A wetland water level measurement method, device, equipment and readable storage medium

Abstract

The present invention provides a wetland water level measurement method, device, equipment and readable storage medium, which relate to the field of synthetic aperture radar interferometry technology. The method also includes the following steps: determining homogeneous pixels of a differential interference fringe pattern, and extracting distributed scatterers based on the homogeneous pixels; performing adaptive terrain filtering on the differential interference fringe pattern after extracting the distributed scatterers. The present invention can determine precise scatterer points and more accurately extract water level information, thereby performing more accurate wetland water level measurement, obtaining more accurate wetland water level changes, and ensuring the accuracy of wetland hydrological information extraction.

Description

Technical Field

[0001] The present invention relates to the field of synthetic aperture radar interferometry technology, and in particular to a wetland water level measurement method, device, equipment and readable storage medium. Background Art

[0002] The uniqueness and importance of wetlands have attracted worldwide attention. Remote sensing technology has been widely used in wetland surveys due to its advantages of high spatial resolution, wide coverage and low labor and economic costs. The relevant applications are mainly concentrated in wetland identification and classification, wetland resource surveys, wetland resource dynamic change surveys and wetland vegetation biomass estimation. However, for wetland ecologists, they are more concerned about water level, water area, vegetation height and vegetation coverage. These hydrological parameters and vegetation parameters directly affect the environment of migratory bird habitat selection. In wetland ecosystems, changes in wetland hydrological conditions such as water level, water volume, flooding frequency and flooding time have always been the focus of wetland protection. They play a leading role in the distribution of wetland organisms and the properties of wetland soils, and together with the growth of wetland vegetation, they affect the habitat environment of wetland animals. Optical remote sensing cannot directly solve the problem of measuring wetland water level and vegetation height. At the same time, vegetation coverage will have a great impact on the accuracy of optical remote sensing wetland water body identification. Radar altimeters have been used to monitor water levels in large inland water bodies such as large lakes and the Amazon forest. However, for small lakes or wetlands, the radar altimeter echo will be deformed due to the small area of ​​the water body, and the influence of the surrounding terrain will cause a large amount of data loss. The existence of the above problems makes remote sensing methods such as optical remote sensing and radar altimeters unable to meet the needs of wetland ecologists for wetland water levels and vegetation heights.

[0003] The backscattered signal of Synthetic Aperture Radar (SAR) is very sensitive to terrain slope, surface roughness and dielectric constant, and can penetrate vegetation to solve the problem of optical data underestimating the water area. It has been widely used in the definition of wetland types, wetland conditions and overflow characteristics. Since SAR not only measures the amplitude of the echo reflected by the ground target, but also records the phase information of the echo, it has a unique advantage in height inversion. Interferometry technology has been widely used in surface deformation monitoring and ground elevation measurement, and has also made great progress in wetland water level monitoring. The launch of new sensors has expanded the application capabilities of SAR, especially the X-band land radar attached digital elevation model (TanDEM-X) dual-base system, which greatly reduces the impact of temporal decorrelation, making wetland vegetation height inversion possible. Therefore, the unique advantages of SAR make it have great potential in the inversion of wetland hydrology and vegetation parameters.

[0004] For the assessment of the health level of wetland systems, it is necessary to continuously obtain long-term hydrological monitoring data. Persistent scatterer (PS) technology and small baseline subset (SBAS) technology obtain long-term series surface deformation through continuous SAR observations. The development of these two technologies has promoted the significant progress of SAR technology application capabilities. However, due to the serious decoherence of ground objects in wetland areas, the surface coverage in wetland areas is very different from that in urban areas. There are no large numbers of artificial buildings, and the main ground objects are vegetation and water surfaces. The main problem is how to determine the precise scatterer points. At the same time, the traditional method is not accurate enough for extracting water level information.

[0005] Therefore, a wetland water level measurement method that can determine precise scatterer points and extract water level information more accurately is an important issue that needs to be urgently addressed in the industry. Summary of the invention

[0006] The present invention provides a wetland water level measurement method, device, equipment and readable storage medium, which are used to solve the defects in the prior art that the precise scatterer points in the area cannot be determined and the accuracy of water level information extraction is insufficient, so as to achieve more accurate wetland water level measurement and obtain more accurate wetland water level changes.

[0007] The present invention provides a wetland water level measurement method, comprising the following steps:

[0008] The synthetic aperture radar image to be processed is registered and calibrated to obtain registered composite data; wherein the composite data is an amplitude image and phase information corresponding to the amplitude image, the composite data is a time series synthetic aperture radar image, and the synthetic aperture radar image to be processed is an image of a wetland area;

[0009] Based on the coherence of the interference image pairs, the composite data are paired and combined to obtain a plurality of interference image pairs of the composite data, and a differential interference fringe pattern is generated according to the interference image pairs;

[0010] Performing phase unwrapping on the differential interference fringe pattern to obtain an unwrapped differential interference fringe pattern;

[0011] Based on the small baseline set technology, the water level change of the unwrapped differential interference fringe pattern is obtained. Before performing phase unwrapping on the differential interference fringe pattern to obtain the unwrapped differential interference fringe pattern, the method further includes the following steps:

[0012] Determining homogeneous pixels of the differential interference fringe pattern, and extracting distributed scatterers based on the homogeneous pixels;

[0013] Adaptive terrain filtering is performed on the differential interference fringe pattern after the distributed scatterer is extracted.

