CN108022259A - Interference SAR complex image corregistration method and system - Google Patents
Interference SAR complex image corregistration method and system Download PDFInfo
- Publication number
- CN108022259A CN108022259A CN201610973301.1A CN201610973301A CN108022259A CN 108022259 A CN108022259 A CN 108022259A CN 201610973301 A CN201610973301 A CN 201610973301A CN 108022259 A CN108022259 A CN 108022259A
- Authority
- CN
- China
- Prior art keywords
- mrow
- image
- munderover
- offset
- registration
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/10—Image acquisition modality
- G06T2207/10032—Satellite or aerial image; Remote sensing
- G06T2207/10044—Radar image
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20016—Hierarchical, coarse-to-fine, multiscale or multiresolution image processing; Pyramid transform
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/30—Subject of image; Context of image processing
- G06T2207/30181—Earth observation
- G06T2207/30184—Infrastructure
Abstract
The invention discloses a kind of method of interference SAR complex image corregistration, including:Master image and auxiliary image are subjected to rough registration, to obtain the first offset;According to the first offset obtained, the part of the master image and the auxiliary image rough registration is taken out, carries out the interpolation of same multiple respectively;By the image after interpolation with specific line-spacing, row away from carrying out piecemeal to obtain multiple sub-blocks;For each sub-block, the second offset is calculated;And using the master image as reference, resampling is carried out according to second offset to each sub-block, to obtain the SAR complex patterns after registration.
Description
Technical field
The present invention relates to SAR complex image corregistrations, are more particularly to used to interfere or chromatograph SAR complex image corregistrations method and be
System.
Background technology
Chromatography synthetic aperture radar technology (abbreviation TomoSAR or chromatography SAR) and interference synthetic aperture radar technology (abbreviation
InSAR or interference SAR) both for repeatedly navigating or multiple antenna single navigated the SAR complex patterns of the areal received
A series of processing is carried out, obtains the means of Three Dimensional Ground information.In this series of processing procedure, the registration of complex pattern is
Obtain one of committed step of Three Dimensional Ground information.Because two width SAR images imaging track, visual angle or the deviation of time, away from
Descriscent and orientation all can must use two width of Same Scene to scheme again there are certain dislocation and distortion before generating interference pattern
As being accurately aligned with, make the same echo point on the pixel corresponding ground of same position in two images, to ensure that interference pattern has
There is higher signal-to-noise ratio.
As it is known to the person skilled in the art, SAR image belong to oblique distance projection coherent imaging, therefore imaging mechanism,
There is very big difference in radiation characteristic and geometrical property with visible images.Therefore, although SAR image registration technique and optics
Remote sensing image registration technology has very big similitude in theory, but compared with conventional optical image, SAR image is more difficult to realize essence
True registration, its main cause are:
(1) due to microwave imaging irradiating angle, irradiation time are different when and irradiate the reason such as change of atural object in scene, lead
SAR is caused to the texture of the detection image of same scene and the difference of gray scale;
(2) coherence imaging system is inevitably present coherent speckle noise to the fine structure of image, resolvability and can
Illustrative destruction;
(3) since the SAR penetration capacitys of different-waveband are different, the description of the image of formation to Same Scene has different
Embody.
SAR image registration method mainly has the method for registering based on geometric properties and the method for registering based on gray feature.
Method for registering based on geometric properties has been obtained for extensive use.Moreover, many random scatter units composition of SAR image
Resolution cell has larger energy and higher spatial frequency, so traditional matching somebody with somebody based on geometric properties there are coherent spot
Quasi- method is no longer appropriate for high-precision interference SAR registration.
Registration Algorithm based on gray scale be usually by the use of two width SAR images grey-level statistics as similitude measurement
Standard, advantage be it is simple and practicable, shortcoming be to noise-sensitive, and when target deforms upon, rotates, size change over when easily lead
Algorithm failure is caused, and calculation amount is excessive, can not handle the registration of big breadth SAR complex patterns in real time.
For example, document " the High Resolution Spaceborne SAR image of Gray and textural characteristics building goes extracting method " proposes
A kind of method for registering images based on gray scale, this method can solve the data of various geomorphic types well, and effect is fine,
But in the algorithm, registering offset elimination of rough difference result has tremendous influence to registration below, and is not also suitable for interference SAR
Registration.
