CN117541929A

CN117541929A - Deformation risk assessment method for large-area power transmission channel of InSAR in complex environment

Info

Publication number: CN117541929A
Application number: CN202311242386.2A
Authority: CN
Inventors: 杨洋; 李孟; 赵蓂冠; 庄文兵; 李晓光; 王红霞; 赵普志; 张博; 杨柱石; 付豪; 王立福
Original assignee: Electric Power Research Institute of State Grid Xinjiang Electric Power Co Ltd
Current assignee: Electric Power Research Institute of State Grid Xinjiang Electric Power Co Ltd
Priority date: 2023-09-25
Filing date: 2023-09-25
Publication date: 2024-02-09

Abstract

The invention discloses a deformation risk assessment method for a large-area power transmission channel of an InSAR in a complex environment, which comprises the following steps: step 1, data collection and arrangement, including InSAR data source selection; collecting basic data; step 2, inSAR pretreatment and interference treatment, comprising the following steps: reading data; TOPS data segmentation; processing the Burst level data; burst splice and sub-band splice; interference treatment; differential interference processing; checking an interference pattern; ESD fine registration and interferogram correction; multi-view processing; step 3, inSAR deformation monitoring, which comprises the following steps: generating an optimal interference network connectivity graph; selecting a high coherence point; selecting a distributed target; constructing a CS/DS network and solving a secondary differential phase model; separating residual phases; and correcting elevation errors and solving deformation. The invention realizes millimeter-level precision earth surface deformation monitoring, comprehensively improves the geological risk monitoring quality effect of the transmission line, and has extremely high guarantee on data timeliness.

Description

Deformation risk assessment method for large-area power transmission channel of InSAR in complex environment

Technical Field

The invention relates to the technical field of geological deformation risk assessment of a power transmission corridor, in particular to a deformation risk assessment method of a large-area power transmission channel of an InSAR in a complex environment.

Background

The method has the advantages that the Xinjiang region is wide, coal resources are rich, extra-high voltage and high voltage transmission lines in the region are numerous, the transmission distance is long, and the lines pass through various natural severe-condition regions such as local sand storm regions, high temperature, high and cold, flood and the like, so that the monitoring and operation and maintenance work of the lines is abnormally heavy.

At present, the safety protection of the power transmission line in the Xinjiang area mainly adopts a video monitoring and manual inspection mode, excessively relies on personnel experience and precautions, lacks a coping method of a scientific system, cannot realize real-time comprehensive online monitoring of the power transmission line, cannot timely early warn and monitor sudden accidents, and can cause that the power transmission line faults cannot be discovered at the first time, the best opportunity for preventing and taking effective measures is lost, and economic loss and social influence caused by the power transmission line faults are enlarged. Due to the influence of external environments such as strong wind, flood, geological disasters and the like, and the reasons such as non-closing construction quality, uneven foundation, accidental collision and the like, the iron tower is very likely to be inclined and deformed, and great danger and hidden danger are brought to the safety of a power grid. At present, the safety condition of part of electric power iron towers is worry, and monitoring means such as manual inspection have the defects of long period, low efficiency and the like, and the iron tower inclination caused by an emergency cannot be timely detected, so that the situation is probably further expanded, and the power supply reliability of a power grid is influenced. At present, conventional monitoring means mainly based on manpower cannot meet the disaster prevention and reduction requirements of a power grid under the wide area energy Internet background, and a comprehensive monitoring and early warning means with high real-time performance, wide monitoring range, short period, high efficiency and low cost is urgently needed, so that power grid equipment and running state data are efficiently and reliably obtained, and the intelligent and informationized operation and detection work is converted into data intelligent driving.

The InSAR technology is utilized, long-time sequence InSAR satellite image data are adopted, large-scale general investigation of geological deformation risks of a power transmission corridor is carried out, and the installation of a Beidou monitoring terminal in a geological deformation risk area is guided; based on big dipper high accuracy location technique, shaft tower slope monitor terminal and geology deformation monitor terminal of development can realize 24 hours unmanned on duty monitoring shaft tower gesture monitoring to in passing through communication network transmission to monitoring system with monitoring data, implement monitoring shaft tower gesture, and for fortune inspection department carries out transmission line shaft tower maintenance and peripheral geology risk hidden danger elimination and provides data support, provides transmission line safety and reliability.

The synthetic aperture radar interferometry (synthetic aperture radar interferometry, inSAR) is used as an active space-to-ground microwave remote sensing technology, is less affected by cloud and rain and other conditions, can acquire ground elevation or deformation information in a large scale, low cost and high spatial resolution, is rapidly developed in the past 30 years, and is widely applied to a plurality of fields such as seismic deformation research, urban settlement monitoring, landslide monitoring, mine deformation monitoring, hydraulic engineering safety monitoring and the like. Many countries and organizations have transmitted or will transmit multiple SAR satellites in recent years, further driving the development of InSAR technology and applications.

However, inSAR also has inherent limitations, such as being easily affected by atmospheric delay errors, space-time coherence noise and the like, and seriously confusing real deformation signals, so as to cause misjudgment on InSAR deformation results; the obtained deformation result is the projection of the earth surface deformation on the line of sight (LOS) and the size and direction of the real deformation are difficult to accurately evaluate.

The InSAR technology utilizes two or more than one SLC images covering the same region at different moments, and obtains interference signals through conjugate multiplication, and is initially used for obtaining a landform map and three-dimensional surface topography data. The InSAR interference signal is related to deformation besides elevation, so that the InSAR interference signal can be used for acquiring tiny earth surface deformation information on the basis of removing the contribution of the terrain phase, and the technology is called differential InSAR (differential InSAR, D-InSAR) technology. The technology obtains the isoseism deformation field of Landers earthquake and brings wide attention to the world scholars. However, errors such as phase-loss interference noise and atmospheric delay seriously reduce the reliability of the D-InSAR technology monitoring, so that the application scene is mainly limited in monitoring areas with larger deformation magnitudes such as earthquake, geological structure movement and volcanic, and the accuracy can only reach the level of decimeters to centimeters generally.

In order to overcome the limitation of the traditional D-InSAR technology and improve the surface deformation fine monitoring capability of the InSAR technology, a series of time-series InSAR (TS-InSAR) technologies, also called multi-temporal InSAR (MT-InSAR), are developed by a plurality of scholars at home and abroad on the basis of the D-InSAR technology, and become hot spots for research in the InSAR field. The core idea of the technology is mainly to extract high-coherence points with stable scattering characteristics by utilizing differential winding or unwrapping interference atlas of the same area, and then reconstruct variable rate and deformation time sequence based on minimum norm theory. The current TS-InSAR technique can be divided into two major categories according to the number of primary images, most typically a single primary image permanent scatterer (persistent scatterer InSAR, PS-InSAR) technique and a multi-primary image small baseline set (smallbaseline subset, SBAS) technique. The PS-InSAR technology is mainly used for modeling the targets of strong-reflection permanent scatterers (persistent scatterer, PS) such as infrastructures, bare rocks and the like in natural environments, so that the technology is often applied to urban areas with denser artificial buildings with better effect. The multi-primary image SBAS technology is mainly aimed at modeling analysis of distributed targets (distributed scatterer, DS). The technology is mainly used for suppressing noise by using multi-view filtering in the early stage, and meanwhile, the data processing efficiency is improved, but the cost is that the spatial resolution is lost, so that the technology is mainly suitable for wide-area earth surface detection, such as seismic source mechanism and large fracture zone research. However, the multi-view SBAS technology is unfavorable for capturing deformation detail characteristics when facing complex scenes and fine monitoring targets. For this reason, SBAS technology based on full resolution (single view) is evolving. In particular, since the second generation of permanent scatterer squesar technology, distributed target InSAR (distributed scattererInSAR, DS-InSAR) technology has become one of the new research hotspots in the international time series InSAR field.

