CN104766280B - A kind of Quality Map phase unwrapping method based on heapsort - Google Patents
A kind of Quality Map phase unwrapping method based on heapsort Download PDFInfo
- Publication number
- CN104766280B CN104766280B CN201510136369.XA CN201510136369A CN104766280B CN 104766280 B CN104766280 B CN 104766280B CN 201510136369 A CN201510136369 A CN 201510136369A CN 104766280 B CN104766280 B CN 104766280B
- Authority
- CN
- China
- Prior art keywords
- phase
- solution
- image
- row
- point
- 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.)
- Expired - Fee Related
Links
Landscapes
- Image Processing (AREA)
Abstract
The invention discloses a kind of Quality Map phase unwrapping method based on heapsort, first pass through flat earth and filtering obtains treating that solution twines phase image, then from treat solution twine phase image in extract phase masses figure, and phase masses figure is ranked up using heapsort method, determine that treating that solution is twined treats that the solution that solution is twined a little is twined smoothly in phase image, finally, order is twined according to solution treat solution and twine a solution and twine, finished until view picture treats that solution twines phase image solution and twined.The present invention is using treating that solution twines the quality information guiding of phase image and twined as treating that solution twines a progress Integral Solution, the method of addition heapsort is ranked up to phase masses figure again simultaneously, so as to which the time required to shortening phase unwrapping, final quick, high-precision completion phase unwrapping generates phase unwrapping image.
Description
Technical field
The invention belongs to InSAR technical field of image processing, more specifically, it is related to a kind of quality based on heapsort
Figure phase unwrapping method.
Background technology
InSAR (interferometric Synthetic Aperture Radar, interference synthetic aperture radar) technology
Initial development purpose is that, in order to carry out mapping, the most successfully application is exactly to obtain high accuracy DEM (Digital
Elevation Model, digital elevation model) data.It is main and with the development of technology, InSAR application fields are constantly being expanded
Including：Ground Deformation detection, moving-target detection, marine charting, forest mapping, flood monitor, Traffic monitoring and glacier are ground
Study carefully.Spaceborne InSAR systems are one and are related to that field is wide, level is more, highly complex big system, its corresponding processing links
It is complex.Spaceborne InSAR system datas processing key step includes：InSAR complex image corregistrations, interferometric phase filtering, interference
Phase unwrapping and DEM etc..Because processing procedure is all linked with one another, the error of any one link processing all will be directly brought into down
In the processing of one step.
InSAR general principle is exactly, while observing or by parallel observation twice, to obtain by two antennas
Ground is presented on the phase difference on SAR (Synthetic Aperture Radar, synthetic aperture radar) image, so that inverting is obtained
To the height value of every.And in actual physical process, the phase difference being presented on image is the wound of true phase difference.
That is, the picture point that actually in interference pattern and ground is associated thinks the π of mould 2.Elevation is obtained in order to inverting, just must
2 π integral multiple must be added to the every bit of SAR image, so as to try to achieve its real phase difference value.Thus phase unwrapping is
One of committed step of InSAR imagings, its order of accuarcy directly influences the precision of DEM or Ground Deformation.
Currently, existing interferometric phase unwrapping method is broadly divided into two major classes：Route complexity, Minimum-Norm Method.Path
The elementary tactics of tracing is that the suitable Integral Solution of selection twines path, prevents phase error from transmitting.The representative of this algorithm is
The Branch cut that Goldstein was proposed in 1988.Branch cut is twined by setting branch tangent line to carry out solution to the residue points of isolation.This
Afterwards, many scholars improve this algorithm.Huntley connects the residue points of opposed polarity using nearest neighbor algorithm, so that
To the branch tangent line of shortest length.Buckland utilizes the Hungary Algorithm in integer programming to connect positive and negative residue points, obtains pairing
The Global optimal solution of branch tangent line.Flynn proposes Mask-Cut algorithms, sets branch tangent line by Quality Map guiding, improves branch
The degree of accuracy that tangent line is set.Xu proposes that algorithm of region growing is twined for solution, and phase range is separated according to quality height, and from
High-quality region is twined to low quality region solution.The problem of Minimum-Norm Method is substantially by phase unwrapping is converted into mathematically
Minimum norm problem, by the L for finding the error for making phase gradient and winding phasepReach that minimum carrys out solution and twines phase.Takajo
Propose to carry out phase unwrapping using least square method at first, and solved using FFT, this method does not introduce Quality Map.Then,
Ghigilia proposes the least square method using Quality Map, improves understanding and twines precision.