[0014] According to the wetland water level measurement method provided by the present invention, the determining of homogeneous pixels of the differential interference fringe pattern and the extraction of distributed scatterers based on the homogeneous pixels specifically include the following steps:

[0015] Obtaining the connectivity number of all pixels in the preset window corresponding to each pixel of the differential interference pattern, and taking the pixels with a connectivity number higher than the preset number as the homogeneous pixels in the preset window; wherein the pixel is the center point of the preset window corresponding to the pixel, and the probability distribution of the multi-phase backscattering coefficient values ​​of the homogeneous pixels between two adjacent pixels in the statistical test is the same;

[0016] Removing the homogeneous pixels that are not connected to the center point, and determining a distributed scatterer block corresponding to the center point;

[0017] The coherence of the distributed scatterer block is obtained, and the distributed scatterers of the distributed scatterer block having a higher coherence than a first coherence are extracted.

[0018] According to the wetland water level measurement method provided by the present invention, after obtaining the coherence of the distributed scatterer block and extracting the distributed scatterers of the distributed scatterer block having a higher coherence than the first coherence, the method further comprises the following steps:

[0019] The flat ground phase and the terrain phase are removed from the distributed scatterer.

[0020] According to the wetland water level measurement method provided by the present invention, the water level change of the unwrapped differential interference fringe pattern is obtained based on the small baseline set technology, which specifically includes the following steps:

[0021] constructing a linear model through the distributed scatterers;

[0022] Based on the L1 norm minimization algorithm, the L2 norm minimization algorithm and the optimal interference network, the relative water level change of the linear model is obtained;

[0023] Based on the L1 norm minimization algorithm, the relative water level change is integrated to obtain a relative water level sequence;

[0024] Comparing the relative water level sequence with the reference water level observation data to obtain a linear offset of the absolute value of the water level;

[0025] The linear offset in the relative water level sequence is removed to obtain an absolute water level sequence.

[0026] According to the wetland water level measurement method provided by the present invention, the relative water level change is integrated based on the L1 norm minimization algorithm to obtain the relative water level sequence step, and the terrain error, atmospheric noise and orbit error are removed; wherein the atmospheric noise is obtained using numerical weather forecasting.

[0027] According to the wetland water level measurement method provided by the present invention, the composite data is paired and combined based on the coherence of the interference image pairs to obtain a plurality of interference image pairs of the composite data, and a differential interference fringe pattern is generated according to the interference image pairs, which specifically includes the following steps:

[0028] Selecting the interference image pairs having higher coherence than the second coherence as the interference image pairs of the composite data, and performing pairing and combination on the composite data to obtain a plurality of the interference image pairs of the composite data;

[0029] Based on the minimum spanning tree method, the basic interference network is constructed;

[0030] Adding the plurality of interference image pairs of the composite data to the basic interference network to obtain the optimal interference network;

[0031] A differential interference fringe pattern is generated according to the plurality of interference image pairs of the composite data.

[0032] According to the wetland water level measurement method provided by the present invention, the synthetic aperture radar image to be processed is registered and calibrated to obtain the registered composite data, which specifically includes the following steps:

[0033] One of the synthetic aperture radar images to be processed is used as a main image, and the other synthetic aperture radar images to be processed are registered to the main image; wherein the main image has a stable scattering characteristic.

[0034] The present invention also provides a wetland water level measuring device, comprising:

[0035] A registration module, used for registering and calibrating the synthetic aperture radar image to be processed to obtain registered composite data; wherein the composite data is an amplitude image and phase information corresponding to the amplitude image, the composite data is a time series synthetic aperture radar image, and the synthetic aperture radar image to be processed is a wetland image;

[0036] A pairing and combining module, used for pairing and combining the composite data based on the coherence of the interference image pairs to obtain a plurality of interference image pairs of the composite data, and generating a differential interference fringe pattern according to the interference image pairs;

[0037] A phase unwrapping module, used for performing phase unwrapping on the differential interference fringe pattern to obtain an unwrapped differential interference fringe pattern;

[0038] A water level measurement module is used to obtain the water level change of the unwrapped differential interference pattern based on the small baseline set technology. Before the phase unwrapping module, the device also includes:

[0039] A scatterer extraction module, used to determine homogeneous pixels of the differential interference fringe pattern, and extract distributed scatterers based on the homogeneous pixels;

[0040] The adaptive filtering module is used to perform adaptive terrain filtering on the differential interference fringe pattern after extracting the distributed scatterer.

[0041] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of any of the wetland water level measurement methods described above are implemented.

[0042] The present invention also provides a non-transitory computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the steps of any of the wetland water level measurement methods described above are implemented.

[0043] The wetland water level measurement method, device, equipment and readable storage medium provided by the present invention determine the homogeneous pixels of the differential interference fringe pattern, extract the distributed scatterers based on the homogeneous pixels, and further determine the scatterers that maintain high coherence, so as to determine the stable scatterers, and eliminate the influence of external errors by adaptive terrain filtering on the differential interference fringe pattern after extracting the distributed scatterers, so as to improve the coherence of the scatterers. Combining the above two points with the existing technology, it is possible to determine the precise scatterer points and extract the water level information more accurately, thereby performing more accurate wetland water level measurement, obtaining more accurate wetland water level changes, and ensuring the accuracy of wetland hydrological information extraction. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0045] Figure 1 The process diagram of the wetland water level measurement method provided by the present invention is as follows Figure 1 ;

[0046] Figure 2 The specific process of step S300 in the wetland water level measurement method provided by the present invention is Figure 1 ;

[0047] Figure 3 The specific process of step S300 in the wetland water level measurement method provided by the present invention is Figure 2 ;

[0048] Figure 4 This is the specific process intention of step S600 in the wetland water level measurement method provided by the present invention;

[0049] Figure 5 This is the specific process intention of step S200 in the wetland water level measurement method provided by the present invention;

[0050] Figure 6 It is a structural schematic diagram of the wetland water level measuring device provided by the present invention;

[0051] Figure 7 The specific structure of the scatterer extraction module in the wetland water level measurement device provided by the present invention is shown in FIG. Figure 1 ;

[0052] Figure 8 The specific structure of the scatterer extraction module in the wetland water level measurement device provided by the present invention is shown in FIG. Figure 2 ;

[0053] Fig. 9 It is a specific structural schematic diagram of a water level measurement module in the wetland water level measurement device provided by the present invention;

[0054] Fig.10 It is a specific structural schematic diagram of the paired combination modules in the wetland water level measurement device provided by the present invention;

[0055] Fig.11 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0056] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0057] Combine the following Figure 1 The wetland water level measurement method of the present invention is described, which is applied to data monitoring and evaluation of wetland areas. The method comprises the following steps:

[0058] S100 , registering and calibrating the SAR image to be processed to obtain registered composite data, where the composite data is an amplitude image and phase information corresponding to the amplitude image.