The registration of SAR image is carried out using DEM and satellite orbit parameter, its registration accuracy can reach Pixel-level, but picture
Plain level registration accuracy is not appropriate for the application of chromatography SAR and interference SAR.
The method of the existing interference SAR registration based on view data mainly have correlation function algorithm, flat wave function method and
Maximum Power estimation method.Wherein, correlation function algorithm and Power estimation method carry out registration using image amplitude information.Averagely fluctuate letter
Number method is to evaluate the effect of registration by assessing interference fringe quality.For the SAR image registration of big breadth, directly using this
A little methods do not ensure that the preferable coherence of whole image.
Therefore, it is necessary to a kind of method and system for realizing interference SAR complex image corregistration, especially meets big breadth interference
The method and system of SAR complex image corregistrations.
The content of the invention
According to one embodiment of present invention, there is provided a kind of method of interference SAR complex image corregistration, including:By master map
Picture and auxiliary image carry out rough registration, to obtain the first offset;According to the first offset obtained, by the master image and institute
The part for stating auxiliary image rough registration is taken out, and carries out the interpolation of same multiple respectively;By the image after interpolation with specific line-spacing, row
Away from carrying out piecemeal to obtain multiple sub-blocks;For each sub-block, the second offset is calculated;And using the master image as ginseng
Examine, resampling is carried out according to second offset to each sub-block, to obtain the SAR complex patterns after registration.
Preferably, described the step of master image and auxiliary image are carried out rough registration, further comprises:Obtain of rough registration
With desired value, wherein orientation and distance corresponding to maximum in the matching desired value to the offset of Pixel-level be
First offset.
Preferably, the matching desired value is related coefficient.
Preferably, second offset of each sub-block is obtained by calculating the compound correlative function of corresponding sub-block
's.
Preferably, the complex correlation function is defined as:
Wherein, s1(m, n) and s2(m, n) represents the master image and auxiliary image respectively;M and N represents to be used to participate in correlation
The window size of calculating;(u, v) represents the offset slided during correlation computations;Represent modulo operation respectively with * and take conjugation to grasp
Make.
According to another embodiment of the invention, there is provided a kind of system of interference SAR complex image corregistration, including:Slightly match somebody with somebody
Quasi-mode block, the rough registration module is configured to master image and auxiliary image carrying out rough registration, to obtain the first offset;Interpolation
Module, the interpolating module are configured to, according to the first offset obtained, the master image and the auxiliary image slightly be matched somebody with somebody
Accurate part is taken out, and carries out the interpolation of same multiple respectively;Piecemeal module, the piecemeal module are configured to the figure after interpolation
As obtaining multiple sub-blocks away from piecemeal is carried out with specific line-spacing, row;Computing module, the computing module are configured to for every
One sub-block, calculates the second offset;And resampling module, the resampling module are configured to using the master image as ginseng
Examine, resampling is carried out according to second offset to each sub-block, to obtain the SAR complex patterns after registration.
Preferably, the rough registration module is further configured to:The matching desired value of rough registration is obtained, wherein described
With the orientation corresponding to the maximum in desired value and distance to the offset of Pixel-level be first offset.
Preferably, the matching desired value is related coefficient.
Preferably, the computing module is configured to:Each is obtained by calculating the compound correlative function of corresponding sub-block
Second offset of sub-block.
According to the detailed description below the disclosure and attached drawing, mesh other to those skilled in the art
, feature and advantage will be apparent.
Brief description of the drawings
Attached drawing illustrates the embodiment of the present invention, and for explaining the principle of the present invention together with specification.In the accompanying drawings:
Fig. 1 shows the block diagram of SAR complex image corregistrations system 100 according to embodiments of the present invention.
Fig. 2 shows the flow chart of SAR complex image corregistration methods according to embodiments of the present invention.
Fig. 3 shows image block schematic diagram according to embodiments of the present invention.
Fig. 4 shows L-band Alos SAR master images according to embodiments of the present invention and auxiliary image.
Fig. 5 shows rough registration related coefficient figure according to embodiments of the present invention.
Fig. 6 shows SAR master images according to embodiments of the present invention and auxiliary image rough registration administrative division map.