It is worth noting that, on InSAR and D-InSAR, (Differential InSAR differential interferometries synthesize hole at present
Footpath radar) data processing phase unwrapping algorithm species it is although more, but universality is poor.Each method has itself
Limitation.Branch cut solution twines speed soon, and still " branch tangent line " setting is improper can cause error propagation.Least square method is easily produced
Error of overall importance, and calculating speed is slower.And because noise and lack sampling cause the problem of phase occurs inconsistent always phase
The difficult point that position solution is twined.It is, therefore, necessary to be improved to existing algorithm, required precision can be met and quickly solution can twine again by finding
Algorithm.
The content of the invention
It is an object of the invention to overcome the deficiencies of the prior art and provide a kind of Quality Map phase unwrapping based on heapsort
Method, to realize that solution takes into account the purpose of speed and precision during twining.
For achieving the above object, a kind of Quality Map phase unwrapping method based on heapsort of the present invention, its feature exists
In comprising the following steps：
(1), SAR image is pre-processed
By two width SAR image conjugate multiplications, interference image is obtained, then flat earth is carried out successively with interference image
Value filtering, obtains treating that solution twines phase image A；
(2), from treating that solution extracts phase masses figure P in twining phase image A
(2.1), according to the phase image A that solution is twined is treated, row winding phase gradient image array Ax is got respectively and row are twined
Around phase gradient image array Ay；
(2.2) institute in row winding phase gradient image array Ax, is traveled through a little, to calculate and treat that solution twines phase image A distances
To the phase masses m of everyi,j；
Wherein, γ represents the size of window；SymbolRepresent round numbers；Axi,jExpression, which is expert at, winds phase gradient image moment
In battle array Ax, the i-th row, the pixel value of jth row point, i=1,2 ..., m, m is row winding phase gradient image array Ax line number,
J=1,2 ..., n, n wind phase gradient image array Ax columns for row；
Again by phase masses m a littlei,jConstitute row winding phase masses figure Ax*；
(2.3), institute a little, calculates and treats that solution twines phase image A orientation in traversal row winding phase gradient image array Ay
To the phase masses n of everyi,j；
Wherein, Ayi,jRepresent in row winding phase gradient image array Ay, the i-th row, the pixel value of jth row point；
Again by phase masses n a littlei,jConstitute row winding phase masses figure Ay*；
(2.4) phase masses figure P, is calculated
P=α1Ax*+α2Ay*；
Wherein, α1And α2For factor of influence, and meet α1+α2=1；
(3) most raft, is set up
(3.1), the optional some starting point twined as solution in treating that solution twines image A, and by the starting point labeled as having solved
Twine point Q；
(3.2), centered on solution twines point Q, finding distance, solution twines the nearest points of point Q, i.e. Q neighborhoods of a point, then judge
Whether solution is twined for certain point in the neighborhood, if the solution is twined, gives up the point；If the no solution of point is twined, it is put into storehouse,
Until having judged the institute in neighborhood a little；
(3.3), all points in storehouse have been generated according to quality size in Quality Map P using computer science heapsort method
Full binary tree, sets up most raft；
(4), the root node progress solution to most raft is twined
(4.1), the root node of most raft is exchanged with last leaf node, while the root node is removed into most raft,
And corresponding click-through line phase solution is twined in treating that solution twines image A to the root node；
(4.2), remaining node it will be retained in storehouse in most raft, then with the root node of most raft in step (4.1)
Starting point is twined as the solution chosen in step (3.1), and returns to step (3.2), is repeated the above steps, until treating that solution twines image
All point solutions twine completion in A.
What the goal of the invention of the present invention was realized in：
Quality Map phase unwrapping method of the invention based on heapsort, first passes through flat earth and filtering obtains waiting to solve
Phase image is twined, then from treating that solution extracts phase masses figure in twining phase image, and phase masses figure is entered using heapsort method
Row sequence, that is, determine that treating that solution is twined treats that the solution that solution is twined a little is twined smoothly in phase image, finally, is twined according to solution and sequentially treats solution and twine a little
Solution is twined, and is finished until view picture treats that solution twines phase image solution and twined.The present invention utilizes the quality information guiding picture for treating that solution twines phase image
Treat that solution twines a progress Integral Solution and twined, while the method for addition heapsort is ranked up to phase masses figure again, so as to shorten phase solution
The time required to twining, final quick, high-precision completion phase unwrapping generates phase unwrapping image.