[0059] In this embodiment, in step S100, the polarization mode of the SAR image to be processed adopts horizontal transmission and horizontal reception (HH), and the obtained data is a single looking complex (SLC) image. Therefore, the composite data is a continuous L-band HH polarization (time series) SAR image, and the SAR image to be processed is an image of a wetland area.

[0060] Since the coastal wetland is located at the junction of land and sea, low-coherence surface targets such as the ocean and mudflats occupy a large area on the SAR image. In this embodiment, in order to ensure the accuracy of registration, the registration method is improved. Specifically, step S100 specifically includes the following steps:

[0061] Assuming that the SAR image to be processed has M+1 scene SAR effects, one of the SAR images to be processed is taken as the main image. According to the orbit parameters, pulse repetition frequency, time information and external digital elevation model (DEM) and other information, the other SAR images to be processed (auxiliary images) are registered to the main image, and the pixels between the main and auxiliary images are registered. The gross error elimination method is used to ensure the accuracy of the registration and achieve sub-pixel registration. Among them, the main image, as the basis of registration, has stable scattering characteristics.

[0062] It is understandable that the specific registration accuracy can be set according to actual conditions.

[0063] S200 , based on the coherence of the interference image pairs, perform pairing and combination on the composite data, i.e., perform complex conjugate operation, to obtain a plurality of interference image pairs of the composite data, and generate a differential interference fringe pattern based on the interference image pairs.

[0064] S300, determining the Statically Homogenous Pixel (SHP) of the differential interference pattern, and extracting the Distributed Scatterer (DS) based on the SHP.

[0065] S400, performing adaptive terrain filtering on the differential interference pattern after DS extraction.

[0066] S500, performing phase unwrapping on the differential interference pattern to obtain an unwrapped differential interference pattern;

[0067] S600, based on SBAS technology, obtains the water level change of the unwrapped differential interference pattern.

[0068] The wetland water level measurement method of the present invention is different from the traditional SAR technology in that steps S300 and S400 are added. Step S300 is used to further determine a scatterer that maintains a relatively high coherence, thereby determining a stable scatterer. Step S400 is used to eliminate the influence of external errors, thereby improving the coherence of the scatterer. By combining steps S300 and S400 with the traditional SAR technology, it is possible to determine precise scatterer points and more accurately extract water level information, thereby performing more accurate wetland water level measurements, obtaining more precise wetland water level changes, and ensuring the accuracy of wetland hydrological information extraction.

[0069] Taking the reed wetland as an example, it is a typical DS. The echo signal of each resolution unit on the SAR image is the vector sum of the backscattering of all independent scattering targets in the resolution unit, showing obvious Gaussian distribution characteristics. The DS backscattering energy is relatively low compared to PS, and usually occupies several adjacent pixels on the SAR image. The scattering characteristics of these adjacent pixels have the same probability distribution. The statistical homogeneous pixels (HP) with probability distribution are confirmed by performing a statistical test on the differential interference pattern. By performing a statistical analysis on the registered L-band HH polarization composite data, it is tested whether the multi-phase backscattering coefficient values ​​of two adjacent pixels belong to the same distribution in the statistical test, and the SHP around each pixel can be identified.

[0070] Combine the following Figure 2 The wetland water level measurement method of the present invention is described, and step S300 specifically includes the following steps:

[0071] S310. Obtain the connectivity number of all pixels in the preset window corresponding to each pixel of the differential interference pattern. Specifically, use the Anderson-Darling (AD) test method to test, and take the pixels with a connectivity number higher than the preset number as the SHP in the preset window and retain them; wherein the pixel is the center point of the preset window corresponding to it, and the probability distribution of the multi-phase backscattering coefficient values ​​of the SHP between two adjacent pixels in the statistical test is the same.

[0072] In this embodiment, the preset window is a 11×11 window, and the DS is determined by traversing all pixels through the preset window.

[0073] S320 , removing the SHP that is not connected to the center point, and determining the DS block corresponding to the center point to ensure the spatial connectivity of the DS block corresponding to the center point of the preset window.

[0074] S330, obtaining the coherence of the DS block, and extracting the distributed scatterers of the distributed scatterer block having a coherence higher than the first coherence.

[0075] Preferably, the first coherence is 0.3, that is, if the coherence of the entire DS block is less than 0.3, it is treated as a decoherent block rather than a DS block, and the decoherent block will not participate in subsequent steps.

[0076] Combine the following Figure 3 The wetland water level measurement method of the present invention is described. After step S330, the method further includes the following steps:

[0077] S340. Remove the flat ground phase and terrain phase from DS to eliminate the influence of DEM error and improve the coherence of DS.

[0078] In step S300, a preset window centered on each pixel is first defined, and then a two-sample Gang of Four (GoF) test at a given significance level is applied between each pixel in each preset window, and all pixels that can be considered as SHPs are selected, and pixels that are not directly connected to the center point through other SHPs are discarded, and finally all SHPs connected to the center point are used for subsequent steps. Step S300 determines the SHPs with the same scattering statistical distribution pixel by pixel based on the significance level and connectivity principle, and determines the DS with the connectivity number as the threshold.

