CN106528498B  The point spread function extracting method of fixed star remote sensing image  Google Patents
The point spread function extracting method of fixed star remote sensing image Download PDFInfo
 Publication number
 CN106528498B CN106528498B CN201610957719.3A CN201610957719A CN106528498B CN 106528498 B CN106528498 B CN 106528498B CN 201610957719 A CN201610957719 A CN 201610957719A CN 106528498 B CN106528498 B CN 106528498B
 Authority
 CN
 China
 Prior art keywords
 fixed star
 point spread
 spread function
 column pixel
 pixel
 Prior art date
Links
 241000287181 Sturnus vulgaris Species 0.000 title claims abstract description 96
 238000004364 calculation methods Methods 0.000 claims abstract description 17
 239000000284 extracts Substances 0.000 claims abstract description 11
 230000003068 static Effects 0.000 claims abstract description 11
 239000011159 matrix materials Substances 0.000 claims description 15
 238000005070 sampling Methods 0.000 claims description 13
 230000017105 transposition Effects 0.000 claims description 3
 239000002956 ash Substances 0.000 claims 1
 241001269238 Data Species 0.000 abstract description 3
 239000000203 mixtures Substances 0.000 abstract description 2
 238000003384 imaging method Methods 0.000 description 7
 238000005516 engineering processes Methods 0.000 description 4
 238000000605 extraction Methods 0.000 description 3
 238000000034 methods Methods 0.000 description 3
 230000003287 optical Effects 0.000 description 2
 239000008264 clouds Substances 0.000 description 1
 238000010586 diagrams Methods 0.000 description 1
 238000009792 diffusion process Methods 0.000 description 1
 238000003012 network analysis Methods 0.000 description 1
Classifications