Meanwhile, the Quality Map phase unwrapping method of the invention based on heapsort also has the advantages that：
(1), by orientation and introducing from distance to factor of influence, can it is determined that during phase masses figure preferably with reality
The influence of noise of border imaging process middle-range descriscent and orientation matches, and so as to preferably instruct phase unwrapping, lifts phase
The degree of accuracy that position solution is twined；
(2), Heap algorithm is a kind of sorting by computer algorithm of stabilization；The method of heapsort and specific data without
Close, thus after the phase masses figure of SAR image is obtained, it is not necessary to worry the irregular conditions of phase masses diagram data；Although heap
Sequence advantage in the case where data are small is not obvious, but is but highly suitable for the sequence of larger data, and SAR image is substantially all
It is the great image of data volume, thus the processing speed of SAR image can be preferably lifted using the method for heapsort；
(3), because maximum heap meets the property that root node is more than left and right leaf node, thus its root node is maximum all the time
Value, only needs to take out root node when choosing phase masses peak every time, also, the characteristics of due to heapsort itself,
When adding new neighborhood point every time into storehouse, generation most raft that can be quickly.
Brief description of the drawings
Fig. 1 is the flow chart of the Quality Map phase unwrapping method of the invention based on heapsort；
Fig. 2 is that Iranian bar nurse area is carried out pretreated treating that solution twines phase image；
Fig. 3 is the explanation figure for asking for row winding phase gradient image array；
Fig. 4 is to treat that solution twines the phase masses calculating schematic diagram of phase image middle-range descriscent；
Fig. 5 is to treat the Quality Map that solution is extracted in twining phase image from Iranian bar nurse area；
Fig. 6 is the schematic diagram for choosing the starting point that solution is twined；
Fig. 7 is the explanation figure by taking known four points of vertex neighborhood as an example；
Fig. 8 is the maximum heapsort process schematic to four neighborhood points in Fig. 7；
Fig. 9 is the schematic diagram for adding new neighborhood point and setting up most raft；
Figure 10 is the phase unwrapping result figure based on Quality Map heapsort method shown in Fig. 2；
Figure 11 is two width SAR image sectional drawings of Phoenix Arizona State Phoenix；
Figure 12 is the interference pattern that the SAR image shown in Figure 11 is obtained after conjugate multiplication；
Figure 13 is the Quality Map that the interference pattern shown in Figure 12 is obtained by processing；
Figure 14 is knot figure of the Quality Map after phase unwrapping shown in Figure 13；
Figure 15 is the elevation map in Phoenix Arizona State Phoenix area.
Embodiment
The embodiment to the present invention is described below in conjunction with the accompanying drawings, so as to those skilled in the art preferably
Understand the present invention.Requiring particular attention is that, in the following description, when known function and design detailed description perhaps
When can desalinate the main contents of the present invention, these descriptions will be ignored herein.
Embodiment
Fig. 1 is the flow chart of the Quality Map phase unwrapping method of the invention based on heapsort.
In the present embodiment, as shown in figure 1, the Quality Map phase unwrapping method based on heapsort of the invention, including it is following
Step：
S1, acquisition interference image
By two width SAR image conjugate multiplications, i.e., the corresponding pixel conjugate multiplication of two width SAR images is interfered respectively
Image；
In the present embodiment, same target area is imaged twice by satellite-borne SAR, obtains two width SAR images.
S2, to interference image carry out flat earth, that is, obtain coherent image；
S3, to coherent image carry out medium filtering, obtain treating that solution twines phase image A；
It is regional (BAM) in Iranian bar nurse in the present embodiment, size is located for 2823 × 1973 ASAR datagrams by pre-
After reason, that is, step S1~step S3 processing is sequentially passed through, still with the presence of interference fringe, as shown in Fig. 2 this interference fringe is produced
Raw the reason for is the result of phase winding.In order to carry out the reconstruction extraction of depth displacement, must just phase be carried out to this image
Position solution is twined, and recovers real phase difference.
S4, from treating that solution extracts phase masses figure P in twining phase image A
S4.1), according to the phase image A that solution is twined is treated, row winding phase gradient image array Ax is got respectively and row are twined
Around phase gradient image array Ay；
Specific method is：
To treat that solution twines the pixel value of all pixels point in phase image A and is mapped to [- 0.5 ,+0.5] from [0,2 π], to reflecting
Penetrate image A1；
Ergodic Maps image A1, mapping graph is asked for as A1In often row in two neighboring pixel pixel value difference Δ (i, j),
I.e.：
Δ (i, j)=A (i, j+1)-A (i, j)
Judge whether Δ (i, j) belongs to [- 0.5 ,+0.5]；
Wherein, i be row coordinate, i=1,2 ..., m, m is to treat that solution twines the line number of image；J be row coordinate, j=1,2 ..., n,
N is to treat that solution twines the columns of image；
In the present embodiment, as shown in figure 3, traveling through this image, calculating Δ ' (i, j).