[0079] Combine the following Figure 4 The wetland water level measurement method of the present invention is described, and step S600 specifically includes the following steps:

[0080] S610. Construct a linear model through DS.

[0081] S620 , obtaining a relative water level change of a linear model based on an L1 norm minimization algorithm, an L2 norm minimization algorithm and an optimal interference network.

[0082] In each unwrapped differential interference pattern, each coherent pixel (x, y) is i To time t j The water level changes can be estimated by formula (1), which is:

[0083]

[0084] Among them, θ inc is the incident angle of SAR, λ is the wavelength of SAR, Indicates the difference between the coherent pixel (x, y) and the reference point (x 0 ,y 0 ) phase change, n is the noise phase, and the water level survey point is selected as the reference point (x 0 ,y 0 ), the corresponding water level changes can be generated using the water level survey point data

[0085] Deformation parameter estimation has always been a hot spot and difficulty in long-term interferometry. The existing algorithm usually adopts the L2 norm minimization algorithm, which often results in a large number of deformation estimation errors in non-urban areas. In step S620, based on the SBAS technology, the full-resolution differential interferometry fringe phase image processed by adaptive terrain filtering is unwrapped to perform full-resolution interferometry phase estimation. On the basis of maintaining the interferometry phase of the point scatterer, the interferometry phase quality of the DS is improved, and the DS with high coherence coefficient is selected based on the full-resolution differential interferometry fringe phase image. Specifically, the low-pass part of the deformation phase and the residual terrain error (residual phase) are estimated by the least squares algorithm (L2 norm minimization algorithm); for the differential interferometry fringe phase image with the residual terrain error (residual phase) removed, the improved simplex method is used to perform L1 norm minimization calculation to form a parameter error map detection and eliminate gross errors. On this basis, the L2 norm minimization algorithm is used to solve the deformation phase to improve the accuracy and stability of the deformation phase estimation; the singular value decomposition (SVD) algorithm is used to solve the high-resolution nonlinear deformation part.

[0086] S630: Based on the L1 norm minimization algorithm, the relative water level change is integrated to obtain a relative water level sequence.

[0087] In this method, the relative water level change is extracted from the differential interference pattern after adaptive terrain filtering based on the L1 norm minimization algorithm. Specifically, on each differential interference pattern, all coherent pixels (x, y) are unwrapped with a pixel of known water level as a reference. For each coherent pixel (x, y), the optimal interference network forms formula (2), which is:

[0088] BΔh=CΔφ+Δh 0 +n (2)

[0089] in, Δφ is the unwrapping value of the phase value of the differential interference pattern after adaptive terrain filtering, B is the matrix defined by the optimal interference network formed, Δh is the relative water level change vector at adjacent acquisition times on the coherent pixel (x, y), and Δh 0is the relative water level change vector at the water level survey point, and n represents the noise phase, including the removal of terrain error, atmospheric noise and orbit error. Since the wetland area is generally flat (0-5 meters above sea level), the terrain error influence of DEM can be ignored, and the orbit error (translated as linear phase fringes on the differential interference pattern) can be removed by removing the linear trend from the differential interference pattern that has not been unwrapped. In SBAS technology, it is usually assumed that the atmospheric effect is completely unrelated to the change in deformation information over time. The atmospheric effect can be estimated and eliminated by low-pass filtering in the spatial dimension and high-pass filtering in the time dimension. However, since water level changes and atmospheric disturbances are both high-frequency signals in the time dimension, it is impossible to use the atmospheric phase removal method in wetland applications. Therefore, the final water level time series may contain a certain degree of atmospheric noise. In this embodiment, numerical weather forecast (NWP) is used to reduce the atmospheric impact during the processing process. Preferably, the atmospheric noise estimated by NWP is 1.2-4.3cm in the wetland area, i.e., within the range of the SAR image to be processed.

[0090] The SVD algorithm can obtain the L2 norm solution of the L2 norm minimization algorithm. However, in coastal wetlands, there are often some non-correlated areas, such as open water, which separate the areas with high coherence. This often introduces phase unwrapping errors, that is, phase jumps between different areas. The L2 norm minimization algorithm often performs poorly in detecting these phase jumps in the unwrapped data. On the contrary, for the problem that phase unwrapping errors often occur and are difficult to detect in non-urban areas, the L1 norm minimization algorithm can provide a more robust phase inversion solution. The L1 norm solution of relative water level change is shown in formula (3), which is:

[0091]

[0092] in, It is the L1 norm solution, that is, the L1 norm minimization algorithm can give the solution of relative water level change by integration.

[0093] S640: Compare the relative water level sequence with the reference water level observation data to obtain a linear offset of the absolute value of the water level.

[0094] S650: Remove the linear offset in the relative water level sequence to obtain an absolute water level sequence.

[0095] It should be noted that only the relative water level changes can be obtained from phase unwrapping, so it is necessary to calibrate the Synthetic Aperture Radar (SAR, InSAR) observations with ground hydrological observations to obtain absolute water level estimates. To estimate the absolute water level, the relative water level changes collected at adjacent times must first be integrated to obtain a relative water level sequence. The relative sequence is then linked to the reference water level observations to obtain the absolute water level sequence. There is always a linear offset between the relative water level sequence generated by InSAR and the reference water level observations. By comparing the InSAR observed water level sequence with the water level observations, the linear offset can be estimated. After removing the linear offset, the absolute water level sequence inverted by InSAR is generated.

[0096] When pairing imaging pairs, the traditional method only uses the maximum time baseline and space baseline constraints, such as setting an appropriate time baseline and vertical baseline, but does not judge the coherence of the interference image pairs. Therefore, after all the interference image pairs are subjected to differential interference pattern generation, there will inevitably be some interference image pairs with low coherence. The differential interference pattern and unwrapping results generated by these interference image pairs are not ideal. Therefore, these interference image pairs need to be eliminated before subsequent processing can be carried out.