The above described time series InSAR technique uses only single polarization data for the surface deformation inversion. With the new generation of SAR satellites mostly having multi-polarization data acquisition capability, some scholars extend a new technology branch, namely polarization timing InSAR (PolPSI) technology, on the basis of the timing InSAR technology. The technical proposal further weakens incoherence caused by factors such as aliasing of various scattering mechanisms or orientation of structures in urban areas, and improves the density of deformation monitoring points. In addition to the above-mentioned widely-used InSAR techniques, some advanced InSAR techniques, such as pixel offset tracking (pixel offset tracking, POT) techniques, multi-aperture InSAR (MAI) techniques, SAR tomography (SAR tomograph) techniques, and whistle data azimuth sub-band overlap region interference (burst overlap interferometry, BOI) techniques, have been developed in recent years. It should be noted that in theory POT technology is based on amplitude data rather than interference phase information and therefore should not belong to the InSAR technology. Because of different technical principles, the monitoring precision is slightly different, so that the actual application needs to be determined according to the situation.

The current practice in international mainstream is to use a ground survey mapping method, and to perform site layout based on expert experience. The traditional method has great limitation, cannot be comprehensively examined, and is time-consuming and labor-consuming. The InSAR method can not only solve the problems, but also efficiently consult historical data and make overall analysis arrangement by combining the existing data.

At present, geological disaster monitoring of a power transmission line mainly utilizes means such as ground installation Beidou GNSS monitoring stations and the like, and high-precision mm-level real-time monitoring of geological changes and disasters can be realized. However, in practical applications, there are mainly the following problems: firstly, the ground monitoring equipment can only realize punctiform monitoring, has high hardware cost and cannot be applied in a large range; secondly, ground monitoring equipment can only monitor the current geological state and cannot evaluate the overall long-period change trend of geology; thirdly, the monitoring precision is limited in a complex environment, and the monitoring effect of geological disasters is seriously affected.

Disclosure of Invention

The invention aims to provide a deformation risk assessment method for a large-area power transmission channel of a complex environment InSAR, which is used for solving the problem that large-scale, high-precision and long-time sequence monitoring deformation cannot be carried out in complex terrains, weather and climates.

In order to achieve the above purpose, the deformation risk assessment method for the large-area power transmission channel of the InSAR in the complex environment, disclosed by the invention, utilizes long-time sequence SAR satellite images to carry out large-scale geological deformation monitoring on a power transmission corridor, realizes general investigation and monitoring of geological deformation settlement risk along the power transmission corridor, and comprises the following steps:

Step 1, data collection and arrangement, including (1) InSAR data source selection; (2) basic data collection;

step 2, inSAR pretreatment and interference treatment, comprising the following steps: (1) data reading; (2) TOPS data segmentation (TOPS split); (3) Burst level data processing; (4) Burst splice and sub-band splice; (5) interference treatment; (6) differential interference processing; (7) interferogram inspection; (8) ESD fine registration and interferogram correction; (9) multiview processing;

step 3, inSAR deformation monitoring, which comprises the following steps: generating an optimal interference network connectivity graph; (2) high coherence point selection; (3) distributed object selection; (4) Constructing a CS/DS network and solving a secondary differential phase model; (5) residual phase separation; and (6) correcting elevation errors and solving deformation.

Further, the InSAR satellite data source selection criteria in the InSAR data source selection in the step 1 comprise the steps of obtaining a high-quality interferogram, performing differential interference processing, and ensuring the consistency of an observation target within a time interval and avoiding a phase loss interference phenomenon, wherein the SLC image pair is required to meet the requirement of a critical base line distance in space; in the data selection process, the type of a data source, satellite orbit, spatial resolution, time interval and economy are considered; the basic data collection means for collecting basic geographic information data of a research area, and comprises the steps of collecting DLG, DOM, DEM data of 3D of the research area, and processing and analyzing InSAR ground subsidence; and collecting field measurement data of a research area, and laying a data foundation for InSAR ground subsidence monitoring work.

Further, the step 2 of data reading includes converting SAR data into a standard format, wherein the original data format of Sentinel-1 is a tiff file, converting the tiff file into a binary file, and simultaneously obtaining corresponding parameter files, wherein 3 subband image data are three independent files, respectively performing data conversion and parameter extraction, and then updating track parameter information by using a fine track data file; the TOPS data segmentation comprises the independent processing of each burst data, wherein the data segmentation is to separate independent burst data files from each sub-band data; the Burst level data processing comprises the steps of compensating a linear frequency modulation signal introduced by Doppler frequency change, namely, performing declassification processing before the registration process is performed; after registration processing, compensating the removed deramp phase again, namely, reramp operation, and then carrying out subsequent interference processing; and estimating the offset between the main image and the auxiliary image by means of external DEM data and fine track information by using a geometric registration method, then carrying out interpolation resampling, and registering the auxiliary image under a main image reference system to obtain the registration precision required by coarse registration.

Further, the Burst stitching and sub-band stitching in the step 2 includes stitching registered Burst data, removing the overlapping area and the black edge of the invalid value, and generating complete sub-band image data; the overlapping area needs to be reserved additionally for later checking whether jump errors exist between adjacent bursts of the interferograms; after the Burst data of each sub-band are spliced, three independent sub-band data are spliced, overlapping areas among the sub-bands are removed, and complete SLC image data are generated; the interference processing comprises the steps of carrying out interference processing on the registered SLC images, carrying out conjugate calculation on the complex images, and simultaneously obtaining corresponding interference parameter information; the differential interference processing comprises that interference fringes caused by land leveling phases and terrain phases mask deformation information, land leveling and simulated terrain phases are generated by using DEM and track parameters, and then the land leveling and simulated terrain phases are subtracted from an interference pattern, so that a differential interference pattern is finally obtained.

Further, in the step 2, the interferogram detection includes that in order to make the registration accuracy meet the requirement of the registration accuracy of the TOPS mode, phase jump existing in the generated interferogram is eliminated, a manual discrimination method is adopted, or interference processing is performed by using overlapping areas among bursts, and phase difference of the overlapping areas is calculated to judge; the ESD fine registration and interferogram correction comprises the steps of obtaining interference pair combinations with registration errors through interferogram inspection, carrying out fine registration processing on the interferograms by using an ESD registration method, converting the relative offset of the interference pairs into the offset relative to a main image, correcting the auxiliary image, regenerating a correct interferogram, converting the offset into an offset phase corresponding to an SAR image, and directly compensating the interferogram; the multi-view processing comprises multi-view filtering of a group of differential interferograms which are finally generated, noise influence is restrained, part of data quantity is reduced, and large-range InSAR data processing workload is reduced.

Further, the generating of the optimal interference network connectivity map in the step 3 includes combining the N SAR images of the research area into an ideal small baseline set according to the space-time baseline threshold value to form the optimal interference network connectivity map; in order to reduce the space-time decoherence influence caused by long baselines, in the interference pair screening process, a group of interference pairs is initially determined by utilizing a small baseline threshold value, then the initially selected interference combination is further screened, and the optimal interference pair combination is determined by adopting a method based on coherence estimation.

Further, the selecting of the high coherence point in the step 3 includes selecting a ground target with stable scattering characteristics as a high coherence target, firstly starting from the identification of the high coherence point (CS), generating an average coherence coefficient map according to the previous step, and screening according to a certain threshold; the distributed target selection comprises the steps of carrying out homogeneous pixel inspection on a non-high coherence area by utilizing an AD inspection method, selecting a distributed target, estimating a coherence matrix of the non-high coherence area according to the inspected homogeneous pixel area, estimating an optimal phase sequence of the coherence matrix by utilizing a maximum likelihood estimation method, calculating interference combinations of the distributed targets according to phase trigonometry and the optimal phase sequence after all the distributed targets are obtained, and finally fusing the distributed target high coherence targets to form a CT/DS phase set.