 G—PHYSICS
 G06—COMPUTING; CALCULATING; COUNTING
 G06F—ELECTRIC DIGITAL DATA PROCESSING
 G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
 G06F17/10—Complex mathematical operations
 G06F17/15—Correlation function computation including computation of convolution operations
Abstract
Description
Technical field
The present invention relates to Signals in Satellite Data Stream and technical field of image processing, and in particular, to a kind of fixed star remote sensing image Point spread function extracting method.
Background technique
With the raising of remote sensing satellite application technology level, the image quilt that highprecision fixed statellite is imaged over the ground is usually required that It is accurately positioned, and is registrated to the nominal grid of the earth.But satellite platform and remote sensing instrument are changed by space mechanics environment, thermal environment Etc. factors influence, satellite can have orbital drift and posture and be directed toward deviation, and instrument can generate geometry deformation, influence remote sensing images Positioning and registration accuracy.Satellite image is navigated and is registrated " technology can solve problems: satellite sweeps the two dimension of remote sensing instrument It retouches mechanism and carries out the compensation of realtime angular increment, Instrument observation point is guided in the motion track of earth's surface to preset path, thus real Existing imaging based navigation, makes remote sensing image registration to the nominal grid of the earth.The instrument geometry deformation of static remote sensing satellite inorbit period is not It can directly measure, sensitive imaging is carried out to fixed star using instrument, obtains the motion profile that fixed star streaks in detector array, By extract star motions rule and characteristic parameter, with fixed star inertial system physical location comparison can calculate remote sensing instrument Rail deflection.In order to improve deflection solving precision, need to obtain the accurate characteristic parameter of target fixed star, point spread function is Important parameter.
Mainly there are 3 kinds of methods to calculate the point spread function of imaging system at present.According to system design criteria and network analysis Model calculates point spread function；Highdefinition picture method of comparison needs a variety of imagers that resolution ratio is different；Using with feature The Image estimation point spread function of object.Method 1 requires the model of imaging system accurate, and otherwise the projection accuracy of point spread function is very It is difficult to guarantee；Method 2 needs additional imager to assist, it is clear that is not suitable for satellite image and navigates and be registrated fixed star point spread function Extraction；Method 3 is handled the image with characteristic body, is easily achieved in engineering, and common method is blade method.
Yang Lihong, Zhao Bianhong, Zhang Xingxiang, Ren Jianyue are " estimation of point spread function Gauss curve fitting restores with remote sensing images " Adjacent two pieces of ground by having knife edge in image is referred in (see " Chinese Optical ", 2012, the phase of volume 5 the 2nd) paper The point spread function of object estimation imaging system.For fixed statellite imaging system, directly fixed star is imaged, can not be formed with class Like the feature object image of knife edge, therefore this method is not suitable for satellite image yet and navigates and be registrated fixed star point spread function Extraction.
The method of the present invention Engineering Oriented is practical, is integrated by the area to fixed star image and pixel overlapping, according to perseverance The gray value of the suitable position and pixel of star image and pixel, the antipoint spread function for releasing fixed star.This method makes full use of More packet image datas when fixed star streaks, are fitted by multigroup data, computational accuracy greatly improved.Present invention can apply to The imaging positioning and registration arrangement development research and development and industry of the static remote sensing satellites such as China's second generation stationary weather satellite wind and cloud four Business application process.
Summary of the invention
For the defects in the prior art, the object of the present invention is to provide a kind of point spread functions of fixed star remote sensing image to mention Take method.
The point spread function extracting method of the fixed star remote sensing image provided according to the present invention, includes the following steps:
Step 1: requiring selection selection polynomial with the point spread function of fitting of a polynomial fixed star, and according to fitting precision Order；
Step 2: according to the point spread function of fixed star, and the relative position of fixed star image and column pixel being combined to derive picture of falling out The calculation formula of first total gray value；
Step 3: the continuous exposure of the upper a certain column pixel of static remote sensing satellite detectors face battle array is passed in and out using fixed star image Light data extracts the total gray value of column pixel with the variation relation of distance L, and wherein L indicates fixed star image center and column pixel center Distance；
Step 4: the column pixel that the total gray value calculation formula of column pixel and step 3 derived in conjunction with step 2 extract is total Gray value goes out the undetermined coefficient in point spread function using least square method the Fitting Calculation with the variation relation of distance L.
Preferably, the step 1 includes:
The point spread function of fixed star is fitted with multinomial, when the order of polynomial fitting is higher, the mistake of fitting Difference is smaller, while the undetermined coefficient introduced can also increase；
Specifically, the point spread function of selection fourth order polynomial fitting fixed star, calculation formula are as follows:
F (r)=ar^{4}+b·r^{3}+c·r^{2}+d·r+e (1)
In formula: f (r) indicates fitting function, a, b, c, d, the undetermined coefficient of e representative polynomial, r expression to fixed star dot center Distance.
Preferably, the step 2 includes:
Step 2.1: establishing the coordinate system between fixed star image and column pixel, be to sit with the center of fixed star image specifically Origin is marked, the direction of motion of fixed star image is xaxis, and the vertical line of the direction of motion is yaxis, then the center point coordinate of fixed star image is (0,0), if column pixel dimension size is D, fixed star image center is L at a distance from column pixel center, then the two of column pixel side It is written as x=Ld/2 and x=L+d/2；
Step 2.2: four pixels up and down centered on fixed star image being integrated, then the integral boundary of yaxis is written as x The calculation formula of=LD/2 and y=2D, the gray value of the column pixel are as follows:
In formula: h indicates that the gray value of respective column pixel, function f () are the point spread function that fourth order polynomial indicates, becomes Measuring x indicates that fixed star image East and West direction coordinate, variable y indicate fixed star image northsouth coordinate.
Preferably, the step 3 includes: upper a certain using fixed star image disengaging static remote sensing satellite detectors face battle array The continuous exposure data of column pixel extract the total gray value of column pixel with the variation relation of distance L, and it is public that abbreviation obtains following calculating Formula:
H=f_{1}(L)·a+f_{2}(L)·b+f_{3}(L)·c+f_{4}(L)·d+f_{5}(L)·e (3)
The formula (3) gives the correlation of the total gray value and point spread function, L of column pixel, in formula: f_{1}(L) table Show the influence that fourth order point spread function integrates pixel, f_{2}(L) influence that third rank point spread function integrates pixel is indicated, f_{3}(L) influence that secondorder point spread function integrates pixel, f are indicated_{4}(L) indicate that the first rank point spread function integrates pixel Influence, f_{5}(L) influence for indicating that zeroth order point spread function integrates pixel is respectively indicated.
Preferably, the step 4 includes:
The continuous exposure data of the upper a certain column pixel of static remote sensing satellite detectors face battle array are passed in and out using fixed star image, The total gray value of column pixel is extracted with the variation relation of distance L, is denoted as h=F (L)；
In conjunction with formula (3), release
F (L)=f_{1}(L)·a+f_{2}(L)·b+f_{3}(L)·c+f_{4}(L)·d+f_{5}(L)·e (4)
By L=L_{1},L_{2}...L_{n}( 3D=L≤3D) brings above formula into, then
It enables
Formula (5) is abbreviated as Ay=B, is fitted, is obtained by least square method:
Y=(A^{T}A)^{1}A^{T}B (9)
In formula: F (L) indicates variation relation of the total gray value of pixel with distance L, L_{1}It indicates in sampling fixed star image for the first time The heart is at a distance from column pixel center, L_{n}Nth sampling fixed star image center is indicated at a distance from column pixel center, subscript n takes Value is sampling number, and A indicates that gray count matrix, y indicate that point spread function coefficient matrix, B indicate gray scale sampling matrix, A^{T}Table Show the transposition of gray count matrix, (A^{T}A)^{1}Indicate the operation of gray count inverse of a matrix；
After acquiring polynomial undetermined coefficient, function f (r) is brought into, then obtain point diffusion of the fixed star image in detector Function.
Compared with prior art, the present invention have it is following the utility model has the advantages that
1, the method in the present invention is integrated by the area to fixed star image and pixel overlapping, according to fixed star image and The suitable position of pixel and the gray value of pixel, the antipoint spread function for releasing fixed star；This method takes full advantage of fixed star and draws Outofdate more packet image datas, are fitted by multigroup data, computational accuracy greatly improved.
2, the present invention avoids the integral operation to Bessel function complexity compared with existing technology and means, reduces Calculation amount, is more suitable for engineer application.
3, the method in the present invention is eliminated by the fitting of multiframe data by detector optical imagery, circuit noise, satellite Asterism position high frequency error caused by the factors such as high dither improves satellite image and navigates and be registrated fixed star point spread function Extraction accuracy.
Detailed description of the invention
Upon reading the detailed description of nonlimiting embodiments with reference to the following drawings, other feature of the invention, Objects and advantages will become more apparent upon:
Fig. 1 is the flow diagram of the point spread function extracting method of fixed star remote sensing image provided by the invention；
Fig. 2 is the matched curve of second order polynomial fit Bessel function；
Fig. 3 is the matched curve that fourth order polynomial is fitted Bessel function；
Fig. 4 is the matched curve that six rank multinomials are fitted Bessel function；
Fig. 5 is the process schematic that fixed star image streaks a certain column pixel.
Specific embodiment
The present invention is described in detail combined with specific embodiments below.Following embodiment will be helpful to the technology of this field Personnel further understand the present invention, but the invention is not limited in any way.It should be pointed out that the ordinary skill of this field For personnel, without departing from the inventive concept of the premise, several changes and improvements can also be made.These belong to the present invention Protection scope.
The point spread function extracting method of the fixed star remote sensing image provided according to the present invention includes the following steps:
Step S1: replacing the point spread function of Bessel function fitting fixed star with multinomial, and consider the precision of fitting Calculation amount selects polynomial order；