A), calculate Δ ' (1,1)：Δ (1,1)=0.2-0.1=0.1 is first calculated, because -0.5≤0.1≤0.5, therefore
Δ ' (1,1)=Δ (1,1)=0.1.
B), calculate Δ ' (1,2)：Δ (1,2)=- 0.4-0.2=-0.6 is first calculated, because -0.6 ＜ -0.5, therefore
Δ ' (1,2)=Δ (1,2)+1=0.4.
C), calculate Δ ' (1,3)：Δ (1,3)=0.5- (- 0.4)=0.9 is first calculated, because 0.9 >=0.5, therefore Δ '
(1,3)=Δ (1,3) -1=-0.1.
By all pixel value difference Δs ' (i, j) constitute the matrixes of m rows n-1 row, then the value of the (n-1)th row in matrix be assigned to
N-th row, constitute row winding phase gradient image array Ax；
The method for obtaining row winding phase gradient image array Ay is：
Ergodic Maps image A1, mapping graph is asked for as A1The pixel value difference of two neighboring pixel in middle each column
I.e.：
JudgeWhether belong to [- 0.5 ,+0.5]；
The computational methods of the pixel value difference of pixel herein are identical with row winding phase gradient image array, are only to ask
The difference of pixel in each column, therefore repeat no more.
By all pixel value differencesConstitute the matrix of m-1 rows n row, then the value of m-1 rows in matrix is assigned to the
M rows, constitute row winding phase gradient image array Ay.
S4.2 institute in row winding phase gradient image array Ax), is traveled through a little, to calculate and treat that solution twines phase image A distances
To the phase masses m of everyi,j；
Wherein, γ represents the size of window；SymbolRepresent round numbers；Axi,jExpression, which is expert at, winds phase gradient image
In matrix A x, the i-th row, the pixel value of jth row point, i=1,2 ..., m, m wind phase gradient image array Ax row for row
Number, j=1,2 ..., n, n winds phase gradient image array Ax columns for row；
Again by phase masses m a littlei,jConstitute row winding phase masses figure Ax*；
In the present embodiment, as shown in figure 4, each grid represents a point in matrix A x, solid line grid constitutes matrix
Ax, dotted line format is the point supplemented outside matrix A x boundary points, and the corresponding numerical value of dotted line format is 0.
It is illustrated below exemplified by putting 8 three points of 1, point 2 and point and being divided into three kinds of situations.
A) boundary point, is asked for：The distance of grid 1 (i.e. the 1st row the 1st row) is to phase masses, then centered on grid 1, γ
During=3*3, periphery comprising five dotted line formats and 2,7,8 three solid line grids, can the side of calculating using above-mentioned computing formula
Phase masses m of the lattice 1 in 3*3 windows1,1。
B) boundary point, is asked for：The distance of grid 2 (i.e. the 1st row the 2nd row) is to phase masses, then centered on grid 2, γ
During=3*3, periphery comprising three dotted line formats and 1,3,7,8,9 five solid line grids, can equally calculate grid 2 in 3*3
The m of phase masses in window1,2。
C) non-boundary point, is asked for：The distance of grid 8 (i.e. the 2nd row the 2nd row) is to phase masses, then centered on grid 8,
During γ=3*3, periphery only comprising 1,2,3,7,9,13,14,15 8 solid line grids, can also equally calculate grid 8 in 3*3
The m of phase masses in window2,2。
S4.3), institute a little, calculates and treats that solution twines phase image A orientation in traversal row winding phase gradient image array Ay
To the phase masses n of everyi,j；
Wherein, Ayi,jRepresent in row winding phase gradient image array Ay, the i-th row, the pixel value of jth row point；
This, the phase masses n that orientation is each puti,jComputational methods it is identical to calculating with distance, will not be repeated here；
Again by phase masses n a littlei,jConstitute row winding phase masses figure Ay*；
S4.4 phase masses figure P), is calculated
P=α1Ax*+α2Ay*；
Wherein, α1And α2For factor of influence, and meet α1+α2=1；In the present embodiment, α1=α2=0.5, such as Fig. 5 institutes
Show, to image regional BAM after above-mentioned steps are handled, you can obtain phase masses figure P；
S5, determination treat that solution twines the starting point that the solution in image A is twined
The starting point optionally a little twined in treating that solution twines image A as solution, it is high that the starting point is located at phase masses as far as possible
Region, and the starting point is twined into point Q labeled as solution；
S6, the non-solution in Q vertex neighborhoods is twined a little it is put into storehouse
Centered on solution twines point Q, finding distance, solution twines the nearest points of point Q, i.e. Q neighborhoods of a point, then judge the neighborhood
Whether solution is twined for middle certain point, if the solution is twined, and gives up the point；If the no solution of point is twined, it is put into storehouse, until sentencing