Further, the constructing of the CS/DS network and the solving of the secondary differential phase model in the step 3 comprise the steps of constructing a network for CS/DS points after a CS/DS initial selection point set of high coherence points is obtained, estimating deformation parameters by using a differential measurement method, and improving the resolving precision and stability by adopting a multi-level network, wherein the primary network adopts a global triangular network, and the overall resolving error can be reduced by using a closed network adjustment; the second hierarchical network is added, and an irregular closed network formed by local reference points is used as a group of additional global constraint conditions, so that error transfer in the network resolving process can be greatly reduced; and carrying out secondary difference on adjacent CS/DS phases, estimating to obtain differential measurement parameters, and then converting the relative quantity of the differential measurement parameters into absolute quantity of global parameters by using a network adjustment resolving method, thereby obtaining a linear deformation rate diagram and a residual elevation diagram.

Further, the residual phase separation in the step 3 includes separating each noise phase component by space-time filtering using the time-space domain characteristics of the residual phase, and extracting the atmospheric delay phase, and the residual phase separation can be further divided into three sub-steps of residual phase unwrapping, residual phase separation and atmospheric phase estimation.

Further, the correction of the elevation error and the solution of the deformation amount in the step 3 include correcting the original phase, taking into account the influence of the initial values of the linear velocity and the elevation error on iteration, filtering the residual phase after obtaining the initial values of the linear velocity and the elevation error, using the phase difference between the residual phase and the filtered phase after the filtering is completed, regarding the phase difference as the influence of the atmosphere and the noise, regarding the rest as the nonlinear deformation amount, adding the nonlinear deformation amount into the linear deformation amount, and obtaining the real deformation amount; and establishing an observation equation on the selected coherent target, solving by adopting a SVD singular value decomposition method to obtain the sedimentation rate of each time interval, and integrating the sedimentation rate of each time interval in a time domain to obtain the accumulated deformation quantity of each time point.

The method of the invention has the following advantages:

The InSAR satellite technology has the advantages of all-day, all weather, high precision, large range, long period, low cost and the like, and can be applied to monitoring of earth surface deformation and settlement of a power transmission corridor: (1) all weather and earth observation capability all the day, and is not affected by weather; (2) the monitoring precision is high, and millimeter-level deformation information can be measured; (3) the monitoring range is wide, and hundreds of square kilometers can be monitored at a time; (4) the monitoring density is high, and more than 1 ten thousand observation point data can be obtained in each square kilometer of urban areas; (5) the repetition frequency is high, the continuous monitoring capability is strong, macroscopic static information can be provided, and quantitative dynamic information can be given out; (6) the cost is low, a monitoring network does not need to be established, some potential or unknown target deformation information can be identified, and deformation conditions of a plurality of years can be provided.

By utilizing the characteristics of InSAR all-day, all weather, high precision, large range, long period, low cost and the like, the geological deformation monitoring and processing method of the InSAR technology under the complex natural earth surface condition is adopted to realize the earth surface deformation monitoring with millimeter-level precision; and carrying out large-scale general investigation on geological deformation risks of the transmission corridor by utilizing the InSAR satellite long-time sequence image data, guiding the installation of Beidou monitoring terminals in geological deformation risk areas, constructing a 'heaven-earth' three-dimensional monitoring system, and comprehensively improving geological risk monitoring quality and efficiency of the transmission line. The timeliness of the data is extremely high, and the settlement/change amplitude of the past stage can be obtained within 15 days after the acquisition of the remote sensing data (the time is changed to the acquisition of the satellite precise orbit). This is not comparable to the conventional method.

Drawings

FIG. 1 is an InSAR elevation measurement schematic diagram;

FIG. 2 is a schematic diagram of a two-rail process flow;

FIG. 3 is a schematic diagram of a three-rail method;

FIG. 4 is a three-rail process flow diagram;

FIG. 5 is a schematic diagram of a line tower tilt monitoring;

FIG. 6 is a roadmap of InSAR earth surface deformation monitoring technology;

FIG. 7 is a flowchart of a Sentinel-1TOPS data interferometry process;

FIG. 8 is a flowchart of an ESD registration algorithm;

FIG. 9 is a DS Point selection diagram;

FIG. 10 is a plot of CS-Point and DS-Point fusion.

Detailed Description

The technical solution of the present invention will be clearly and completely described in conjunction with the specific embodiments, but it should be understood by those skilled in the art that the embodiments described below are only for illustrating the present invention and should not be construed as limiting the scope of the present invention. All other embodiments, which can be made by one of ordinary skill in the art without undue burden on the person of ordinary skill in the art based on the embodiments of the present invention, are within the scope of the present invention.

The method is characterized in that firstly, based on the advantages of a time sequence InSAR technology, such as a large range, full-day time, all-weather, high precision, long period, planar monitoring and the like, the SAR image covering the power transmission line area is subjected to time sequence differential interference treatment, and the information of the space position, the deformation range, the deformation rate, the accumulated deformation and the like of the deformation abnormal area in the area is extracted. And positioning the abnormal region through spatial analysis with the power transmission line, and determining the deformed abnormal region in the covered power transmission line region based on the deformation information and the spatial analysis result. And comprehensively evaluating deformation risk factors of the large-area power transmission channel by adopting a wavelet analysis method in combination with influence factors such as geology, weather, groundwater and the like under the complex environmental condition, judging whether the change characteristics of the deformation factors are harmful or not, and taking measures to deal with abnormality, so that effective evaluation and predictive analysis of the deformation risk of the large-area power transmission channel under the complex environment are realized.

MEMS sensing technology

MEMS sensors, i.e. microelectromechanical systems (Microelectro Mechanical Systems), are the leading-edge research area of multi-disciplinary intersection developed on the basis of microelectronics. Over forty years of development, it has become one of the major technological areas of worldwide attention. The method relates to various disciplines and technologies such as electronics, machinery, materials, physics, chemistry, biology, medicine and the like, and has wide application prospect. By 2010, about 600 units of MEMS have been developed and produced worldwide, and hundreds of products including miniature pressure sensors, acceleration sensors, micro-inkjet printheads, digital micromirror displays have been developed, with MEMS sensors accounting for a significant proportion. MEMS sensors are novel sensors fabricated using microelectronics and micromachining techniques. Compared with the traditional sensor, the sensor has the characteristics of small volume, light weight, low cost, low power consumption, high reliability, suitability for mass production, easy integration and realization of intelligence. At the same time, feature sizes on the order of microns allow it to perform functions not possible with some conventional mechanical sensors.

The silicon micro-acceleration sensor is a micro-mechanical sensor that is the second to enter the market after the micro-pressure sensor. The main types are piezoresistive, capacitive, force balanced and resonant. Of these, the most attractive is the force balance accelerometer, a typical product of which is the AGXL50 type reported by Kuehnel et al in 1994.

Angular velocity is typically measured using a gyroscope. Conventional gyroscopes measure angular velocity using the characteristic that a high-speed rotating object has to maintain its angular momentum. The gyroscope has high precision, but has complex structure, short service life and high cost, is generally only used in navigation, and is difficult to apply in a general motion control system. In fact, the angular velocity sensor may find wide application in fields such as automotive traction control systems, camera stabilization systems, medical instruments, military instruments, sports machinery, computer inertial mice, military, etc., if not cost-limited. Common micromechanical sensors of angular velocity have a double gimbal structure, a cantilever structure, a tuning fork structure, a vibrating ring structure, etc. However, the precision of the micromechanical gyroscope realized is still less than 10 °/h, which is far from the 0.1 °/h required for inertial navigation systems.

Satellite remote sensing technology

Each pixel in the image acquired by the spaceborne SAR system contains both radar back-scatter intensity information for the ground resolution element and phase information related to the slant range (distance from the radar platform to the imaging point). Subtracting the phase values of corresponding pixels of two radar images covering the same area can obtain a phase difference diagram, namely an interference phase diagram (interference pattern). The phase difference information is the contribution sum of the factors such as topography fluctuation, surface deformation (if present) and the like. InSAR is the fact that the interference phase signals with high sensitivity are used for extracting and separating useful information (such as ground surface elevation or ground surface deformation), and is quite different from photogrammetry and visible light and near infrared remote sensing which mainly use image gray level information for reconstructing three-dimensional or extracting information.

