CN105467390A - Bridge deformation close range monitoring method based on foundation InSAR - Google Patents

Abstract

The invention discloses a bridge deformation close range monitoring method based on a foundation InSAR, belonging to the engineering measurement field. The method comprises: obtaining two pairs of SAR images before and after bridge deformation through an InSAR, performing coherence processing on the SAR images to obtain two phase coherence images before and after bridge deformation; and performing difference processing on the two SAR images, finally obtaining a differential phase, i.e., a deformation phase, and obtaining bridge deformation according to the deformation phase. The InSAR measurement technology is all-weather and full-time, and possesses the characteristics of high resolution and continuous space coverage; accordingly, the method overcomes the disadvantages of single-point measurement, low resolution and susceptibility to extraneous interference, and can perform real-time and long-term monitoring on a bridge, and provide basis and guidance for bridge maintenance and management decision.

Description

A kind of bridge deformation based on ground InSAR closely monitoring method

Technical field

The present invention relates to supervisory system, be specifically related to a kind of bridge deformation based on ground interference synthetic aperture radar closely monitoring method.

Background technology

Bridge, as one of infrastructure that China is important, has vital effect to the development of national economy.Once the security of bridge can not ensure completely, negative effect will be brought to economic development.And along with the development of social progress and science and technology, the constantly progress of bridging technology, bridge structure progressively to the development of light and handy, very thin aspect, meanwhile bridge load-carrying, constantly increase across footpath and bridge deck width, structural shape constantly changes.Traditional deformation monitoring means are all almost based on " point " formula collecting method, intactly can not embody the deformation of bridge comprehensively, spatially can not realize the object of bridge deformation EARLY RECOGNITION and dynamic monitoring, so this just can carry out the method for long-term Real-Time Monitoring more reliably to the health parameters of bridge in the urgent need to one.

Last century the nineties, interference synthetic aperture radar grows up gradually, the miniature deformation on earth's surface can be monitored, there is the advantage of round-the-clock, round-the-clock, broad covered area, increasingly automated and high precision monitor earth's surface distortion, be that effective that measurement of the level and GPS measure supplements, long term monitoring bridge deformation and health degree thereof have clear superiority.

Show after deliberation, because bridge deformation monitoring is a long-term process, cause the quality of image of SAR to receive the impact of several factors, the impact of especially season and weather conversion.So in the link of data processing, season be added and weather condition is corrected, filtering to monitor useless data to bridge deformation, thus make measurement more accurate.

Summary of the invention

1. object: problem to be solved by this invention is in existing monitoring method, provides a kind of bridge deformation based on ground interference synthetic aperture radar closely monitoring method.

2. technical scheme; The present invention is a kind of bridge deformation based on ground interference synthetic aperture radar closely monitoring method, and the concrete steps of the program are as follows:

Step one: the pre-service of SAR image

Non-stationary image is obtained in order to prevent the scattering coefficient on bridge floor from differing comparatively large, needing to carry out pre-filtering process to image---orientation, to filtering and distance to filtering, namely reduces because of the mistake correlation effect of echo bearing to Doppler frequency center caused by different and baseline introducing.

Step 2: SAR Image registration

Due to the existence of baseline, two antennas are all had any different to the angle of target area and oblique distance, thus two width SAR image distance to orientation to all there is certain dislocation and the anglec of rotation, cause identical impact point position in two images to be differed.The object of image registration is exactly make the same point in the some corresponding ground scene of same position in complex pattern, improves the signal to noise ratio (S/N ratio) of follow-up interference treatment, makes interference fringe more clear.

Step 3: go Horizon effect

Elliptical earth phase does not comprise the information of elevation change, and makes the striped of interferometric phase become close, affects the effect of phase unwrapping, so need before this to remove elliptical earth phase striped.

Step 4: interferometric phase filtering

Instrument work and environment can produce additive noise, and there is distinctive multiplicative noise in SAR.So in order to ensure the quality that subsequent phase solution twines, must noise-removed filtering, make phase place more clear.

Step 5: phase unwrapping

Electronic equipment can only record the phase place of 0 to 2 π, is namely wound around phase place, and real interferometric phase needs to add 2n π on the basis being wound around phase place, so must pass through this step of phase unwrapping, obtains the true phase corresponding with Terrain Elevation.

3. advantage and effect:

The defect of one aspect of the present invention can overcome " point " formula collecting method, carries out omnibearing monitoring to bridge, has round-the-clock, round-the-clock, not by atmospheric propagation and climate effect, and its measuring accuracy can reach a millimeter rank.The whole process of bridge deformation can be observed on the other hand by interferogram, can Timeliness coverage bridge whether impaired, be convenient to it and safeguard.