The institute broken in neighborhood is a little；
In the present embodiment, as shown in fig. 6, a point in the lattice representative image of each square；In image
Point through being twined by solution marks light grey, not twined by solution point mark Dark grey, and the maximum point of selection wherein quality is twined as solution
Starting point, and by this point labeled as light grey, the i.e. signified point of arrow in Fig. 5；
S7, foundation most raft
Complete y-bend is generated according to quality size in Quality Map P to all points in storehouse using computer science heapsort method
Tree, sets up most raft；
Its specific steps is described as follows below：
S7.1 most raft) is set up
If the array of adjacent column is { r in storehouse1,r2,...,rn, n is the number of element in array, and most raft needs to meet
Following property：
OrWherein
In the present embodiment, the array of adjacent column is：A [4]={ 0.89,0.93,0.81,0.85 }, it sets up most raft
Step is：
S7.1.2), according to array a [4]={ 0.89,0.93,0.81,0.85 }, a binary tree is generated at random；
In the present embodiment, as shown in Fig. 7 (c), to originate in vertex neighborhood exemplified by four points, a y-bend is generated at random
Tree, shown in binary tree structure such as Fig. 7 (a), then this storage organization of four points in the calculation, i.e. adjacent column, such as Fig. 7 (b) institutes
Show；
S7.2.1), adjusted since last non-leaf nodes of binary tree；Its regulation rule is：Take this non-leaf
Maximum of points and this non-leaf nodes are exchanged in node, left child node and right child node, until all non-leaf nodes all
Untill the definition for meeting most raft, that is, generate complete binary tree；
In the present embodiment, as shown in figure 8, it is 0.93, its left sub- section to choose last non-leaf nodes in 8 (a)
Point is 0.85, without right child node.Take after wherein maximum of points is exchanged with the non-leaf nodes, be still the binary tree shown in 8 (a).
It is 0.89 to continue to find non-leaf nodes forward, and its left child node is 0.93, and right child node is 0.81, takes wherein maximum of points
After 0.93 exchanges with the non-leaf nodes, complete binary tree is generated, shown in such as Fig. 8 (b)；
S7.2), shown in such as Fig. 8 (c), the root node 0.93 of most raft is exchanged with last leaf node 0.85, and
Root node 0.93 is removed into most raft；
S7.3), shown in such as Fig. 8 (d), remaining node is retained in storehouse；
S7.4), shown in such as Fig. 9 (a), when new node adds storehouse, i.e., when node 0.9,0.7,0.6 adds storehouse,
Then and repeat step S7.1)~step S7.3), until step S7.3) in without remaining node untill, as shown in Fig. 9 (b)；
S8, the root node to most raft carry out solution and twined
S8.1), the root node of most raft is exchanged with last leaf node, while the root node is removed into most raft,
And corresponding click-through line phase solution is twined in treating that solution twines image A to the root node；
S8.2), remaining node will be retained in storehouse in most raft, then with step S8.1) in most raft root node
Starting point is twined as the solution chosen in step S5, and returns to step S5, is repeated the above steps, until treating that solution twines institute in image A
Some point solutions twine completion.
In the present embodiment, to twine phase diagram after phase unwrapping processing of the invention after solution shown in Fig. 2, its result
As shown in Figure 10, exist in image without obvious interference fringe.
Example
In general, the present invention needs the two width SAR datas not changed according to landform to carry out a series of processing
Obtain.And due to Iranian bar nurse area obtain two width SAR images be respectively in BEFORE AND AFTER EARTHQUAKE, thus its disentanglement fruit can not
Enough compared with actual landform.For this reason, it may be necessary to look for more particularly suitable spaceborne data.
Figure 11 is two width SAR image sectional drawings of Phoenix Arizona State Phoenix；
Figure 12 is the interference pattern that the SAR image shown in Figure 11 is obtained after conjugate multiplication；
Figure 13 is the Quality Map that the interference pattern shown in Figure 12 is obtained by processing；
Figure 14 is result figure of the Quality Map after phase unwrapping shown in Figure 13；
Figure 15 is the elevation map in Phoenix Arizona State Phoenix area.
In the present embodiment, as shown in figure 11, two width RADARSAT-2 images of this area are taken in May, 2008 respectively
No. 4 and on May 28th, 2008, picture size are 11471 × 14206.And there are three mountains this area, can more intuitively test
Card, as shown in figure 15.