(1) Interference phase signal

The SAR echo signal of the ground target not only comprises amplitude information A, but also comprises phase information phi, and the backscattering information of each pixel on the SAR image can be expressed as complex Ae ^iφ . The phase information includes distance information of the SAR system from the target and scattering characteristics of the surface target, namely:

in the formula 3.1, 4pi is a two-way distance phase; r is the slant distance between SAR and the target; phi (phi) _obj Is the scattering phase of the ground target.

The images when the ground target point P is imaged twice are respectively as follows:

wherein, c ₁ C is the main image ₂ Is a secondary image. And has the following steps:

the complex interferogram can be obtained by conjugate multiplication of the main image and the auxiliary image, and the complex interferogram is as follows:

wherein, represents taking the conjugate. Is provided withThe interference phase is:

if the scattering properties of the ground target are unchanged during the two imaging, i.e. φ ₁ ＝φ ₂ If the skew is poor, Δr=r ₁ -R ₂ The phase of the interferogram is related only to the path difference of the two observations, namely:

here, theIs the true interference phase. The whole number of turns of the phase obtained in the actual process is unknown, that is, the winding phase, and the winding phase must be unwound in order to obtain the true phase.

Further decomposing the interference phase to obtain:

in the middle ofThe interference phases caused by the earth shape, topography relief, surface deformation, atmosphere and noise are shown, respectively.

(2) InSAR elevation measurement

The geometrical relationship of the repeated orbit InSAR observation is generally shown in fig. 1. S1 and S2 respectively represent a main image sensor and an auxiliary image sensor, B is a base line distance, alpha is a base line distance and a horizontal direction inclination angle, theta is a main image incidence angle, H is the height of the main sensor relative to the ground, R1 and R2 respectively represent a main image inclination distance, P is a ground target point, and the heights are H and P ₀ Is the equidistant point of P on the reference flat ground. For ease of discussion, it is assumed that there is no surface deformation during the master-slave relative acquisition and no atmospheric effects.

Decomposing the base line along the incident direction and perpendicular to the incident direction to obtain a perpendicular base line slant distance B _⊥ And parallel baseline skew B _// :

B _⊥ ＝Bcos(θ-α),B _// =Bsin(θ-α) 3.8

In the far field case, it can be assumed that Δr=b _// The formula can be expressed as:

under the condition that the reference plane is flat, according to the triangular relation, there are

h=H-R ₁ cosθ 3.10

Differential the two sides of the formulas 3.9 and 3.10 respectively, comprising

Substituting the following formula 3.11 into the above formula:

wherein, the left side represents the interference phase difference of adjacent pixels; the first right term indicates the phase caused by the change in elevation of the target, and the second right term indicates the phase caused by the level ground without the change in elevation, which is called the level ground phase. To invert the elevation, the land phase needs to be removed and the relationship between the interference phase and the elevation is directly established.

After the flat ground phase is removed, the direct relation between the elevation and the phase can be obtained, namely

Wherein θ is ₀ =θ - Δθ, representing the equidistant point P on flat ground ₀ Is a main image incident angle of (c). B. Alpha and H can be deduced from the orbit attitude data, and R ₁ The method can be calculated according to related radar parameters in the SAR image header file.

If the reference ellipsoids and the spheres are selected as the reference surfaces, the interference phases after the ground effect removal under different reference surfaces can be respectively obtained as follows:

examination ellipsoid model 3.14

Sphere model 3.15

Wherein H is the satellite platform height; r is (r) _H 、r _h The radius of the earth at the position of the satellite lower point and the target point respectively; r is the slant distance.

(3) InSAR earth surface deformation measurement

Satellite InSAR systems have found wide application in earth's surface deformation detection. To separate out the deformation information, the shape of the earth and topography factors that have a significant impact must be removed from the initial interferometric phase, thus a differential interferometry (DInSAR) method is available. In 1989 Gabriel, the concept of differential interferometry was first introduced, which is a measurement technique that uses two interference images of the same region, one of which is an interference image before deformation and the other of which is an interference image acquired after deformation, and then acquires the surface deformation by differential processing. The traditional DINSAR method mainly comprises a two-rail method, a three-rail method and a four-rail method. For ease of calculation, the following discussion does not consider atmospheric and noise effects.

1) Two-rail method

The basic idea of the two-rail method is to generate an interference pattern by using two images before and after the surface change of the experimental area, and remove the topographic information from the interference pattern to obtain the surface deformation information. The method has the advantages that phase unwrapping of the interference pattern is not needed, and unwrapping difficulty is avoided. The disadvantage is that the method cannot be adopted for areas without DEM data; at the same time of causing the DEM data, new errors may be caused, such as elevation errors of the DEM itself, registration errors of the DEM simulation interference phase and the real SAR pattern, and the like.

The two-rail process flow diagram is shown in fig. 2.

Obtained from formula 3.7:

wherein:the interference phases caused by the earth shape and the topography fluctuation are shown.

The slope distance variable quantity reflecting the surface deformation can be obtained by the following calculation:

2) Three-rail method

The three-rail method is based on the principle that three-view images are utilized to generate two interference patterns, one reflecting the topographic information and the other reflecting the topographic deformation information. The three-rail method has the main advantages that auxiliary DEM data are not needed, the method is particularly important for monitoring the change of some non-topographic data, and the registration between the data is easy to realize; the disadvantage is that the quality of the phase unwrapping will affect the final result.

FIG. 3 is a diagram of a geometric model of a three-rail measurement, where S1 and S2 are the locations where the SAR system images the same region twice without terrain displacement, and the resulting interferometric phase contains only terrain information; s3 is the observation position of the SAR system after the surface deformation. The interference phases obtained by S1 and S3 contain not only the terrain phase but also the phase contribution of the surface deformation.

The two interference phases are respectively

In phi ₁₂ Only contains topographic information; phi (phi) ₁₃ Contains topographic information and deformation information; b (B) _// 、B _/ ' _/ Respectively S ₁ S ₂ And S is ₁ S ₃ θ is the image viewing angle; alpha _1, α ₂ The angles between the base line B, B' and the horizontal direction are respectively; Δd is the deformation displacement of the earth's surface in the satellite line of sight LOS direction. Thus the phase phi caused by displacement of the earth's surface in the LOS direction _d The method comprises the following steps:

the surface displacement deformation is expressed as:

the three-rail process flow is shown in fig. 4.

The four-rail method is similar to the three-rail method, except that the topography interferograms and the deformation interferograms are mutually independent.

(4) InSAR data processing

Based on digital signal processing technology, the data processing process of InSAR can be highly automated to extract surface three-dimensional information and surface deformation results. Before the interferometric data processing is performed, the appropriate interferometric pair and other ancillary data (e.g., external DEM for removal of the topography phase) must be selected. The selection criteria for the interference pair are: the interference baseline can be neither too long nor too short for DEM generation; for deformation monitoring, the shorter the interference baseline, the better. After the valid interferometric data sets are obtained, they are subjected to the necessary processing steps including SAR image registration, interferogram generation, reference plane/topography influence removal, geometric transformations (commonly referred to as geocoding), phase unwrapping, etc.

1) Image registration

The primary problem with extracting topography relief or surface deformation information from multi-phase sarlc images is the accurate registration of images taken along repeated tracks (with slight offset) covering the same region. The registration of the SAR image is to calculate the image coordinate mapping relation between the reference image (main image) and the image to be registered (auxiliary image), and then use the relation to perform coordinate transformation and resampling on the image to be registered. Because the track offset is small (typically around 1 km), while the track height is hundreds of kilometers. Therefore, in the overlapping region of the repeated track images, the coordinate offset between the pairs of identical image points has a certain change rule, and a high-order polynomial can be generally used for fitting.