Accompanying drawing explanation

Fig. 1 be in the present invention interference synthetic aperture radar to the structural representation of bridge monitoring;

Fig. 2 is that in the present invention, InSAR surveys high principle schematic.

Fig. 3 is the process flow diagram of processing procedure of the present invention.

Embodiment

Some operational methods that the present invention relates to

1, high principle is surveyed

Utilize the two width antennas with certain subtense angle to obtain the relevant SAR image of the same area, then calculate height value according to its interferometric phase, as shown in Figure 2.

2, the relation of bridge deformation amount and interferogram phase differential

Antenna A ₁with antenna A ₂represent the position of two slave antennas respectively, the parallax range B of the distance between antenna represents, the angle of baseline and horizontal direction is α.The height of h representative antennas and bridge bottom surface, P is impact point, and R is that P is to antenna A ₁distance, R+ Δ R is that P is to antenna A ₂distance, θ is the reference viewing angle of the first slave antenna, and P point deformation quantity z represents.

Antenna A ₁with antenna A ₂the phase differential of the signal received is:

Due in the electronic system of reality, the phase place of 0 to 2 π can only be recorded, and the phase place of reality is:

$\begin{array}{cc}\mathrm{\φ}=\frac{2\mathrm{\π}\mathrm{\Δ}R}{\mathrm{\λ}}+2N\mathrm{\π}& N=0,\±1,\±2,...\end{array}---\left(2\right)$

Can be in the hope of by Fig. 2

$sin(\mathrm{\α}-\mathrm{\θ})=\frac{{(R+\mathrm{\Δ}R)}^2-R^2-B^2}{2RB}---\left(3\right)$

z＝h-Rcosθ(4)

Ignore Δ R ², can be obtained fom the above equation

$\mathrm{\Δ}R\≈B\sin (\mathrm{\α}-\mathrm{\θ})+\frac{B^2}{2R}---\left(5\right)$

In sum $\frac{d\mathrm{\φ}}{dz}=\frac{4\mathrm{\π}}{\mathrm{\λ}}\frac{B_{\⊥}}{R\sin \mathrm{\θ}}---\left(6\right)$

Above formula reflects the rule of interferometric phase conformal variable change

The present invention is two pairs of SAR imagings by first obtaining before and after bridge deformation, and they are carried out Coherent processing separately, obtains two width phase coherence figure before and after deformation; Again this two width image is carried out difference processing, the differential phase finally obtained, bridge deformation amount can be obtained by phase differential.Its detailed process as shown in Figure 3

The first step: respectively front and back two width figure is carried out image registration.

Use maximum spectrum method, there is peak value to frequency spectrum in the distance of interferometric phase image, particular location is determined by Y-PSNR HLR on certain frequency

$HLR=\frac{MAX\left|{\mathrm{RF}}_{12}(m,n)\right|}{{\Σ}_{m=0}^{M,m\≠m_0}{\Σ}_{n=0}^{N,m\≠m_0}\left|{\mathrm{RF}}_{12}(m,n)\right|}---\left(7\right)$

RF in formula ₁₂for the frequency spectrum of interferometric phase image

The maximum position of HLR just correspond to best registration.

Second step: principal and subordinate's images filter.

3rd step: flat earth.

If the spectrum peak of N point interferometric phase appears at l point place, then corresponding fringe frequency is

$f_{fringe}=\frac{l}{N}\·RSR---\left(8\right)$

Wherein RSR is distance samples rate:

$RSR=\frac{1}{{\mathrm{\Δt}}_r}=\frac{c}{2R_r}---\left(9\right)$

Wherein Δ t _rfor sampling time interval; R _rfor antenna is to the oblique distance of target to distance.At frequency domain by spectrum peak ring shift to zero point, then corresponding time domain phase only pupil filter φ _correct(n)

${\mathrm{\φ}}_{correct}\left(n\right)=\frac{2\mathrm{\π}}{N}nl=\frac{2\mathrm{\π}}{N}n\frac{N\·f_{fringe}}{RSR}=\frac{2{\mathrm{\πf}}_{fringe}}{RSR}n---\left(10\right)$

Because for any contour reference surface, fringe frequency f _fringewith oblique distance fringe frequency f _fringe.rpass be:

$f_{fringe}=f_{fringe.r}\frac{c}{2}---\left(11\right)$

So, ${\mathrm{\φ}}_{correct}\left(n\right)=2{\mathrm{\πf}}_{fringe}\frac{2}{c}{\mathrm{nR}}_r=2{\mathrm{\πf}}_{fringe.r}{\mathrm{nR}}_r---\left(12\right)$ Frequency displacement method is in fact the ring shift by frequency spectrum, introduces the modifying factor equaling relative level land phase place in time domain, thus reaches the object removing elliptical earth phase.