[0097] Combine the following Figure 5 The wetland water level measurement method of the present invention is described, and step S200 specifically includes the following steps:

[0098] S210, selecting interference image pairs with higher coherence than the second coherence as interference image pairs of composite data, performing pairing and combination on the composite data, i.e., performing complex conjugate operation, to obtain a plurality of interference image pairs of the composite data.

[0099] S220. Based on the Minimum Spanning Tree (MST) method, a basic interferometric network is constructed to generate an interferometric image pair connection diagram of all SAR images to be processed.

[0100] S230, adding several interference image pairs of the composite data to the basic interference network to obtain an optimal interference network, and generating an optimal interference image pair connection diagram that balances calculation accuracy and calculation efficiency.

[0101] S240 , generating a differential interference fringe pattern according to a plurality of interference image pairs of the composite data.

[0102] In step S210, by analyzing the coherence of typical interferometric image pairs, the overall coherence of the area to be processed, i.e., the SAR image to be processed, and the temporal variation pattern of the coherence of the seawall are determined. Combined with the spatial baseline decorrelation evaluation function, the coherence of all interferometric image pairs is evaluated, and this is used as the connection weight. In step S230, the interferometric image pairs with high coherence obtained in step S210 are added to the basic interferometric network to form the final optimal interferometric network, which can reduce redundant calculations and ensure the accuracy of wetland hydrological information extraction.

[0103] In step S200, an interference image pair selection method is established to identify the optimal interference network formed on the time-space baseline plane. The basis for constructing the optimal interference network is the MST algorithm and the second coherence (coherence coefficient threshold). In this embodiment, the theoretical coherence coefficient is estimated by formula (4), which is:

[0104]

[0105] Among them, γ spatial represents the spatial component, γ temporal represents the time component, γ doppler represents the Doppler component, γ noise represents the thermal noise component, B perp represents the spatial centroid baseline, Δt represents the time centroid baseline, Δfdc represents the Doppler centroid baseline, B perp_c is the critical baseline corresponding to zero coherence. fdc of ALOS PALSAR (PALSARA is an L-band SAR sensor carried by ALOS micro) in FBD mode c is 1521Hz, T c is the time decay constant, which is 2500 days for wetland areas. The SNR (signal-to-noise ratio) of ALOS PALSAR in FBD mode is 6.95dB.

[0106] Preferably, the second coherence is 0.6, that is, according to formula (4), the interference phase with a theoretical coherence coefficient greater than 0.6 is added to the basic interference network constructed in the space-time plane.

[0107] The wetland water level measuring device provided by the present invention is described below. The wetland water level measuring device described below and the wetland water level measuring method described above can be referenced to each other.

[0108] Combine the following Figure 6 The wetland water level measuring device of the present invention is applied to data monitoring and evaluation of wetland areas, and the device comprises:

[0109] The registration module 100 is used to register and calibrate the SAR image to be processed to obtain the registered composite data, where the composite data is the amplitude image and the phase information corresponding to the amplitude image.

[0110] In this embodiment, in the registration module 100, the polarization mode of the SAR image to be processed adopts HH, and SLC data is obtained. Therefore, the composite data is a continuous L-band HH polarization (time series) SAR image, and the SAR image to be processed is an image of a wetland area.

[0111] Since the coastal wetland is located at the junction of land and sea, low-coherence surface targets such as the ocean and mudflats occupy a large area on the SAR image. In this embodiment, in order to ensure the accuracy of registration, the registration method is improved. Specifically, the registration module 100 specifically includes the following steps:

[0112] Assuming that the SAR image to be processed has M+1 scene SAR effects, one of the SAR images to be processed is taken as the main image. According to the orbit parameters, pulse repetition frequency, time information, external DEM and other information, the other SAR images to be processed (auxiliary images) are registered to the main image, and the pixels between the main and auxiliary images are registered. The gross error elimination method is used to ensure the accuracy of the registration and achieve sub-pixel registration. Among them, the main image, as the basis of registration, has stable scattering characteristics.

[0113] It is understandable that the specific registration accuracy can be set according to actual conditions.

[0114] The pairing and combining module 200 is used to perform pairing and combining, i.e., complex conjugate operation, on the composite data based on the coherence of the interference image pairs to obtain a plurality of interference image pairs of the composite data, and generate a differential interference fringe pattern based on the interference image pairs.

[0115] The scatterer extraction module 300 is used to determine the SHP of the differential interference pattern and extract the DS based on the SHP.

[0116] The adaptive filtering module 400 is used to perform adaptive terrain filtering on the differential interference pattern after the DS is extracted.

[0117] A phase unwrapping module 500 is used to perform phase unwrapping on the differential interference pattern to obtain an unwrapped differential interference pattern;

[0118] The water level measurement module 600 is used to obtain the water level change of the unwrapped differential interference pattern based on the SBAS technology.

[0119] The wetland water level measurement device of the present invention is different from the traditional SAR technology in that a scatterer extraction module 300 and an adaptive filtering module 400 are added. The scatterer extraction module 300 is used to further determine the scatterers that maintain a high coherence, thereby determining the stable scatterers. The adaptive filtering module 400 is used to eliminate the influence of external errors, thereby improving the coherence of the scatterers. By combining the scatterer extraction module 300 and the adaptive filtering module 400 with the traditional SAR technology, it is possible to determine the precise scatterer points and more accurately extract the water level information, thereby performing more accurate wetland water level measurements, obtaining more accurate wetland water level changes, and ensuring the accuracy of wetland hydrological information extraction.