The required image registration accuracy must reach the sub-pel level. Typically, this is performed in two stages, coarse and fine registration. The rough registration can calculate rough offset of the image to be registered relative to the reference image in azimuth and oblique directions by utilizing satellite orbit data or selecting a small amount of characteristic points, and aims to provide initial values for searching homonymous pixels in the accurate registration of the image. The fine registration is firstly based on rough image offset and image matching algorithm, a sufficient number of homonym point pairs uniformly distributed in an overlapping area are searched from the master image and the slave image, and then a polynomial model is used for describing pixel coordinate deviation of two images, namely, the coordinate difference of the homonym point pairs of the master image and the slave image can be expressed as a function expression of the coordinates of the master image. Based on the obtained coordinate offset observables of the same-name image points and a least square algorithm, polynomial model parameters can be solved, so that the establishment of the coordinate transformation relation of the images is completed. And finally, resampling the image to be registered by using the model to sample the slave image into the space of the master image.

2) Interferogram generation

The phase difference map can be easily obtained by subtracting the phases of the pixels corresponding to the resampled sub-image from the main image. In the actual calculation process, the principal and subordinate images are first complex conjugate multiplied, and the mathematical expression is

I(r，t)＝M(r，t)·S(r，t)*

Where M (r, t) and S (r, t) represent complex values of corresponding pixels of the master-slave image, respectively, and I (r, t) represents complex conjugates, and I (r, t) represents the generated interference information, which is also complex. The result is called a complex form of interferogram. Then extracting a phase principal value component diagram from the interference diagram to obtain a phase difference diagram, wherein the interference phase is changed from-p to +p, and a complete change appears as an interference fringe, but the phase integer ambiguity problem exists on each pixel.

3) Reference surface/topography influence removal

The primary differential interference phase map is a harmonic reflection of a variety of factors such as reference trend surface, topography relief, surface displacement, noise, and the like. For topography measurement, interference pairs not including deformation information are generally selected in advance according to prior information to be processed so as to avoid unnecessary trouble, and therefore, the direct phase difference score mainly includes contributions of a reference plane (generally selected as a reference ellipsoid) and topography undulation, and in order to facilitate subsequent phase unwrapping, the phase component of the ellipsoidal reference plane is generally removed from the direct differential phase. Notably, the contribution of the reference ellipsoid is dominant relative to the contribution of the terrain, which is why the primary differential interference phase map appears to appear as a stripe approximately parallel to the track, the longer the effective interference baseline, the denser the interference stripe, the greater the slope of the terrain, the denser the interference stripe, the more complex the terrain, and the more pronounced the change in stripe curvature. When we remove the contribution of the reference surface, the topographic phase stripe appears clearly, which shows a shape consistent with the shape of the topographic contour.

4) Phase unwrapping

In order to obtain the earth elevation or the earth displacement in the radar range direction, we have to determine the phase difference integer number of each pixel in the interferometric phase map, similar to the integer ambiguity determination problem in GPS, called phase unwrapping in InSAR is a key algorithm in interferometric data processing. At present, phase unwrapping algorithms are more, but mainly fall into two categories: (1) an integration method based on path control; (2) least squares based overall solution algorithms. The idea of the integration method is as follows: for each pixel of the winding phase diagram, first-order difference along the row direction and the column direction is obtained, and then the first-order difference is continuously integrated to obtain the unwrapping phase. Since the interferogram has singular points (called points of retention in the complex function), the integration path should be constrained from error propagation of the local interferential phase, so the key of this algorithm is to locate the singular points on a principle and connect them as "firewalls" of the integration path, i.e. they cannot traverse these paths when integrating. The idea of the least squares algorithm is: the method is characterized in that the method is used for integrally solving the phase gradient after unwrapping and the unwrapping in the sense that the sum of squares of differences between the phase gradient and the phase gradient after unwrapping is minimum, and the influence of a singular phase on an unwrapping result can be weakened by using a weighted estimation method.

(5) InSAR error propagation

The primary differential interference phase data and satellite orbit data are utilized to carry out three-dimensional reconstruction of the earth surface; and the surface deformation detection can be performed by utilizing the secondary differential interference treatment. These interferometry analyses require the use of radar system parameters, radar platform attitude (baseline) data, phase observations, and terrain data (subtraction of terrain phases in secondary differencing), etc., and it is apparent that uncertainties or errors in these data can propagate into the interferometric elevation or deformation results. Based on mathematical statistics and measurement error basic theory, the characteristics of several main error sources (i.e., phase observables, baseline data, and topography data) in a satellite radar interferometry system and their effects on elevation and deformation measurements are briefly described herein.

1) Interference phase error

The phase observance in SAR images is the most critical data source in the interferometric processing. Combining two SAR images acquired along different trajectories, the interferometric processing can extract a phase difference map (interferometric phase map) of the corresponding pixel, the interferometric phase of each pixel comprising the following contributions: (1) topography relief, (2) earth displacement projected into the radar line of sight, (3) possible atmospheric effects, (4) noise. The first three show some spatial autocorrelation, and understanding of interference phase noise requires a discussion to be made from the composition of the phase signal in a single SAR image.

When radar imaging, microwave signals emitted by the antenna pass through the atmosphere and interact with the ground surface to be reflected back and recorded by the sensor. For the phase of each pixel of a single SAR image, there are three main contributions: (1) the linear path length of the sensor to the surface resolution element, (2) the path curvature caused by the non-uniform atmospheric medium, and (3) the backscatter phase introduced by interaction of the microwave signal with objects within the surface resolution element. The scattering add-on phase is mainly related to two factors, firstly, random disturbance (such as vegetation growth or wind swing) or chemical characteristic change (such as ionization constant change related to soil humidity) can occur in the ground resolution element, and secondly, for the same resolution element, the track interval (or referred to as space baseline) can cause different radar side view angles and different scattering characteristics. For two SAR images acquired at different times, the respective random additional phase components (noise) are different or uncorrelated, are difficult to cancel out when the phase differences are time-sharing, can cause the interferograms to have no obvious fringes or fringe discontinuities, and the phase integer ambiguity resolution would be very difficult, while the variable atmospheric conditions (air pressure, temperature and relative humidity) may lead to different phase delays, which inconsistencies are manifested both on a time scale and on a spatial scale.

In general, the larger the time interval between acquisition of two SAR images along a repeating trajectory, the more severe the noise of the interference phase, the so-called time-decorrelation, which leads to failure of elevation and deformation measurements (especially vegetation coverage), especially to the monitoring of long-term accumulated deformations, such as pre-and post-earthquake deformations, volcanic movements, which becomes very difficult. Previous studies have shown that long band SAR systems (e.g., JERS-1L band SAR system, 23.5cm wavelength) are more advantageous than short band SAR systems (e.g., ERS-1/2C band SAR system, 5.7cm wavelength) in maintaining time correlation. The larger the orbit space interval between the two SAR images is acquired, the higher the interference phase noise level will be, i.e. the spatial decorrelation, which limits the available number of effective interference pairs. It is also quite difficult to completely subtract the effects of the atmosphere from the interference results due to the lack of high resolution ground meteorological data synchronized with SAR imaging time (the atmospheric delay solution of a sparse GPS permanent tracking station can be used to remove the atmospheric low frequency components).

2) Baseline error

In order to extract surface deformation information from the interferometric phase, we must use baseline parameters to subtract the contributions of the reference trend surface phase and the topography phase. Furthermore, the baseline parameter also needs to be used when calculating the earth elevation from the interferometric phase. Clearly, the baseline parameters play an indispensable role in InSAR data processing and analysis.

While the baseline parameters required for radar interferometry can be calculated directly from satellite orbit data, the accuracy of orbit data determined by existing orbit determination techniques is limited, and thus the baseline parameters contain errors and propagate into the interferometry results.

3) Influence of topographic data errors on deformation observations

In the case of a slight shift in the satellite repetition orbit, i.e., an interference baseline length that is not zero, radar differential interference requires the subtraction of the terrain phase contribution from the "terrain-deformation" interferogram (generated from the two SAR images spanning the deformation period) using terrain data (such as a digital ground elevation model or terrain interference phase). The uncertainty of the terrain data has a direct effect on the error in the deformation results proportional to the baseline length of the "terrain-deformation" interference pair. In order to achieve the required deformation measurement accuracy, the long baseline interference is higher than the short baseline interference in terms of accuracy of the topographic data.