Fig. 7 shows rough registration interference pattern according to embodiments of the present invention.
Fig. 8 shows the major-minor image after interpolation according to embodiments of the present invention.
Embodiment
A kind of method and system of interference SAR complex image corregistration is disclosed according to an embodiment of the invention.It is described below
In, for illustrative purposes, multiple details are elaborated to provide the comprehensive understanding to the embodiment of the present invention.It is however, right
In those skilled in the art it is readily apparent that the embodiment of the present invention can be realized in the case of these no details.
SAR image is two width by the same imageable target in the ground for having the observation of small differential rail to obtain twice
SAR complex patterns.Each echo point in target area corresponds to each pixel point in SAR image.Due to the angle being imaged twice
There are less difference so that and same echo point correspond to different pixel positions in two width SAR images in target area,
I.e. there is offset in the position of each corresponding pixel of two images.Therefore, in interference imaging processing procedure, it is impossible to directly calculate two width
The phase difference of each pixel point in SAR image, but need, to SAR image registration, to make corresponding pixel point position in two images
Put alignment.General thinking is:Using piece image therein as master image, image supplemented by another piece image, according to certain registration side
The offset between two images is calculated in method, then auxiliary image resampling is made master image and adopted again according to offset
The pixel point of same position corresponds to the same target point in ground in auxiliary image after sample.
For the SAR image of big breadth, to realize interference SAR or chromatograph SAR match somebody with somebody alignment request, it is necessary to it is registering to carry out essence,
Reach pixel level registration.But the SAR image of big breadth, in different azimuth and offset apart from upward sub-pixel and differ
Cause, or even in some cases, the offset of Pixel-level is also inconsistent.Therefore, cannot be to whole using correlation function method for registering
Width image is handled using same offset.
In addition, for the SAR image of big breadth, do not merely have to carry out rough registration, and even in the rough registration stage significantly
Wide SAR image also can there are offset it is inconsistent the problem of.
SAR image registration based on complex correlation function
According to an embodiment of the invention, it is proposed that a kind of system of SAR image registration based on complex correlation function and side
Method, the system and method can carry out the registration of SAR complex patterns exactly and in real time, and it is big to be particularly suitable for operand
The registration of big breadth SAR complex patterns.
Fig. 1 shows the block diagram of SAR complex image corregistrations system 100 according to embodiments of the present invention.As shown in Figure 1, system
100 include rough registration module 101, interpolating module 103, piecemeal module 105, computing module 107 and resampling module 109.
The rough registration module 101 is configured to master image and auxiliary image carrying out rough registration, to obtain the first offset.
Specifically, several equally distributed pixels are selected in master image, respectively centered on each pixel, window size is of m × n
With window, in auxiliary image, respectively centered on the pixel of same coordinate position, the search window of a larger window M × N is taken,
Search window must take sufficiently large can find the corresponding window to match with matching window inside it.Then, by row in search window
Row order with different whole pixel-shift amount choose with matching window corresponding window of the same size, respectively calculating matching window with it is each corresponding
The registering desired value of window, can obtain a matching desired values of (M-m+1) × (N-n+1) by calculating, take and wherein match desired value
Maximum correspondence window is the window of matching window registration, the center pel of the matching window center of master image with corresponding to window in auxiliary image
Coordinate position difference is offset.
In one embodiment of the invention, the rough registration module 101 is registering according to the SAR image of complex correlation function
Method obtains the related coefficient of rough registration, wherein the orientation and distance corresponding to maximum in the related coefficient to picture
The offset of plain level is first offset.That is, matching desired value is related coefficient.However, people in the art
Member is it is appreciated that matching desired value can be other desired values.
The interpolating module 103 is configured to according to the first offset obtained, by the master image and the auxiliary figure
As the part of rough registration is taken out, the interpolation of same multiple is carried out respectively.The piecemeal module 105 is configured to the figure after interpolation
As obtaining multiple sub-blocks away from piecemeal is carried out with specific line-spacing, row.The computing module 107 is configured to for each height
Block, calculates the second offset.Specifically, the computing module 107 calculates the phase relation of each sub-block according to compound correlative function
Orientation corresponding to maximum and distance in number to corresponding second offset.The resampling module 109 is configured
Into using the master image as reference, resampling is carried out according to second offset to each sub-block, after obtaining registration
SAR complex patterns.