Good disentanglement fruit should show the elementary contour on three mountains.Due to the limitation of calculator memory, experiment interception
Coordinate (6000,5000) arrives 3000 × 6000 images of coordinate (9000,11000) in image, and the image just covers view picture
Three mountains in image.Figure 12 is interference pattern of this two images after conjugate multiplication；Figure 13 be the image after treatment
Obtained Quality Map, Figure 14 twines phase diagram for final solution；Pass through comparison diagram 14 and Figure 15, here it is apparent that three in Figure 15
The profile on mountain can be high-visible in fig. 14.
Although illustrative embodiment of the invention is described above, in order to the technology of the art
Personnel understand the present invention, it should be apparent that the invention is not restricted to the scope of embodiment, to the common skill of the art
For art personnel, as long as various change is in the spirit and scope of the present invention that appended claim is limited and is determined, these
Change is it will be apparent that all utilize the innovation and creation of present inventive concept in the row of protection.
Claims (5)
1. a kind of Quality Map phase unwrapping method based on heapsort, it is characterised in that comprise the following steps：
(1), SAR image is pre-processed
By two width SAR image conjugate multiplications, interference image is obtained, then carries out to interference image flat earth and intermediate value filter successively
Ripple, obtains treating that solution twines phase image
(2), from treating that solution twines phase imageMiddle extraction phase masses figure P
(2.1), basis treats the phase image that solution is twinedRow winding phase gradient image array Ax and row winding phase are got respectively
Potential gradient image array Ay；
(2.2) institute in row winding phase gradient image array Ax, is traveled through a little, to calculate and treat that solution twines phase imageDistance to
The phase masses m of everyi,j；
Wherein, γ represents the size of window；SymbolRepresent round numbers；Axi,jExpression, which is expert at, winds phase gradient image array Ax
In the i-th row, the pixel value of jth row point, i=1,2 ..., m, m is row winding phase gradient image array Ax line number, j=1,
2 ..., n, n wind phase gradient image array Ax columns for row；
Again by phase masses m a littlei,jConstitute row winding phase masses figure Ax*；
(2.3), institute a little, calculates and treats that solution twines phase image in traversal row winding phase gradient image array AyOrientation
The phase masses n of everyi,j；
Wherein, Ayi,jRepresent the i-th row, the pixel value of jth row point in row winding phase gradient image array Ay；
Again by phase masses n a littlei,jConstitute row winding phase masses figure Ay*；
(2.4) phase masses figure P, is calculated
P=α1Ax*+α2Ay*；
Wherein, α1And α2For factor of influence, and meet α1+α2=1；
(3) most raft, is set up
(3.1), treating that solution twines phase imageIn the optional some starting point twined as solution, and by the starting point labeled as having solved
Twine point Q；
(3.2), centered on solution twines point Q, finding distance, solution twines the nearest points of point Q, i.e. Q neighborhoods of a point, then judge the neighbour
Whether solution is twined for certain point in domain, if the solution is twined, gives up the point；If the no solution of point is twined, it is put into storehouse, until
Judged in neighborhood institute a little；
(3.3) complete two, are generated according to quality size in Quality Map P to all points in storehouse using computer science heapsort method
Fork tree, sets up most raft；
(4), the root node progress solution to most raft is twined
(4.1), the root node of most raft is exchanged with last leaf node, while the root node is removed into most raft, and it is right
The root node is treating that solution twines imageIn it is corresponding click-through line phase solution twine；
(4.2), remaining node will be retained in storehouse in most raft, then using the root node of most raft in step (4.1) as
The solution chosen in step (3.1) twines starting point, and returns to step (3.2), repeats the above steps, until treating that solution twines imageIn
All point solutions twine completion.
2. the Quality Map phase unwrapping method according to claim 1 based on heapsort, it is characterised in that described step
(2.1) in, the method for obtaining row winding phase gradient image array Ax is：
It will treat that solution twines phase imageThe pixel value of middle all pixels point is mapped to [- 0.5 ,+0.5] from [0,2 π], obtains mapping graph
As A1；
Ergodic Maps image A1, mapping graph is asked for as A1In often row in two neighboring pixel pixel value difference Δ (i, j), i.e.,：
Δ (i, j)=A (i, j+1)-A (i, j)
Judge whether Δ (i, j) belongs to [- 0.5 ,+0.5]；
Wherein, i be row coordinate, i=1,2 ..., m, m is to treat that solution twines the line number of image；J is row coordinate, j=1,2 ..., n, and n is
Treat that solution twines the columns of image；
By all pixel value difference Δs ' (i, j) constitute the matrixes of m rows n-1 row, then the value of the (n-1)th row in matrix be assigned into n-th
Row, constitute row winding phase gradient image array Ax；
The method for obtaining row winding phase gradient image array Ay is：
Ergodic Maps image A1, mapping graph is asked for as A1The pixel value difference of two neighboring pixel in middle each columnI.e.：
JudgeWhether belong to [- 0.5 ,+0.5]；
By all pixel value differencesThe matrix of m-1 rows n row is constituted, then the value of m-1 rows in matrix is assigned to m rows,
Constitute row winding phase gradient image array Ay.