4th step: spatial filtering noise reduction

Utilize the stripe centerline of cell neural network parallel behavior quick obtaining interference fringe picture, ask for filter window according to stripe centerline, utilize this filter window to carry out filtering to interference fringe picture.Consider that InSAR Interferometric phase fringe image has 2 π saltus step lines, adopt sine and cosine filter method to realize the filtering of InSAR interferogram.

5th step: phase unwrapping.

If the phase data after phase unwrapping is φ _i,j, then the thought of least square phase unwrapping method minimizes following formula:

$\underset{i=0}{\overset{M-1}{\Σ}}\underset{j=0}{\overset{N}{\Σ}}{({\mathrm{\φ}}_{i+1,j}-{\mathrm{\φ}}_{i,j}-{\mathrm{\Δ}}_{i,j}^x)}^2\underset{i=0}{\overset{M}{\Σ}}\underset{j=0}{\overset{N-1}{\Σ}}{({\mathrm{\φ}}_{i,j+1}-{\mathrm{\φ}}_{i,j}-{\mathrm{\Δ}}_{i,j}^y)}^2---\left(13\right)$

After arranging weights, above formula becomes weighted equation:

$\underset{i=0}{\overset{M-1}{\Σ}}\underset{j=0}{\overset{N}{\Σ}}U(i,j){({\mathrm{\φ}}_{i+1,j}-{\mathrm{\φ}}_{i,j}-{\mathrm{\Δ}}_{i,j}^x)}^2\underset{i=0}{\overset{M}{\Σ}}\underset{j=0}{\overset{N-1}{\Σ}}V(i,j){({\mathrm{\φ}}_{i,j+1}-{\mathrm{\φ}}_{i,j}-{\mathrm{\Δ}}_{i,j}^y)}^2---\left(14\right)$

Wherein

$\begin{array}{c}U(i,j)=\min (w_{i+1,j}^2,w_{i,j}^2)\\ V(i,j)=\min (w_{i,j+1}^2,w_{i,j}^2)\end{array}---\left(15\right)$

Because weighted least-squares cannot solve with fast algorithms such as DCT, FFT, iterative algorithm can only be adopted to solve.Can be expressed as:

$\mathrm{\φ}(i,j)=\frac{U(i,j){\mathrm{\φ}}_{i+1,j}+U(i-1,j){\mathrm{\φ}}_{i-1,j}+V(i,j){\mathrm{\φ}}_{i,j+1}+V(i,j-1){\mathrm{\φ}}_{i,j-1}-c_{i,j}}{U(i,j)+U(i-1,j)+V(i,j)+V(i,j-1)}$

$\left(16\right)$

Wherein c _i,jfor the Laplacian operator of weighting, its expression formula is:

$c_{i,j}=U(i,j){\mathrm{\Δ}}_{i,j}^x-U_x(i-1,j){\mathrm{\Δ}}_{i-1,j}^x+V(i,j){\mathrm{\Δ}}_{i,j}^y-V(i,j-1){\mathrm{\Δ}}_{i,j-1}^y---\left(17\right)$

6th step: differential interferometry

The two width interferograms that front and back obtain are carried out difference, and the differential interferometry figure before and after obtaining, is then reflected to phase differential in bridge deformation amount.

Claims (5)

1. the method using interference synthetic aperture radar to monitor bridge deformation, is characterized in that:

A, interference synthetic aperture radar is arranged on bridge pier, to the without hindrance effect in navigation channel

B, two SAR with certain subtense angle are utilized to carry out imaging

C, maximum spectrum registration method is used to carry out registration to two width SAR image

D, utilization interferogram spectrum offset carry out flat earth to the interferogram generated

E, utilization interference fringe center line filtering method are to interferogram filtering

F, utilization weighted least-squares method are carried out solution to phase place and are twined.

2. method according to claim 1, is characterized in that, step C specifically comprises:

C1, according to complex image to data acquisition interferogram frequency spectrum;

The maximum modulus value of C2, calculation interferogram frequency spectrum and the ratio of other radio-frequency component modulus value sums, this ratio is called evaluation function.

3. method according to claim 1, is characterized in that, step D specifically comprises:

D1, according to interfere the discrete Fourier transformation carried out line by line of reset diagram try to achieve peak-peak frequency be interference fringe distance to main frequency;

D2, according to main frequency center, by the ring shift of frequency domain spectra, introduce in time domain and equal the modifying factor of relative level land phase place.

4. method according to claim 1, is characterized in that, step e specifically comprises:

E1, according to cell neural network parallel processing extract stripe centerline;

The filter window that E2, basis are determined by center line carries out filtering to image.

5. method according to claim 1, is characterized in that, step F specifically comprises:

F1, utilize the phase gradient being wound around phase calculation horizontal direction or vertical direction according to least square method;

F2, carry out integration according to the phase gradient of horizontal direction and vertical direction.