Compound correlative function
When actually carrying out SAR complex image corregistrations, SAR complex patterns are made of many discrete pixels, and orientation
To unknown to offset with distance, the correlation function of SAR complex patterns can be defined as according to the definition of statistics correlation function:
In formula (1), s1(m, n) and s2(m, n) represents the actual SAR complex patterns obtained of two width, it is, before representing respectively
The master image stated and auxiliary image;M and N represents to be used for the window size for participating in correlation calculations;(u, v) represents to slide during correlation computations
Dynamic offset;Represent modulo operation respectively with * and take conjugate operation.
Fig. 2 shows the flow chart 200 of SAR complex image corregistration methods according to embodiments of the present invention.
SAR image rough registration
Rough registration is also known as pixel level registration, it calculates matching by ranks in search window with different whole pixel offsets
Quality of registration evaluation index between window and corresponding data window, thus obtained registration accuracy is within a pixel.Slightly match somebody with somebody
Accurate purpose is so that how overlapping the input window of follow-up sub-pixel registration is as far as possible.
In one embodiment, rough registration module 101 in step S201 using above-mentioned complex correlation function by master image and
Auxiliary image carries out rough registration, to obtain the first offset.Specifically, in rough registration step, rough registration module 101 is according to complex phase
The SAR image registration method of closing property function obtains the related coefficient of rough registration, wherein the maximum institute in the related coefficient is right
The orientation and distance answered to the offset of Pixel-level be first offset (u, v).
SAR complex patterns after rough registration only ensure pixel-shift amount (u, v) as close possible to true excursions amount, but not
It can meet interference SAR and chromatograph the requirement of SAR, it is therefore necessary to carry out essence registration.
SAR image essence registration
Compared with rough registration, since smart registration accuracy requires height, so needing the matching window and search window number to complex pattern
According to doing sub-pixed mapping interpolation.
According to the present invention, in the processing of SAR image essence registration, the SAR complex patterns of the big breadth of rough registration are inserted
Value, then carries out complex correlation function calculating to its piecemeal, obtains corresponding to the relevance function maximum of each piecemeal
Sub-pixel offset, then using master image as reference, resampling is carried out according to piecemeal to auxiliary image.Specific method and step is as schemed
Shown in 2.
In step S203, according to the first offset obtained, the part of master image and auxiliary image rough registration is taken out, point
The interpolation of same multiple is not carried out.It will be understood by those skilled in the art that the method for interpolation include but not limited to Fourier's interpolation method,
Closest interpolation method, bilinear interpolation, 4 cubic convolution interpolation methods and B-spline interpolation method etc..In practical applications, can be with
Interpolation algorithm is selected according to the needs of interpolation precision and algorithm calculation amount.
In order to make image registration more accurate, it is necessary to carry out piecemeal processing to image.In step S205, by the image after interpolation
With specific line-spacing, arrange away from piecemeal is carried out to obtain multiple sub-blocks.It will be understood by those skilled in the art that piecemeal is smaller, precision
It is higher.Fig. 3 shows image block schematic diagram according to embodiments of the present invention.By the image after interpolation in the first step with certain
Line-spacing, arranges away from piecemeal is carried out, as shown in Figure 3.M in figuresAnd NsEach piece of columns and line number during piecemeal are represented respectively.In order to scheme
As registration accuracy reaches sub-pixel, it is necessary to which every piece that second step has been divided calculates offset again.
In step S207, for each sub-block, the compound correlative function of each sub-block is calculated to obtain each sub-block
The second offset.Due to major-minor image, rough registration, every piece of offset will not be typically smaller than rough registration and obtained very greatly
The offset obtained.
In step S209, into row interpolation resampling.Specifically, according to the second of the sub-block calculated in step S207 the offset
Amount, using master image as reference, resampling is carried out to each sub-block according to the second offset.The result of resampling is the knot of registration
Fruit.