3. the Quality Map phase unwrapping method according to claim 1 based on heapsort, it is characterised in that described step
(3.1) in, the starting point selection that solution is twined is in the high region of phase masses.
4. the Quality Map phase unwrapping method according to claim 1 based on heapsort, it is characterised in that described step
(3.3) in, the step of setting up most raft is：
4.1) most raft is set up
If the array of adjacent column is in storehouse For the number of element in array, and it meets following property：
OrWherein
Then setting up most raft step is：
(4.1.1), according to arrayA binary tree is generated at random；
(4.1.2), the adjustment since last non-leaf nodes of binary tree；Its regulation rule is：Take root node, Zuo Zijie
Maximum is exchanged with root node in point and right child node, until the definition that all non-leaf nodes all meet most raft is
Only, that is, complete binary tree is generated；
4.2), the root node of most raft and last leaf node are exchanged, and root node is removed into most raft；
4.3), remaining node is retained in storehouse；
4.4) when new node, being added into storehouse, and repeat step 4.1)~step 4.3), until step 4.3) in without residue
Untill node；
5. the Quality Map phase unwrapping method according to claim 4 based on heapsort, it is characterised in that described root section
Point is quality highest point in the storehouse.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510136369.XA CN104766280B (en) | 2015-03-26 | 2015-03-26 | A kind of Quality Map phase unwrapping method based on heapsort |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201510136369.XA CN104766280B (en) | 2015-03-26 | 2015-03-26 | A kind of Quality Map phase unwrapping method based on heapsort |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN104766280A CN104766280A (en) | 2015-07-08 |
| CN104766280B true CN104766280B (en) | 2017-07-18 |
Family
ID=53648092
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201510136369.XA Expired - Fee Related CN104766280B (en) | 2015-03-26 | 2015-03-26 | A kind of Quality Map phase unwrapping method based on heapsort |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104766280B (en) |
Families Citing this family (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105719253A (en) * | 2016-01-20 | 2016-06-29 | 桂林电子科技大学 | Kalman filtering phase unwrapping method having heapsort function in embedded manner |
| CN105738896B (en) * | 2016-02-25 | 2018-06-19 | 内蒙古工业大学 | A kind of ground SAR Slope with multi-step interferometric phase unwrapping methods and device |
| CN107092022B (en) * | 2017-04-21 | 2019-08-02 | 哈尔滨工业大学 | Region filter quality based on InSAL guides phase unwrapping method |
| CN107341136B (en) * | 2017-06-30 | 2021-06-18 | 广州酷狗计算机科技有限公司 | Method and device for generating sequence index number and storage medium |
| CN109472834B (en) * | 2018-10-23 | 2023-04-14 | 桂林电子科技大学 | Kalman filtering phase expansion method based on wavelet transformation |
| CN109884634B (en) * | 2019-03-20 | 2020-08-04 | 中南大学 | InSAR phase unwrapping method and system based on hierarchical network construction mode |
| CN111562579B (en) * | 2020-03-05 | 2022-03-04 | 华北水利水电大学 | MIMO InSAR phase unwrapping method based on integer programming model |
| CN112505782B (en) * | 2020-12-23 | 2021-07-30 | 中国石油大学(北京) | Interference reference plane reconstruction method and system for radiation mode correction in four-dimensional earthquake |
| CN113311433B (en) * | 2021-05-28 | 2022-08-02 | 北京航空航天大学 | InSAR interferometric phase two-step unwrapping method combining quality map and minimum cost flow |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6107953A (en) * | 1999-03-10 | 2000-08-22 | Veridian Erim International, Inc. | Minimum-gradient-path phase unwrapping |
| CN101655357A (en) * | 2009-09-11 | 2010-02-24 | 南京大学 | Method for acquiring phase gradient correlated quality diagram for two-dimensional phase unwrapping |
| CN102096910A (en) * | 2011-01-25 | 2011-06-15 | 南京大学 | Method for acquiring weighted gradient quality map for two-dimensional phase unwrapping |
| CN103325092A (en) * | 2013-03-13 | 2013-09-25 | 中国科学院电子学研究所 | Method and device for generating two-dimensional phase disentanglement quality picture |
-
2015
- 2015-03-26 CN CN201510136369.XA patent/CN104766280B/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6107953A (en) * | 1999-03-10 | 2000-08-22 | Veridian Erim International, Inc. | Minimum-gradient-path phase unwrapping |