[0120] Taking the reed wetland as an example, it is a typical DS. The echo signal of each resolution unit on the SAR image is the vector sum of the backscattering of all independent scattering targets in the resolution unit, showing obvious Gaussian distribution characteristics. The DS backscattering energy is relatively low compared to the PS, and usually occupies several adjacent pixels on the SAR image. The scattering characteristics of these adjacent pixels have the same probability distribution. The statistical HP with probability distribution is confirmed by performing a statistical test on the differential interference pattern. By performing a statistical analysis on the registered L-band HH polarization composite data, it is tested whether the multi-phase backscattering coefficient values ​​of two adjacent pixels belong to the same distribution in the statistical test, and the SHP around each pixel can be identified.

[0121] Combine the following Figure 7 Describing the wetland water level measurement device of the present invention, the scatterer extraction module 300 specifically includes:

[0122] The homogeneous pixel determination unit 310 is used to obtain the connectivity number of all pixels in the preset window corresponding to each pixel of the differential interference pattern. Specifically, the Anderson-Darling (AD) test method is used for testing, and pixels with a connectivity number higher than the preset number are regarded as SHPs in the preset window and retained; wherein the pixel is the center point of the preset window corresponding to it, and the probability distribution of the multi-phase backscattering coefficient values ​​of the SHP between two adjacent pixels in the statistical test is the same.

[0123] In this embodiment, the preset window is a 11×11 window, and the DS is determined by traversing all pixels through the preset window.

[0124] The removal unit 320 is used to remove the SHP that is not connected to the center point and determine the DS block corresponding to the center point to ensure the spatial connectivity of the DS block corresponding to the center point of the preset window.

[0125] The coherence determination unit 330 is used to obtain the coherence of the DS block and extract the distributed scatterers of the distributed scatterer block having a higher coherence than the first coherence.

[0126] Preferably, the first coherence is 0.3, that is, if the coherence of the entire DS block is less than 0.3, it is treated as a decoherent block rather than a DS block, and the decoherent block will not participate in subsequent steps.

[0127] Combine the following Figure 8 The wetland water level measuring device of the present invention is described. After the coherence determination unit 330, the device further includes:

[0128] The error removal unit 340 is used to remove the flat ground phase and the terrain phase from the DS to eliminate the influence of the DEM error and improve the coherence of the DS.

[0129] In the scatterer extraction module 300, a preset window centered on each pixel is first defined, and then a two-sample GoF test at a given significance level is applied between each pixel in each preset window, and all pixels that can be considered as SHPs are selected, and pixels that are not directly connected to the center point through other SHPs are discarded, and finally all SHPs connected to the center point are used for subsequent steps. The scatterer extraction module 300 determines the SHPs with the same scattering statistical distribution pixel by pixel based on the significance level and connectivity principle, and determines the DS with the connectivity number as a threshold.

[0130] Combine the following Fig. 9 The wetland water level measuring device of the present invention is described, and the water level measuring module 600 specifically includes:

[0131] The model building unit 610 is used to build a linear model through DS.

[0132] The relative water level change acquisition unit 620 is used to acquire the relative water level change of the linear model based on the L1 norm minimization algorithm, the L2 norm minimization algorithm and the optimal interference network.

[0133] Deformation parameter estimation has always been a hot spot and difficulty in long-term interferometry. Existing algorithms usually use L2 norm minimization algorithms, which often result in a large number of deformation estimation errors in non-urban areas. In the relative water level change acquisition unit 620, based on the SBAS technology, the full-resolution differential interference fringe phase image processed by adaptive terrain filtering is unwrapped to perform full-resolution interference phase estimation. On the basis of maintaining the interference phase of the point scatterer, the interference phase quality of the DS is improved, and a DS with a high coherence coefficient is selected based on the full-resolution differential interference fringe phase image. Specifically, the low-pass part of the deformation phase and the residual terrain error (residual phase) are estimated by the least squares algorithm (L2 norm minimization algorithm); for the differential interference fringe phase image with the residual terrain error (residual phase) removed, the improved simplex method is used to perform L1 norm minimization calculation to form a parameter error map detection and eliminate gross errors. On this basis, the L2 norm minimization algorithm is used to solve the deformation phase to improve the accuracy and stability of the deformation phase estimation; the SVD algorithm is used to solve the high-resolution nonlinear deformation part.

[0134] The relative water level sequence acquisition unit 630 is used to integrate the relative water level changes based on the L1 norm minimization algorithm to obtain the relative water level sequence.

[0135] The SVD algorithm can obtain the L2 norm solution of the L2 norm minimization algorithm. However, in coastal wetlands, there are often some non-correlated areas, such as open water, which separate the areas with high coherence. This often introduces phase unwrapping errors, that is, phase jumps between different areas. The L2 norm minimization algorithm often performs poorly in detecting these phase jumps in the unwrapped data. In contrast, the L1 norm minimization algorithm can provide a more robust phase inversion solution for the problem that phase unwrapping errors often occur and are difficult to detect in non-urban areas.

[0136] The linear offset acquisition unit 640 is used to compare the relative water level sequence with the reference water level observation data to acquire the linear offset of the absolute value of the water level.

[0137] The absolute water level sequence acquisition unit 650 is used to remove the linear offset in the relative water level sequence to obtain the absolute water level sequence.

[0138] It is important to note that only relative water level changes can be obtained from phase unwrapping, so the InSAR observations need to be calibrated with ground-based hydrological observations to obtain an absolute water level estimate. To estimate the absolute water level, the relative water level changes collected at adjacent times are first integrated to obtain a relative water level sequence. The relative sequence is then linked to the reference water level observations to obtain the absolute water level sequence. There is always a linear offset between the relative water level sequence generated by InSAR and the reference water level observations. By comparing the InSAR observed water level sequence with the water level observations, the linear offset can be estimated. After removing the linear offset, the absolute water level sequence inverted by InSAR is generated.