Experimental result and analysis
In practice, the present invention uses L-band ALOS PALSAR complex patterns, and big breadth image has been carried out to be based on complex phase
The method for registering images research of closing property function, rough registration offset is obtained in rough registration using complex correlation function maxima,
Then by the image of rough registration into row interpolation, the sub-pixel that piecemeal obtains each piecemeal using complex correlation function maxima is inclined
Shifting amount, then using main picture as reference, resampling is carried out according to piecemeal to auxiliary image.Truthful data handling result is demonstrated according to this
The validity of the method for invention.
Fig. 4 shows L-band Alos SAR master images according to embodiments of the present invention and auxiliary image.L-band is used herein
Alos SAR datas are tested.The two width SAR images for choosing 512*512 are handled, and major-minor image is as shown in Figure 4.
Rough registration
The related coefficient of rough registration is obtained according to the SAR image registration method of complex correlation function.Fig. 5 is shown according to this
The rough registration related coefficient figure of inventive embodiments.As shown in Figure 5.Related coefficient maximum is 0.8848, side corresponding at this time
Position range pixel level offset is (134,44).
According to the offset of rough registration, the rough registration result of major-minor image is obtained.Fig. 6 is shown according to embodiments of the present invention
SAR master images and auxiliary image rough registration administrative division map.The overlapping region of major-minor image is as shown in Figure 6.Overlapping region is carried out just
Step interference, obtains interference fringe.Fig. 7 shows rough registration interference pattern according to embodiments of the present invention.Can from the registration result of Fig. 7
To see, the interference fringe of image is may be seen indistinctly, but and unintelligible.The part that major-minor image rough registration overlaps is retrieved,
As essence with punctual major-minor image.The size of main and auxiliary image after the rough registration of this experiment interception is 378*468.
Smart registration
The first step, using Fourier's interpolation method, to the major-minor image after rough registration, it carries out 10 times of interpolation, interpolation respectively
The results are shown in Figure 7.Major-minor image orientation Range Profile vegetarian refreshments after interpolation is respectively 3780 and 4680.
Second step, piecemeal is carried out to the image after interpolation.Herein using azran to being all 1000 pixels as spacing
Carry out piecemeal.
3rd step, calculates the offset of each sub-block.The calculating of correlation is carried out in each sub-block, due to being to rough registration
The essence registration again that image carries out afterwards, so offset will not be very big.Using the center of each sub-block as reference, using correlation
The 1/10 sub-pix offset that function method for registering obtains each sub-block is shown in Table 1, and each sub-block calculates the related coefficient obtained such as
Shown in table 2.
1/10 sub-pixel offset of 1 each sub-block of table
The corresponding related coefficient of 2 each sub-block of table
4th step, resampling.For the master image of 10 times of interpolation, 10 demultiplication samplings are directly carried out;For auxiliary image, according to
1/10 sub-pixel offset of each sub-block re-starts 10 demultiplications sampling.Fig. 8 shows according to embodiments of the present invention insert
Major-minor image after value.The interference fringe picture obtained is as shown in Figure 8.Can be intuitively from the result of rough registration and essence registration
Arrive, the interference fringe after progress essence registration is very clear.
According to an embodiment of the invention, a kind of system and method for simple SAR complex image corregistrations are realized.The present invention's
The system and method for SAR complex image corregistrations are real-time, accuracy is high, and have very strong antijamming capability and robustness.
The registration accuracy of the system and method can reach sub-pixel, suitable for interference SAR image and chromatograph matching somebody with somebody for SAR image
Standard, is particularly suitable for big breadth SAR complex image corregistrations.
Above-described embodiment is only the preferred embodiment of the present invention, is not intended to limit the invention.To those skilled in the art
It is readily apparent that without departing from the spirit and scope of the present invention, various repair can be carried out to the embodiment of the present invention
Change and change.Therefore, the invention is intended to cover to fall into all within the scope of the present invention as defined by the appended claims and repair
Change or modification.
Claims (10)
1. a kind of method of interference SAR complex image corregistration, including:
Master image and auxiliary image are subjected to rough registration, to obtain the first offset;
According to the first offset obtained, the part of the master image and the auxiliary image rough registration is taken out, is carried out respectively
The interpolation of same multiple;
By the image after interpolation with specific line-spacing, row away from carrying out piecemeal to obtain multiple sub-blocks;
For each sub-block, the second offset is calculated;And
Using the master image as reference, resampling is carried out according to second offset to each sub-block, after obtaining registration
SAR complex patterns.