| CN101655357A (en) * | 2009-09-11 | 2010-02-24 | 南京大学 | Method for acquiring phase gradient correlated quality diagram for two-dimensional phase unwrapping |
| CN102096910A (en) * | 2011-01-25 | 2011-06-15 | 南京大学 | Method for acquiring weighted gradient quality map for two-dimensional phase unwrapping |
| CN103325092A (en) * | 2013-03-13 | 2013-09-25 | 中国科学院电子学研究所 | Method and device for generating two-dimensional phase disentanglement quality picture |
Non-Patent Citations (7)
| Title |
|---|
| 《Multilevel quality-guided phase unwrapping algorithm for real-time three-dimensional shape reconstruction》;Song Zhang等;《APPLIED OPTICS》;20070101;第46卷(第1期);全文 * |
| 《New method of generating phase quality map for InSAR》;ZHONG Heping等;《Journal of Remote Sensing》;20110430;第15卷(第4期);全文 * |
| 《一种基于新质量图引导的干涉相位快速解缠方法》;黄柏圣等;《系统仿真学报》;20100228;第22卷(第2期);全文 * |
| 《一种基于质量指导的InSAR相位解缠快速实现方法》;李芳芳等;《雷达学报》;20120630;第1卷(第2期);全文 * |
| 《基于构造边的精确快速相位解缠算法》;陆军等;《光电子 激光》;20150131;第26卷(第1期);全文 * |
| 《相位质量图在二维相位解缠中的应用探讨》;孟智强等;《计算机应用与软件》;20060228;第23卷(第2期);全文 * |
| 《质量图引导的相位解缠在不同地形的对比分析》;刘伟科等;《测绘地理信息》;20130630;第38卷(第3期);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN104766280A (en) | 2015-07-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN104766280B (en) | A kind of Quality Map phase unwrapping method based on heapsort | |
| CN103198480B (en) | Based on the method for detecting change of remote sensing image of region and Kmeans cluster | |
| CN104318270A (en) | Land cover classification method based on MODIS time series data | |
| CN110348324B (en) | Flood real-time flooding analysis method and system based on remote sensing big data | |
| CN114581784B (en) | Construction method of long-time-sequence yearly mangrove remote sensing monitoring product | |
| CN112949414B (en) | Intelligent surface water body drawing method for wide-vision-field high-resolution six-satellite image | |
| Xia et al. | High-resolution mapping of water photovoltaic development in China through satellite imagery | |
| CN110415265A (en) | Terraced fields extraction method based on unmanned plane high accuracy DEM data | |
| Kang et al. | Land use and land cover change and its impact on river morphology in Johor River Basin, Malaysia | |
| CN103268587A (en) | Method for obtaining urban building land information by means of imitated building land index | |
| CN104835196A (en) | Vehicular infrared image colorization and three-dimensional reconstruction method | |
| Chen et al. | Mapping three-dimensional morphological characteristics of tidal salt-marsh channels using UAV structure-from-motion photogrammetry | |
| CN109784205A (en) | A kind of weeds intelligent identification Method based on multispectral inspection image | |
| CN115393741A (en) | Ground feature classification artificial intelligence identification method and system based on unmanned aerial vehicle low-altitude sampling | |
| Liu et al. | A shape-matching cropping index (CI) mapping method to determine agricultural cropland intensities in China using MODIS time-series data | |
| CN116012723A (en) | Wetland type extraction method and device based on time sequence remote sensing image and electronic equipment | |
| CN117611996A (en) | Grape planting area remote sensing image change detection method based on depth feature fusion | |
| CN114462579B (en) | Soil organic carbon content estimation method based on terrain and remote sensing data | |
| CN114494586B (en) | Lattice projection deep learning network broadleaf branch and leaf separation and skeleton reconstruction method | |
| CN116612245A (en) | Beach topography construction method, system and storage medium based on video image | |
| Chen et al. | Neural classification of SPOT imagery through integration of intensity and fractal information | |
| Yan et al. | Examining the expansion of Spartina alterniflora in coastal wetlands using an MCE-CA-Markov model | |
| Yuan et al. | Capturing small objects and edges information for cross-sensor and cross-region land cover semantic segmentation in arid areas | |
| CN112966657A (en) | Remote sensing automatic classification method for large-scale water body coverage | |
| Litinsky | Visualization of Landsat image spectral space as a method of boreal ecosystems geomatic modeling (on the example of Eastern Fennoscandia) |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| EXSB | Decision made by sipo to initiate substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20170718 Termination date: 20200326 |