Regardless of whether the terrain errors are random or systematic, the overall appearance is that the magnitude of the deformation errors increases as the magnitude of the terrain errors increases, but they have different negative effects on the differential interference signal analysis. The random topography errors can lead to significant reduction of the correlation coefficient of the differential interferogram, the reduction of the correlation coefficient is almost in linear relation with the noise level of the topography elevation, the higher the topography random error noise level is, the more serious the decorrelation is, and the topography decorrelation further increases the difficulty of processing interference data such as phase unwrapping.

Systematic topography errors hardly reduce interference correlation, but severely distort the true deformation measurements. On the premise that the accuracy information of the topographic data is unknown, the influence range of the topographic system error on the deformation result and the size of the deformation error introduced by the topographic system error are difficult to determine, so that the accuracy of the deformation result is particularly important to check and analyze after the interference processing is finished, such as the mutual comparison with the external ground deformation observation (such as GPS data) result.

In general, systematic errors in existing DEMs dominate, and therefore, prior to differential processing, the interferometric DEM or the topography interference phase map must be moderately spatially filtered.

Principle of monitoring inclination of line pole tower

By installing the solar data transmission base station and the tower inclination sensor on the tower body of the power transmission line, the inclination state of the tower is monitored in real time, so that the working state and the environmental state of the tower are controlled in real time, and the safety level and the operation and maintenance working efficiency of the power transmission line are improved. The solar data transmission base station collects environmental and working state information collected by the tower inclination sensor in real time and sends the information to the monitoring cloud platform through a special safety wireless channel. Aiming at the condition that no communication resource and no power supply are arranged on the tower, the project adopts a solar wireless convergence base station to communicate with a wireless public network, and alternative ADSS/OPGW optical fiber communication and Beidou short message emergency communication are adopted, so that communication with a monitoring cloud platform is realized, as shown in fig. 5.

The wireless convergence base station sends the collected sensor data to a security firewall of a monitoring center through a public network VPN, then the sensor data enter a cloud convergence server and an application server, the convergence server realizes information processing, information format normalization and data distribution, the application server realizes data storage, alarming and information display, and the unified cloud platform is suitable for full-power-grid operation data monitoring and supports multi-service mobile application. The wireless convergence base station encrypts the acquired sensor data through SM1/SM2/SM7 chip hardware and then sends the encrypted sensor data to a security firewall of a monitoring center, then the sensor data enter an internal decryption machine for decryption, the decrypted data are transmitted to a monitoring convergence server and an application server, the convergence server realizes information processing, information format normalization and data storage, and the remote application and the mobile APP application adopt AES soft encryption or special hardware encryption dogs to realize information transmission, so that information security is ensured.

The tower position and the inclination sensor are a full-function composite node in the wireless sensor network, the node is integrated with a BDS/GPS dual-mode positioning module, the functions of tower body positioning and wireless network time alignment can be completed, the high-precision inclination sensor is integrated, the inclination of the tower body can be monitored in real time, the node is integrated with 2 different channel links to respectively complete uplink and downlink communication, one LoRa long-distance channel is used for completing the long-distance communication with a wireless convergence base station and other full-function nodes, the other 2.4GHz micro-power wireless channel is used for completing the communication with other micro-power sensor nodes, the hybrid networking is realized, and the system complexity and the total possession cost are reduced. In order to realize private secret wireless communication transmission network, a national secret SM7 encryption chip is arranged in a wireless data transmission base station, so that possible network attack is stopped, and then the encrypted network attack is transmitted to a background information processing system. The micro-power wireless communication of the sensor and the like is generally non-sensitive information due to the fact that the acquired state information is short in transmission distance, the information is encrypted by adopting the SM7 encryption chip with low power consumption and then transmitted to the wireless data transmission base station, the whole process information safety control is achieved, and the information safety of the whole information sensing system is ensured.

InSAR satellite-based geological deformation monitoring of power transmission corridor

The InSAR technology is a novel space earth observation technology based on radar remote sensing, and the technology utilizes interference phase difference of two or more Synthetic Aperture Radar (SAR) images, can monitor large-area micro ground deformation with high precision, and realizes geometric measurement of ground deformation millimeter level. The InSAR satellite technology has the advantages of all-day, all weather, high precision, large range, long period, low cost and the like, and can be applied to monitoring of earth surface deformation and settlement of a power transmission corridor: (1) all weather and earth observation capability all the day, and is not affected by weather; (2) the monitoring precision is high, and millimeter-level deformation information can be measured; (3) the monitoring range is wide, and hundreds of square kilometers can be monitored at a time; (4) the monitoring density is high, and more than 1 ten thousand observation point data can be obtained in each square kilometer of urban areas; (5) the repetition frequency is high, the continuous monitoring capability is strong, macroscopic static information can be provided, and quantitative dynamic information can be given out; (6) the cost is low, a monitoring network does not need to be established, some potential or unknown target deformation information can be identified, and deformation conditions of a plurality of years can be provided.

By utilizing the characteristics of InSAR all-day, all-weather, high precision, large scale, long period, low cost and the like, the key technology of geological deformation monitoring and processing of the InSAR technology under the complex natural earth surface condition is researched, the earth surface deformation monitoring with millimeter level precision is realized, the large scale general survey of the geological deformation risk of a power transmission corridor is carried out by utilizing the long-time sequence image data of an InSAR satellite, the Beidou monitoring terminal is guided to be installed in the geological deformation risk area, the earth three-dimensional monitoring system is constructed, and the geological risk monitoring quality and efficiency of a power transmission line are comprehensively improved.

According to the method, based on geological deformation monitoring of the InSAR satellites, the long-time sequence InSAR satellite images are utilized to conduct large-scale geological deformation monitoring on the power transmission corridor, so that geological deformation settlement risk census and monitoring along the power transmission corridor are realized. Overall technical route, as shown in fig. 6.

(1) Data collection and arrangement

1) InSAR data source selection

InSAR satellite data source selection criteria: to obtain high-quality interferograms and perform differential interference processing, the coherence problem of source data, including time coherence and space coherence, is to be solved, that is, the SLC image pair must meet the requirement of critical base line distance in space, ensure the consistency of an observation target in a time interval, and avoid the phenomenon of losing coherence. In the process of data selection, the factors such as the type of data source, satellite orbit, spatial resolution, time interval and economy are considered.

According to the technical requirements, fully analyzing, adopting Sentinel 1A (Sentinel-1A) satellite data, wherein technical parameters of Sentinel 1AInSAR data are shown in table 1, and carrying out large-scale geological deformation monitoring by utilizing an InSAR measurement technology.

TABLE 1 sentinel 1AInSAR data technical parameters

2) Basic data collection

Collecting basic geographic information data of a research area, including 3D data (DLG, DOM, DEM) of the area and the like, for processing and analyzing InSAR ground subsidence; and collecting field measurement data of a research area, and laying a solid data foundation for InSAR ground subsidence monitoring work.

(2) InSAR preprocessing and interference processing

For Sentinel-1IW mode data, the image preprocessing flow is as follows: data preparation, IW data segmentation (IW split), burst extraction, doppler parameter calculation, deskew phase calculation, deskew processing, DEM registration, anticlockwise deskew, burst stitching, subband stitching and post-registration image. The pretreatment flow is shown in fig. 7.

1) Data reading

SAR data of different satellites have different data formats, and the SAR data are required to be converted into standard formats for subsequent processing. The original data format of Sentinel-1 is a tiff file, the tiff file is assembled and converted into a binary file, and corresponding parameter files are obtained at the same time, wherein 3 subband image data are three independent files, and data conversion and parameter extraction are respectively carried out. Then, the track parameter information is updated using the fine track data file.