2. according to the method described in claim 1, wherein, described the step of master image and auxiliary image are carried out rough registration, is further
Including:
Obtain rough registration matching desired value, wherein it is described matching desired value in maximum corresponding to orientation and distance to
The offset of Pixel-level be first offset.
3. according to the method described in claim 2, wherein, the matching desired value is related coefficient.
4. according to the method described in claim 1, wherein, second offset of each sub-block is by calculating corresponding son
The compound correlative function of block and obtain.
5. according to the method described in claim 1, wherein, the complex correlation function is defined as:
<mrow>
<mi>&rho;</mi>
<mrow>
<mo>(</mo>
<mi>u</mi>
<mo>,</mo>
<mi>v</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mrow>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>m</mi>
<mo>=</mo>
<mn>0</mn>
</mrow>
<mrow>
<mi>m</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</munderover>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>n</mi>
<mo>=</mo>
<mn>0</mn>
</mrow>
<mrow>
<mi>N</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</munderover>
<msub>
<mi>s</mi>
<mn>1</mn>
</msub>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>,</mo>
<mi>n</mi>
<mo>)</mo>
</mrow>
<msubsup>
<mi>s</mi>
<mn>2</mn>
<mo>*</mo>
</msubsup>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>+</mo>
<mi>u</mi>
<mo>,</mo>
<mi>n</mi>
<mo>+</mo>
<mi>v</mi>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<msqrt>
<mrow>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>m</mi>
<mo>=</mo>
<mn>0</mn>
</mrow>
<mrow>
<mi>m</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</munderover>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>n</mi>
<mo>=</mo>
<mn>0</mn>
</mrow>
<mrow>
<mi>N</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</munderover>
<mo>|</mo>
<msub>
<mi>s</mi>
<mn>1</mn>
</msub>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>,</mo>
<mi>n</mi>
<mo>)</mo>
</mrow>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
</mrow>
</msqrt>
<msqrt>
<mrow>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>m</mi>
<mo>=</mo>
<mn>0</mn>
</mrow>
<mrow>
<mi>m</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</munderover>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>n</mi>
<mo>=</mo>
<mn>0</mn>
</mrow>
<mrow>
<mi>N</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</munderover>
<mo>|</mo>
<msub>
<mi>s</mi>
<mn>1</mn>
</msub>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>,</mo>
<mi>n</mi>
<mo>)</mo>
</mrow>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
</mrow>
</msqrt>
</mrow>
</mfrac>
</mrow>
Wherein, s1(m, n) and s2(m, n) represents the master image and auxiliary image respectively;M and N represents to be used to participate in correlation calculations
Window size;(u, v) represents the offset slided during correlation computations;Represent modulo operation respectively with * and take conjugate operation.
6. a kind of system of interference SAR complex image corregistration, including:
Rough registration module, the rough registration module are configured to master image and auxiliary image carrying out rough registration, inclined to obtain first
Shifting amount;
Interpolating module, the interpolating module are configured to according to the first offset obtained, by the master image and described auxiliary
The part of image rough registration is taken out, and carries out the interpolation of same multiple respectively;
Piecemeal module, the piecemeal module are configured to the image after interpolation with specific line-spacing, row away from carrying out piecemeal to obtain
Much a sub-blocks;
Computing module, the computing module are configured to, for each sub-block, calculate the second offset;And
Resampling module, the resampling module are configured to using the master image as reference, to each sub-block according to described the
Two offsets carry out resampling, to obtain the SAR complex patterns after registration.
7. system according to claim 6, wherein, the rough registration module is further configured to:
Obtain rough registration matching desired value, wherein it is described matching desired value in maximum corresponding to orientation and distance to
The offset of Pixel-level be first offset.
8. system according to claim 7, wherein, the matching desired value is related coefficient.
9. system according to claim 6, wherein, the computing module is configured to:
Second offset of each sub-block is obtained by calculating the compound correlative function of corresponding sub-block.