[0139] When pairing imaging pairs, the traditional method only uses the maximum time baseline and space baseline constraints, such as setting an appropriate time baseline and vertical baseline, but does not judge the coherence of the interference image pairs. Therefore, after all the interference image pairs are subjected to differential interference pattern generation, there will inevitably be some interference image pairs with low coherence. The differential interference pattern and unwrapping results generated by these interference image pairs are not ideal. Therefore, these interference image pairs need to be eliminated before subsequent processing can be carried out.

[0140] Combine the following Fig.10 Describing the wetland water level measurement device of the present invention, the pairing combination module 200 specifically includes:

[0141] The pairing and combining unit 210 is used to select interference image pairs with higher coherence than the second coherence as interference image pairs of composite data, and perform pairing and combining, i.e., complex conjugate operation, on the composite data to obtain a plurality of interference image pairs of the composite data.

[0142] The first construction unit 220 is used to construct a basic interference network based on the MST method to generate an interference image pair connection diagram of all SAR images to be processed.

[0143] The second construction unit 230 is used to add several interference image pairs of the composite data to the basic interference network to obtain the optimal interference network and generate an optimal interference image pair connection diagram with a balance between calculation accuracy and calculation efficiency.

[0144] The interference pattern generating unit 240 is used to generate a differential interference pattern according to a plurality of interference pattern pairs of the composite data.

[0145] In the pairing and combination unit 210, by analyzing the coherence of typical interference image pairs, the overall coherence of the area to be processed, i.e., the SAR image to be processed, and the law of how the coherence of the seawall changes over time are determined. Combined with the spatial baseline decorrelation evaluation function, the coherence of all interference image pairs is evaluated, and this is used as the connection weight. In the second construction unit 230, the interference image pairs with high coherence obtained by the pairing and combination unit 210 are added to the basic interference network to form a final optimal interference network, which can reduce redundant calculations and ensure the accuracy of wetland hydrological information extraction.

[0146] In the pairing combination module 200, an interference image pair selection method is established to identify the optimal interference network formed on the time-space baseline plane. The basis for constructing the optimal interference network is the MST algorithm and the second coherence (coherence coefficient threshold).

[0147] Preferably, the second coherence is 0.6, and the interference phase with a calculated theoretical coherence coefficient greater than 0.6 is added to the basic interference network constructed in the space-time plane.

[0148] Fig.11 An example of a physical structure diagram of an electronic device is shown in FIG. Fig.11 As shown, the electronic device may include: a processor 810, a communication interface 820, a memory 830 and a communication bus 840, wherein the processor 810, the communication interface 820 and the memory 830 communicate with each other through the communication bus 840. The processor 810 may call the logic instructions in the memory 830 to execute the wetland water level measurement method, which includes the following steps:

[0149] S100, registering and calibrating the SAR image to be processed to obtain registered composite data; wherein the composite data is an amplitude image and phase information corresponding to the amplitude image, the composite data is a time series SAR image, and the SAR image to be processed is an image of a wetland area;

[0150] S200, based on the coherence of the interference image pairs, the composite data are paired and combined to obtain a plurality of interference image pairs of the composite data, and a differential interference fringe pattern is generated according to the interference image pairs;

[0151] S300, determining homogeneous pixels of the differential interference fringe pattern, and extracting distributed scatterers based on the homogeneous pixels;

[0152] S400, performing adaptive terrain filtering on the differential interference fringe pattern after the distributed scatterer is extracted.

[0153] S500, performing phase unwrapping on the differential interference pattern to obtain an unwrapped differential interference pattern;

[0154] S600, based on the small baseline set technology, obtain the water level change of the unwrapped differential interference pattern.

[0155] In addition, the logic instructions in the above-mentioned memory 830 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or the part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0156] On the other hand, the present invention also provides a computer program product, the computer program product includes a computer program stored on a non-transitory computer-readable storage medium, the computer program includes program instructions, when the program instructions are executed by a computer, the computer can execute the wetland water level measurement method provided by the above methods, the method comprising the following steps:

[0157] S100, registering and calibrating the SAR image to be processed to obtain registered composite data; wherein the composite data is an amplitude image and phase information corresponding to the amplitude image, the composite data is a time series SAR image, and the SAR image to be processed is an image of a wetland area;

[0158] S200, based on the coherence of the interference image pairs, the composite data are paired and combined to obtain a plurality of interference image pairs of the composite data, and a differential interference fringe pattern is generated according to the interference image pairs;

[0159] S300, determining homogeneous pixels of the differential interference fringe pattern, and extracting distributed scatterers based on the homogeneous pixels;

[0160] S400, performing adaptive terrain filtering on the differential interference fringe pattern after the distributed scatterer is extracted.

[0161] S500, performing phase unwrapping on the differential interference pattern to obtain an unwrapped differential interference pattern;

[0162] S600, based on the small baseline set technology, obtain the water level change of the unwrapped differential interference pattern.

[0163] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is implemented when the computer program is executed by a processor to perform the wetland water level measurement method provided above, the method comprising the following steps:

[0164] S100, registering and calibrating the SAR image to be processed to obtain registered composite data; wherein the composite data is an amplitude image and phase information corresponding to the amplitude image, the composite data is a time series SAR image, and the SAR image to be processed is an image of a wetland area;

[0165] S200, based on the coherence of the interference image pairs, the composite data are paired and combined to obtain a plurality of interference image pairs of the composite data, and a differential interference fringe pattern is generated according to the interference image pairs;

[0166] S300, determining homogeneous pixels of the differential interference fringe pattern, and extracting distributed scatterers based on the homogeneous pixels;

[0167] S400, performing adaptive terrain filtering on the differential interference fringe pattern after the distributed scatterer is extracted.

[0168] S500, performing phase unwrapping on the differential interference pattern to obtain an unwrapped differential interference pattern;

[0169] S600, based on the small baseline set technology, obtain the water level change of the unwrapped differential interference pattern.