With the improvement of the resolution of the SAR image, the data size also grows in geometric level, and the coverage area is 250 x 250km, and each scene of SLC image is about 3.7GB by taking a Sentinel-1 radar image (strip map,5 m resolution) as an example. In order to improve the data processing efficiency and reduce the storage space of the intermediate processing result, it is very necessary to select the region of interest.

2) TOPS data segmentation (TOPS split)

Because the burst data in each sub-band is not continuously acquired, there is overlap and black edge in the burst data, and there is also a difference in doppler frequency, and each burst data needs to be processed independently. Thus, the data splitting step is to separate an independent burst data file from each sub-band data.

3) Burst level data processing

After the independent Burst data file is acquired from the previous step, the following SAR image registration steps are all processed based on Burst level data. Because the Sentinel-1 data swings toward the antenna, the Doppler frequency linearly changes along with burst time, and the constant is no longer maintained, and at the moment, the interpolation and filtering of the data can cause great problems, the linear frequency modulation signal introduced by the Doppler frequency change must be compensated first, namely, the declining deramp is performed first before the registration process is performed, and specific calculation formulas can refer to ESA manual. After the registration process, the removed deramp phase needs to be compensated again, i.e. the deramp operation is performed, and then the subsequent interference process is performed.

Because Sentinel-1 has higher requirements on SAR image registration, multiple registration operations are generally required to be carried out by adopting a plurality of registration methods, and generally, coarse registration is firstly carried out by adopting a correlation operation or geometric registration algorithm, and then ESD fine registration operation is carried out subsequently. The most used geometric registration method in many open source software estimates the offset between the main image and the auxiliary image by means of external DEM data and fine track information, then carries out interpolation resampling, registers the auxiliary image to the reference frame of the main image, and can obtain higher registration precision in the rough registration step.

4) Burst splice and subband splice

And splicing the registered burst data, removing the overlapped area and the black edge (invalid value), and generating complete sub-band image data. In addition, the overlapping area needs to be reserved for later checking whether jump errors exist between adjacent bursts of the interferograms. And after splicing the Burst data of each sub-band, splicing the three independent sub-band data, removing the overlapping area between the sub-bands, and generating complete SLC image data.

5) Interference processing

And performing interference processing on the registered SLC images, namely performing conjugate calculation on the complex images, and acquiring corresponding interference parameter information.

6) Differential interference processing

Interference fringes caused by the land leveling phase and the topography phase mask deformation information, the land leveling phase and the simulated topography phase are generated by using the DEM and the track parameters, and then the phase is subtracted from the interference pattern, so that a differential interference pattern is finally obtained.

7) Interferogram inspection

The registration accuracy of the method is difficult to meet the registration accuracy requirement of a TOPS mode by using a traditional correlation operation and geometric registration algorithm, the phase jump problem exists in the generated interferogram, and the method can be judged by adopting a manual judgment method or performing interference treatment by using overlapping areas among bursts and calculating the phase difference of the overlapping areas.

8) ESD fine registration and interferogram correction

The flow of the ESD registration algorithm is shown in fig. 8, and the interference pair combination with registration error is obtained through interferogram inspection. And carrying out fine registration processing on the interferograms by using an ESD registration method, and converting the relative offset of the interference pair into the offset relative to the main image. Two correction methods exist, one is to correct the auxiliary image and regenerate the correct interferogram, and the other is to convert the offset into the offset phase corresponding to the SAR image by using the offset, and then directly compensate the interferogram.

9) Multi-view processing

And performing multi-view filtering on the finally generated differential interferograms to inhibit noise influence, and simultaneously reducing a part of data volume and reducing the large-range InSAR data processing workload.

(3) InSAR deformation monitoring embodiment

The time sequence InSAR processing flow mainly comprises: interferogram generation, deformation parameter estimation and deformation time sequence generation. Generating an interference pattern, namely determining a proper interference pair combination from the time sequence SLC image and DEM data which are ready to be registered, and generating a corresponding multi-view interference pattern and a corresponding coherence coefficient pattern file; the deformation parameter estimation is carried out, firstly, high coherence points are selected from the interference diagram, then, the time sequence analysis is carried out on the interference phases of the coherence points, various methods are utilized to improve the precision and stability of deformation calculation results, and finally, linear deformation rate products are generated; and generating a deformation time sequence, disentangling and atmospheric filtering the residual phase, thus obtaining a nonlinear deformation result, generating a final deformation time sequence product after geocoding, and converting the output result into an output format. The specific algorithm flow is as follows:

1) Optimal interference network connectivity graph generation

And combining N SAR images of the research area into a proper small baseline set according to the space-time baseline threshold value to form an optimal interference network connected graph. In order to reduce the space-time decoherence influence caused by long baselines, in the interference pair screening process, a group of interference pairs is initially determined by utilizing a small baseline threshold value, and then the initially selected interference combination is further screened, and an optimal interference pair combination is determined by adopting a method based on coherence estimation. Since the direct use of coherence for filtering will result in the operation of the coherence coefficient reaching (N-1) as the number of images N increases! The amount of calculation is too great. Therefore, the short baseline threshold is used for initial selection, and the interference pair number is usually about 4N, so that only the coherence coefficients of the 4N interference combinations need to be estimated.

2) High coherence point selection

Ground targets with stable scattering properties are chosen as high coherence targets. Similar to the PSI technique, the average coherence coefficient map is generated according to the previous step, starting from high coherence point (CS) identification, and screening is performed according to a certain threshold.

3) Distributed object selection

In order to ensure that the number of monitoring points in the project research area is sufficient, homogeneous pixel inspection is carried out on a non-high coherence area by using an AD inspection method, and a distributed target is selected. And estimating the coherence matrix of the homogeneous pixel region according to the checked homogeneous pixel region, and estimating the optimal phase sequence of the coherence matrix by using a maximum likelihood estimation method. After all distributed targets are acquired, the interference combination of the distributed targets is calculated according to the phase trigonometry and the optimal phase sequence. And finally, fusing the distributed target high coherence targets to form a CT/DS phase set. The DS Point selection is shown in FIG. 9.

4) Construction of CS/DS network and secondary differential phase model solving

After the high coherence point CS/DS initial selection point set is obtained, in order to solve the initial selection CS/DS points, a network is needed to be built for the CS/DS points, and a differential measurement method is used for deformation parameter estimation. In order to give consideration to network calculation efficiency and precision, various factors are comprehensively considered, and a multi-level network is adopted to improve the calculation precision and stability. The primary network adopts a global triangular network, and the closed network adjustment can be utilized to reduce the overall solution error. Because the large-scale interferogram has larger error space changes such as atmosphere and the like, the influence of space related errors cannot be completely eliminated by utilizing a differential measurement method, if a Delaunay network is directly utilized, the solution errors can be transmitted outwards step by step in space, and the final accumulated errors can even cover the real deformation result.

And performing secondary difference on adjacent CS/DS phases, estimating and obtaining differential measurement parameters, and converting the relative quantity of the differential measurement parameters into absolute quantity of global parameters by using a network adjustment resolving method, thereby obtaining a linear deformation rate diagram and a residual elevation diagram. The CS-Point and DS-Point fusion is shown in FIG. 10.

5) Residual phase separation

And (3) utilizing the time-space domain characteristics of the residual phase, separating each noise phase component through space-time filtering, and extracting the atmospheric delay phase. This step can be divided into three sub-steps, residual phase unwrapping, residual phase separation and atmospheric phase estimation.

6) Correction of elevation error and deformation solving

The linear velocity and elevation correction can effectively reflect the distribution of the deformation field. Thus, the original phase needs to be corrected, mainly considering the influence of the initial value of the linear velocity and the elevation error on iteration. After the two terms are obtained, the residual phase is filtered, the residual phase and the phase difference after the filtering are utilized after the filtering is completed, the residual phase and the phase difference after the filtering can be considered as the influences of the atmosphere and the noise, and the rest is the nonlinear deformation. By adding this to the linear deformation, the true deformation can be obtained. And establishing an observation equation on the selected coherent target. Solving by adopting a SVD singular value decomposition method to obtain the sedimentation rate of each time interval. And integrating the sedimentation rate of each period in a time domain to obtain the accumulated deformation quantity of each time point.