Step S2: column are derived in conjunction with the relative position of fixed star image and column pixel according to the point spread function of undetermined coefficient The total gray scale value expression of pixel；
Step S3: the continuous exposure of the upper a certain column pixel of static remote sensing satellite detectors face battle array is passed in and out using fixed star image Light data extracts the total gray value of column pixel with the variation relation of distance L；
Step S4: the column pixel extracted in conjunction with the step S2 column pixel gray scale value expression derived and step 3 is always grey Angle value goes out the undetermined coefficient of point spread function using least square method the Fitting Calculation.
With fitting of a polynomial point spread function
Point spread function generally indicates that the Bessel function curve in attached drawing 2 describes fixed star and visiting with Bessel function The Luminance Distribution characteristic in device is surveyed, that is, the distance for arriving asterism center is closer, and brightness is higher；Conversely, the distance to asterism center is got over Far, brightness is lower.
In fact, the Bessel function curve in attached drawing 2 can use fitting of a polynomial, therefore can be by this section of Bessel function Write as the polynomial form of undetermined parameter.From fitting of a polynomial result it is found that polynomial order is higher, the error of fitting is got over It is small, but polynomial undetermined coefficient is also increase accordingly, and increases operand.From following table as can be seen that being fitted with fourth order polynomial When Bessel function, fitting average deviation is 0.0001985, solves the error that pixel gray value introduces with fourth order polynomial and is about 0.01985%, much smaller than the sampling error of pixel gray value.Therefore the point spread function of fourth order polynomial fitting fixed star can be used Number, expression formula are
F (r)=ar^{4}+b·r^{3}+c·r^{2}+d·r+e (1)
A in above formula, b, c, d, e are polynomial undetermined coefficient, and r is the distance to asterism center.
The different rank multinomials of table 1 are fitted mean square deviation
Total gray scale value expression of column pixel is derived using point spread function
When fixed star image streaks column pixel, the relationship signal of fixed star image and column pixel is as shown in Fig. 3.It establishes following Coordinate system: using the center of fixed star image as coordinate origin, the direction of motion of fixed star image is xaxis, and the vertical line of the direction of motion is y Axis, then the center point coordinate of fixed star image is (0,0).If pixel dimension size is D, fixed star image center and column pixel center Distance is L, then the two of column pixel side can be written as x=LD/2 and x=L+D/2.In view of the size of fixed star image and column pixel It is similar, in order to fully consider influence of the fixed star image to column pixel, four pixels up and down centered on fixed star image are carried out Integral, then the integral boundary of yaxis can be written as y=2D and y=2D, and the gray scale value expression of the column pixel is writeable to be
Function f () in above formula is the point spread function that fourth order polynomial indicates, variable x indicates fixed star image East and West direction Coordinate, variable y indicate fixed star image northsouth coordinate, can obtain after abbreviation
H=f_{1}(L)·a+f_{2}(L)·b+f_{3}(L)·c+f_{4}(L)·d+f_{5}(L)·e (3)
The formula (3) gives the correlation of the total gray value and point spread function, L of column pixel.In above formula, f_{1}(L) Indicate the influence that fourth order point spread function integrates pixel, f_{2}(L) shadow that third rank point spread function integrates pixel is indicated It rings, f_{3}(L) influence that secondorder point spread function integrates pixel, f are indicated_{4}(L) indicate the first rank point spread function to pixel product The influence divided, f_{5}(L) influence for indicating that zeroth order point spread function integrates pixel is respectively indicated.
Least square method solves point spread function undetermined coefficient
The continuous exposure of the upper a certain column pixel of static remote sensing satellite detectors face battle array is passed in and out first with fixed star image Data extract the total gray value of column pixel with the variation relation of distance L, are denoted as h=F (L).
Convolution (3), can obtain
F (L)=f_{1}(L)·a+f_{2}(L)·b+f_{3}(L)·c+f_{4}(L)·d+f_{5}(L)·e (4)
By L=L_{1},L_{2}...L_{n}( 3D≤L≤3D) brings above formula into, can obtain
It enables
Formula (5) is abbreviated as Ay=B, is fitted, is obtained by least square method:
Y=(A^{T}A)^{1}A^{T}B (9)
In formula: F (L) indicates variation relation of the total gray value of pixel with distance L, L_{1}It indicates in sampling fixed star image for the first time The heart is at a distance from column pixel center, L_{n}Nth sampling fixed star image center is indicated at a distance from column pixel center, subscript n takes Value is sampling number, and A indicates that gray count matrix, y indicate that point spread function coefficient matrix, B indicate gray scale sampling matrix, A^{T}Table Show the transposition of gray count matrix, (A^{T}A)^{1}Indicate the operation of gray count inverse of a matrix.
After acquiring undetermined coefficient, function f (r) is brought into, be point spread function of the fixed star image in detector.
Specific embodiments of the present invention are described above.It is to be appreciated that the invention is not limited to abovementioned Particular implementation, those skilled in the art can make a variety of changes or modify within the scope of the claims, this not shadow Ring substantive content of the invention.In the absence of conflict, the feature in embodiments herein and embodiment can any phase Mutually combination.
Claims (2)
Priority Applications (1)
Application Number  Priority Date  Filing Date  Title 