10. system according to claim 6, wherein, the complex correlation function is defined as:
<mrow>
<mi>&rho;</mi>
<mrow>
<mo>(</mo>
<mi>u</mi>
<mo>,</mo>
<mi>v</mi>
<mo>)</mo>
</mrow>
<mo>=</mo>
<mfrac>
<mrow>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>m</mi>
<mo>=</mo>
<mn>0</mn>
</mrow>
<mrow>
<mi>m</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</munderover>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>n</mi>
<mo>=</mo>
<mn>0</mn>
</mrow>
<mrow>
<mi>N</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</munderover>
<msub>
<mi>s</mi>
<mn>1</mn>
</msub>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>,</mo>
<mi>n</mi>
<mo>)</mo>
</mrow>
<msubsup>
<mi>s</mi>
<mn>2</mn>
<mo>*</mo>
</msubsup>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>+</mo>
<mi>u</mi>
<mo>,</mo>
<mi>n</mi>
<mo>+</mo>
<mi>v</mi>
<mo>)</mo>
</mrow>
</mrow>
<mrow>
<msqrt>
<mrow>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>m</mi>
<mo>=</mo>
<mn>0</mn>
</mrow>
<mrow>
<mi>m</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</munderover>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>n</mi>
<mo>=</mo>
<mn>0</mn>
</mrow>
<mrow>
<mi>N</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</munderover>
<mo>|</mo>
<msub>
<mi>s</mi>
<mn>1</mn>
</msub>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>,</mo>
<mi>n</mi>
<mo>)</mo>
</mrow>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
</mrow>
</msqrt>
<msqrt>
<mrow>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>m</mi>
<mo>=</mo>
<mn>0</mn>
</mrow>
<mrow>
<mi>m</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</munderover>
<munderover>
<mo>&Sigma;</mo>
<mrow>
<mi>n</mi>
<mo>=</mo>
<mn>0</mn>
</mrow>
<mrow>
<mi>N</mi>
<mo>-</mo>
<mn>1</mn>
</mrow>
</munderover>
<mo>|</mo>
<msub>
<mi>s</mi>
<mn>1</mn>
</msub>
<mrow>
<mo>(</mo>
<mi>m</mi>
<mo>,</mo>
<mi>n</mi>
<mo>)</mo>
</mrow>
<msup>
<mo>|</mo>
<mn>2</mn>
</msup>
</mrow>
</msqrt>
</mrow>
</mfrac>
<mo>-</mo>
<mo>-</mo>
<mo>-</mo>
<mrow>
<mo>(</mo>
<mn>1</mn>
<mo>)</mo>
</mrow>
</mrow>
In formula (1), s1 (m, n) and s2(m, n) represents the master image and auxiliary image respectively;M and N represents to be used to participate in correlation
The window size of calculating;(u, v) represents the offset slided during correlation computations;Represent modulo operation respectively with * and take conjugation to grasp
Make.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610973301.1A CN108022259A (en) | 2016-11-04 | 2016-11-04 | Interference SAR complex image corregistration method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201610973301.1A CN108022259A (en) | 2016-11-04 | 2016-11-04 | Interference SAR complex image corregistration method and system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN108022259A true CN108022259A (en) | 2018-05-11 |
Family
ID=62083667
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201610973301.1A Pending CN108022259A (en) | 2016-11-04 | 2016-11-04 | Interference SAR complex image corregistration method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108022259A (en) |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109658446A (en) * | 2018-10-30 | 2019-04-19 | 武汉珈和科技有限公司 | A kind of high-resolution remote sensing image geometrical registration method and device |
CN111640147A (en) * | 2020-06-04 | 2020-09-08 | 北京无线电测量研究所 | SAR subimage registration method, device and computer equipment for step-by-step frequency band splicing |
CN112068136A (en) * | 2020-09-14 | 2020-12-11 | 广东省核工业地质局测绘院 | Azimuth deformation monitoring method based on amplitude offset |
CN112288781A (en) * | 2018-08-22 | 2021-01-29 | 深圳市真迈生物科技有限公司 | Image registration method, apparatus and computer program product |
CN113781529A (en) * | 2021-09-24 | 2021-12-10 | 中国科学院精密测量科学与技术创新研究院 | Wide-area SAR complex image sequence rapid registration method adopting twice blocking strategy |
Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104933673A (en) * | 2015-06-26 | 2015-09-23 | 西安电子科技大学 | Interference SAR (Synthetic Aperture Radar) image precise registration method based on resolution search sub-pixel offset |
-
2016
- 2016-11-04 CN CN201610973301.1A patent/CN108022259A/en active Pending