While the invention has been described in detail in the foregoing general description and specific examples, it will be apparent to those skilled in the art that modifications and improvements can be made thereto. Accordingly, such modifications or improvements may be made without departing from the spirit of the invention and are intended to be within the scope of the invention as claimed.

Claims

1. A deformation risk assessment method for a large-area power transmission channel of an InSAR in a complex environment utilizes long-time sequence InSAR satellite images to carry out large-scale geological deformation monitoring on a power transmission corridor, realizes geological deformation settlement risk census and monitoring along the power transmission corridor, and comprises the following steps:

2. The method for evaluating the deformation risk of the large-area transmission channel in the complex environment InSAR according to claim 1, wherein the step 1 is characterized in that the InSAR data source selection data standard comprises the steps of obtaining a high-quality interferogram and performing differential interference processing, wherein SLC images are required to meet the requirement of critical base line distance in space, the consistency of an observation target is ensured within a time interval, and the phase loss phenomenon is avoided; in the data selection process, the type of a data source, satellite orbit, spatial resolution, time interval and economy are considered; the basic data collection means for collecting basic geographic information data of a research area, and comprises the steps of collecting DLG, DOM, DEM data of 3D of the research area, and processing and analyzing InSAR ground subsidence; and collecting field measurement data of a research area, and laying a data foundation for InSAR ground subsidence monitoring work.

3. The method for evaluating deformation risk of large-area transmission channel in complex environment InSAR according to claim 1, wherein the data reading in the step 2 comprises the steps of converting SAR data into a standard format, converting the original data format of Sentinel-1 into a tiff file, converting the tiff file into a binary file, simultaneously acquiring corresponding parameter files, wherein 3 sub-band image data are three independent files, respectively performing data conversion and parameter extraction, and then updating track parameter information by using a fine track data file; the TOPS data segmentation comprises the independent processing of each burst data, wherein the data segmentation is to separate independent burst data files from each sub-band data; the Burst level data processing comprises the steps of compensating a linear frequency modulation signal introduced by Doppler frequency change, namely, performing declassification processing before the registration process is performed; after registration processing, compensating the removed deramp phase again, namely, reramp operation, and then carrying out subsequent interference processing; and estimating the offset between the main image and the auxiliary image by means of external DEM data and fine track information by using a geometric registration method, then carrying out interpolation resampling, and registering the auxiliary image under a main image reference system to obtain the registration precision required by coarse registration.

4. The method for evaluating deformation risk of the large-area transmission channel in the complex environment InSAR according to claim 3, wherein the Burst splicing and sub-band splicing in the step 2 comprises splicing registered Burst data, removing black edges of an overlapping area and invalid values, and generating complete sub-band image data; the overlapping area needs to be reserved additionally for later checking whether jump errors exist between adjacent bursts of the interferograms; after the Burst data of each sub-band are spliced, three independent sub-band data are spliced, overlapping areas among the sub-bands are removed, and complete SLC image data are generated; the interference processing comprises the steps of carrying out interference processing on the registered SLC images, carrying out conjugate calculation on the complex images, and simultaneously obtaining corresponding interference parameter information; the differential interference processing comprises that interference fringes caused by land leveling phases and terrain phases mask deformation information, land leveling and simulated terrain phases are generated by using DEM and track parameters, and then the land leveling and simulated terrain phases are subtracted from an interference pattern, so that a differential interference pattern is finally obtained.

5. The method for evaluating the deformation risk of the large-area transmission channel in the complex environment InSAR according to claim 4, wherein the interferogram inspection in the step 2 comprises eliminating phase jump existing in the generated interferogram in order to enable the registration precision to meet the registration precision requirement of a TOPS mode, and judging by adopting a manual judging method or performing interference processing by utilizing overlapping areas among bursts and calculating the phase difference of the overlapping areas; the ESD fine registration and interferogram correction comprises the steps of obtaining interference pair combinations with registration errors through interferogram inspection, carrying out fine registration processing on the interferograms by using an ESD registration method, converting the relative offset of the interference pairs into the offset relative to a main image, correcting the auxiliary image, regenerating a correct interferogram, converting the offset into an offset phase corresponding to an SAR image, and directly compensating the interferogram; the multi-view processing comprises multi-view filtering of a group of differential interferograms which are finally generated, noise influence is restrained, part of data quantity is reduced, and large-range InSAR data processing workload is reduced.

6. The method for evaluating deformation risk of the large-area transmission channel in the complex environment InSAR according to claim 1, wherein the generation of the optimal interference network connectivity map in the step 3 comprises the steps of combining N SAR images of a research region into an ideal small baseline set according to a space-time baseline threshold value to form the optimal interference network connectivity map; in order to reduce the space-time decoherence influence caused by long baselines, in the interference pair screening process, a group of interference pairs is initially determined by utilizing a small baseline threshold value, then the initially selected interference combination is further screened, and the optimal interference pair combination is determined by adopting a method based on coherence estimation.

7. The method for evaluating deformation risk of large-area transmission channels in complex environment InSAR according to claim 6, wherein the selecting of high coherence points in step 3 includes selecting a ground target with stable scattering characteristics as a high coherence target, first starting from high coherence point (CS) identification, generating an average coherence coefficient map according to the previous step, and screening according to a certain threshold; the distributed target selection comprises the steps of carrying out homogeneous pixel inspection on a non-high coherence area by utilizing an AD inspection method, selecting a distributed target, estimating a coherence matrix of the non-high coherence area according to the inspected homogeneous pixel area, estimating an optimal phase sequence of the coherence matrix by utilizing a maximum likelihood estimation method, calculating interference combinations of the distributed targets according to phase trigonometry and the optimal phase sequence after all the distributed targets are obtained, and finally fusing the distributed target high coherence targets to form a CT/DS phase set.

8. The method for evaluating deformation risk of the large-area transmission channel in the complex environment InSAR according to claim 7, wherein the construction of the CS/DS network and the solution of the secondary differential phase model in the step 3 comprise the steps of constructing a network for CS/DS points after a high coherence point CS/DS initial selection point set is obtained, performing deformation parameter estimation by using a differential measurement method, and adopting a multi-level network to improve the resolving precision and stability, wherein the primary network adopts a global triangular network, and the overall resolving error can be reduced by using a closed network adjustment; the second hierarchical network is added, and an irregular closed network formed by local reference points is used as a group of additional global constraint conditions, so that error transfer in the network resolving process can be greatly reduced; and carrying out secondary difference on adjacent CS/DS phases, estimating to obtain differential measurement parameters, and then converting the relative quantity of the differential measurement parameters into absolute quantity of global parameters by using a network adjustment resolving method, thereby obtaining a linear deformation rate diagram and a residual elevation diagram.

9. The method for evaluating deformation risk of large-area transmission channels in complex environment InSAR according to claim 8, wherein the residual phase separation in the step 3 comprises the sub-steps of utilizing the time-space domain characteristics of residual phases, separating each noise phase component through space-time filtering, and extracting the atmospheric delay phase, wherein the residual phase separation can be further divided into three sub-steps of residual phase unwrapping, residual phase separation and atmospheric phase estimation.

10. The method for evaluating the deformation risk of the transmission channel in the large area of the complex environment InSAR according to claim 9, wherein the correction of the elevation error and the solution of the deformation amount in the step 3 comprise correcting the original phase, taking the influence of the initial values of the linear velocity and the elevation error on iteration into consideration, filtering the residual phase after the initial values of the linear velocity and the elevation error are obtained, recognizing the difference between the residual phase and the filtered phase as the influence of the atmosphere and the noise after the filtering is completed, adding the difference into the linear deformation amount to obtain the real deformation amount; and establishing an observation equation on the selected coherent target, solving by adopting a SVD singular value decomposition method to obtain the sedimentation rate of each time interval, and integrating the sedimentation rate of each time interval in a time domain to obtain the accumulated deformation quantity of each time point.