[0170] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0171] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0172] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for measuring wetland water level, comprising the following steps: The synthetic aperture radar image to be processed is registered and calibrated to obtain registered composite data; wherein the composite data is an amplitude image and phase information corresponding to the amplitude image, the composite data is a time series synthetic aperture radar image, and the synthetic aperture radar image to be processed is an image of a wetland area; Based on the coherence of the interference image pairs, the composite data are paired and combined to obtain a plurality of interference image pairs of the composite data, and a differential interference fringe pattern is generated according to the interference image pairs; Determining homogeneous pixels of the differential interference fringe pattern, and extracting distributed scatterers based on the homogeneous pixels; Performing adaptive terrain filtering on the differential interference fringe pattern after extracting the distributed scatterer; Performing phase unwrapping on the filtered differential interference fringe pattern to obtain an unwrapped differential interference fringe pattern; Based on the small baseline set technology, obtaining the water level change of the unwrapped differential interference pattern; The method of obtaining the water level change of the unwrapped differential interference pattern based on the small baseline set technology specifically includes the following steps: constructing a linear model through the distributed scatterers; Based on the L1 norm minimization algorithm, the L2 norm minimization algorithm and the optimal interference network, the relative water level change of the linear model is obtained; Based on the L1 norm minimization algorithm, the relative water level change is integrated to obtain a relative water level sequence; Comparing the relative water level sequence with the reference water level observation data to obtain a linear offset of the absolute value of the water level; The linear offset in the relative water level sequence is removed to obtain an absolute water level sequence.

2. The wetland water level measurement method according to claim 1, characterized in that: The determining of homogeneous pixels of the differential interference fringe pattern and extracting distributed scatterers based on the homogeneous pixels specifically includes the following steps: Obtaining the connectivity number of all pixels in the preset window corresponding to each pixel of the differential interference pattern, and taking the pixels with a connectivity number higher than the preset number as the homogeneous pixels in the preset window; wherein the pixel is the center point of the preset window corresponding to the pixel, and the probability distribution of the multi-phase backscattering coefficient values ​​of the homogeneous pixels between two adjacent pixels in the statistical test is the same; Removing the homogeneous pixels that are not connected to the center point, and determining a distributed scatterer block corresponding to the center point; The coherence of the distributed scatterer block is obtained, and the distributed scatterers of the distributed scatterer block having a higher coherence than a first coherence are extracted.

3. The wetland water level measurement method according to claim 2, characterized in that: After obtaining the coherence of the distributed scatterer block and extracting the distributed scatterers of the distributed scatterer block having a coherence higher than the first coherence, the method further comprises the following steps: The flat ground phase and the terrain phase are removed from the distributed scatterer.

4. The wetland water level measurement method according to claim 1, characterized in that: In the step of integrating the relative water level change based on the L1 norm minimization algorithm to obtain the relative water level sequence, terrain error, atmospheric noise and orbit error are removed; wherein the atmospheric noise is obtained by numerical weather forecasting.

5. The wetland water level measurement method according to claim 1, characterized in that: The method of pairing and combining the composite data based on the coherence of the interference image pairs to obtain a plurality of interference image pairs of the composite data, and generating a differential interference fringe pattern according to the interference image pairs specifically includes the following steps: Selecting the interference image pairs having higher coherence than the second coherence as the interference image pairs of the composite data, and performing pairing and combination on the composite data to obtain a plurality of the interference image pairs of the composite data; Based on the minimum spanning tree method, the basic interference network is constructed; Adding the plurality of interference image pairs of the composite data to the basic interference network to obtain the optimal interference network; A differential interference fringe pattern is generated according to the plurality of interference image pairs of the composite data.

6. The wetland water level measurement method according to claim 1, characterized in that: The process of registering and calibrating the synthetic aperture radar image to be processed to obtain the registered composite data specifically includes the following steps: One of the synthetic aperture radar images to be processed is used as a main image, and the other synthetic aperture radar images to be processed are registered to the main image; wherein the main image has a stable scattering characteristic.

7. A wetland water level measuring device, comprising: A registration module (100) is used to register and calibrate the synthetic aperture radar image to be processed to obtain registered composite data; wherein the composite data is an amplitude image and phase information corresponding to the amplitude image, the composite data is a time series synthetic aperture radar image, and the synthetic aperture radar image to be processed is an image of a wetland area; A pairing and combining module (200) is used to pair and combine the composite data based on the coherence of the interference image pairs to obtain a plurality of interference image pairs of the composite data, and to generate a differential interference fringe pattern based on the interference image pairs; A phase unwrapping module (500) is used to perform phase unwrapping on the differential interference fringe pattern to obtain an unwrapped differential interference fringe pattern; The water level measurement module (600) is used to obtain the water level change of the unwrapped differential interference pattern based on the small baseline set technology, characterized in that before the phase unwrapping module (500), the device further comprises: A scatterer extraction module (300), used to determine homogeneous pixels of the differential interference fringe pattern, and extract distributed scatterers based on the homogeneous pixels; An adaptive filtering module (400) is used to perform adaptive terrain filtering on the differential interference fringe pattern after extracting the distributed scatterer; The water level measurement module specifically includes: A model building unit, for building a linear model through distributed scatterers; A relative water level change acquisition unit, used for acquiring the relative water level change of the linear model based on an L1 norm minimization algorithm, an L2 norm minimization algorithm and an optimal interference network; A relative water level sequence acquisition unit is used to integrate the relative water level change based on an L1 norm minimization algorithm to obtain a relative water level sequence; A linear offset acquisition unit is used to compare the relative water level sequence with the reference water level observation data to obtain the linear offset of the absolute value of the water level; The absolute water level sequence acquisition unit is used to remove the linear offset in the relative water level sequence to obtain the absolute water level sequence.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the steps of the wetland water level measurement method according to any one of claims 1 to 6 are implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the wetland water level measurement method according to any one of claims 1 to 6 are implemented.

Images