Patent Citations (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN104933673A (en) * | 2015-06-26 | 2015-09-23 | 西安电子科技大学 | Interference SAR (Synthetic Aperture Radar) image precise registration method based on resolution search sub-pixel offset |
Non-Patent Citations (1)
Title |
---|
朱超猛: "高精度InSAR配准及相位解缠技术研究", 《中国优秀硕士学位论文全文数据库 信息科技辑,2016年第3期,I136-2421页》 * |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112288781A (en) * | 2018-08-22 | 2021-01-29 | 深圳市真迈生物科技有限公司 | Image registration method, apparatus and computer program product |
CN109658446A (en) * | 2018-10-30 | 2019-04-19 | 武汉珈和科技有限公司 | A kind of high-resolution remote sensing image geometrical registration method and device |
CN111640147A (en) * | 2020-06-04 | 2020-09-08 | 北京无线电测量研究所 | SAR subimage registration method, device and computer equipment for step-by-step frequency band splicing |
CN111640147B (en) * | 2020-06-04 | 2023-10-31 | 北京无线电测量研究所 | SAR sub-image registration method, device and computer equipment for step frequency band stitching |
CN112068136A (en) * | 2020-09-14 | 2020-12-11 | 广东省核工业地质局测绘院 | Azimuth deformation monitoring method based on amplitude offset |
CN113781529A (en) * | 2021-09-24 | 2021-12-10 | 中国科学院精密测量科学与技术创新研究院 | Wide-area SAR complex image sequence rapid registration method adopting twice blocking strategy |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN108022259A (en) | Interference SAR complex image corregistration method and system | |
Chen et al. | SAR and multispectral image fusion using generalized IHS transform based on à trous wavelet and EMD decompositions | |
US20100141508A1 (en) | Method and system for forming an image with enhanced contrast and/or reduced noise | |
CN101826157B (en) | Ground static target real-time identifying and tracking method | |
CN105425216A (en) | Image-segmentation-based registration method of polarized InSAR image in repeated passing | |
CN106204539B (en) | A method of the inverting urban architecture object sedimentation based on Morphological Gradient | |
CN104007439A (en) | Interferential circular SAR elevation estimation processing method | |
CN112711021B (en) | Multi-resolution InSAR (interferometric synthetic Aperture Radar) interactive interference time sequence analysis method | |
CN103886600B (en) | Continuous phase plate machining error identification method | |
CN108983231B (en) | Interferometric video measuring method based on video synthetic aperture radar | |
CN105180852B (en) | GB SAR deformation monitoring methods based on triple steppings | |
Ferro et al. | A novel technique for the automatic detection of surface clutter returns in radar sounder data | |
CN103809180A (en) | Azimuth pre-filtering processing method for Interferometric Synthetic Aperture Radar (InSAR) topographic survey | |
Douxchamps et al. | Stereoscopic and velocimetric reconstructions of the free surface topography of antidune flows | |
CN108876829A (en) | SAR high registration accuracy method based on Nonlinear Scale Space Theory and radial basis function | |
CN112505647A (en) | Moving target azimuth speed estimation method based on sequential sub-image sequence | |
Peterson et al. | Registration of multi-frequency SAR imagery using phase correlation methods | |
CN115291215B (en) | Long-time-sequence two-dimensional deformation rapid resolving method based on lifting orbit SAR satellite | |
Refice et al. | On the use of anisotropic covariance models in estimating atmospheric DInSAR contributions | |
Kirk et al. | Evaluating stereo dtm quality at jezero crater, mars with hrsc, ctx, and hirise images | |
CN111222544A (en) | Ground simulation test system for influence of satellite flutter on camera imaging | |
Imperatore et al. | Contribution of super resolution to 3D reconstruction from pairs of satellite images | |
Kim et al. | An efficient automatic georegistration technique for hige-resolution spaceborne SAR image fusion | |
CN108010070B (en) | A kind of scene-change detecting method and device based on the sequential image of SAR | |
Wang et al. | Motion compensation technique for UGV based forward looking InSAR |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180511 |
|
RJ01 | Rejection of invention patent application after publication |