and adopts a suitable way to obtain the large-scale disentangled phase, which further includes: Step 3, using the branch cut method to preprocess the interferogram to generate branch cut lines; Step 4, establish a multi-baseline InSAR geometric model according to the multiple interferograms, and build a Chinese remainder theorem mathematical model according to the multi-baseline InSAR geometric model; Step 5, according to the preprocessed distribution of branch tangents, avoid branch tangents and adopt a reasonable approach, use the Chinese remainder theorem to unwrap the interferogram, and obtain phase unwrapping information in a large area. 4. The multi-baseline phase unwrapping method considering the robust Chinese remainder theorem of branch tangents according to claim 1, characterized in that, for the distribution area of branch tangents in the interferogram, the branch cutting method is used to generate a local range The branch tangent line in , and use the Chinese remainder theorem to unwrap again to obtain the unwrapping information of the local area, which further includes: Step 6: Extract one by one with an adaptive rectangular frame for the branch tangent distribution area in the interferogram; Step 7: Use the branch cutting method to reselect the expansion point for the interference phase in the frame, and generate the branch tangent line in the local range; Step 8: Filtering the phase information of the points on the branch tangent line in the local range; Step 9: Unwrap the interferometric phase after filtering in the rectangular frame using the Chinese remainder theorem again to obtain the unwrapped information of the local area. 5. The multi-baseline phase unwrapping method considering the robust Chinese remainder theorem of branch tangents according to claim 3, characterized in that, said step 4, further comprising: The parameter expressions of the multi-baseline InSAR system geometric model are as follows:

In the formula, h represents the height difference, λ represents the wavelength, θ represents the side viewing angle, α represents the horizontal angle of the baseline, B _i represents different baseline lengths, R represents the oblique distance between the phase center and the target point, and _ki represents the interference of different points A fuzzy number for the phase. If B ₀ =[B ₁ ,B ₂ ,…,B _n ] is set, where [] means to find the least common multiple, and the modulus length m _i =B ₀ /B _i is set at the same time, the interferometric phase can be expressed by height difference:

If set

Then the fuzzy numbers to be solved corresponding to the interferometric phase of each point in the interferogram under different baselines satisfy the following equation: T＝w _i +k _i m _i Corresponding to N interferograms formed by N sets of baselines, each corresponding point is composed of N sets of equations, and in each set of equations, m _i is determined, w _i is different, and T is the same. According to mathematical theory , that is, the congruence equations are constructed, and the final fuzzy number _ki to be obtained is the process of solving the mathematical model of the Chinese remainder theorem. 6. The multi-baseline phase unwrapping method considering the robust Chinese remainder theorem of branch tangents according to claim 3, characterized in that, said step 5, further comprising: according to the preprocessed branch tangent distribution, for branch tangent derivatives In areas with more abrupt terrain changes, the Chinese remainder theorem is used to unwrap the interferogram to obtain phase unwrapped information in a large area. 7. The multi-baseline phase unwrapping method considering the robust Chinese remainder theorem of branch tangents according to claim 4, characterized in that, in the step 6, the size of the interferogram is M×N, and the size of the rectangular frame is u×v, the size of the adaptive rectangular frame satisfies the formula: u＝Floor[0.1×M+Ω] v=Floor[0.1×N+Ω] Among them, Ω is the signal-to-noise ratio of the interferogram. 8. The multi-baseline phase unwrapping method considering the robust Chinese remainder theorem of branch tangents according to claim 4, characterized in that, the filtering process in the step 8 adopts median filtering process.

Images