CN201610957719.3A CN106528498B (en)  20161103  20161103  The point spread function extracting method of fixed star remote sensing image 
Applications Claiming Priority (1)
Application Number  Priority Date  Filing Date  Title 

CN201610957719.3A CN106528498B (en)  20161103  20161103  The point spread function extracting method of fixed star remote sensing image 
Publications (2)
Publication Number  Publication Date 

CN106528498A CN106528498A (en)  20170322 
CN106528498B true CN106528498B (en)  20190118 
Family
ID=58326946
Family Applications (1)
Application Number  Title  Priority Date  Filing Date 

CN201610957719.3A CN106528498B (en)  20161103  20161103  The point spread function extracting method of fixed star remote sensing image 
Country Status (1)
Country  Link 

CN (1)  CN106528498B (en) 
Families Citing this family (1)
Publication number  Priority date  Publication date  Assignee  Title 

CN108414462B (en) *  20180210  20201009  中国科学院国家天文台  Lowresolution fixed star continuous spectrum automatic fitting method based on template matching 
Citations (7)
Publication number  Priority date  Publication date  Assignee  Title 

CN101441764A (en) *  20081231  20090527  中国资源卫星应用中心  MTFC remote sensing image restoration method 
CN102116626A (en) *  20091231  20110706  北京控制工程研究所  Prediction and correction method of node of star point track image 
CN102779333A (en) *  20120710  20121114  武汉大学  Optical image restoration method based on Kalman filter 
WO2013019180A1 (en) *  20110729  20130207  Raytheon Company  Method and system for vicarious spatial characterization of a remote image sensor 
CN105092156A (en) *  20150921  20151125  上海卫星工程研究所  Triaxial air bearing table highprecision mass center balance adjusting device and method 
CN105572692A (en) *  20151216  20160511  上海卫星工程研究所  Satellite image navigation and registering fullphysical testing device and testing method 
CN105758400A (en) *  20160215  20160713  上海卫星工程研究所  Extracting method for fixed star sensitive east and west parameters of stationary satellite imaging navigation and registration 
Family Cites Families (2)
Publication number  Priority date  Publication date  Assignee  Title 

WO2006055471A2 (en) *  20041115  20060526  Imaging Diagnostic Systems, Inc.  Apparatus and method for acquiring timeresolved measurements 
US7636155B2 (en) *  20070118  20091222  Dcg Systems, Inc.  System and method for resolving photoemission from semiconductor devices 

2016
 20161103 CN CN201610957719.3A patent/CN106528498B/en active IP Right Grant
Patent Citations (7)
Publication number  Priority date  Publication date  Assignee  Title 

CN101441764A (en) *  20081231  20090527  中国资源卫星应用中心  MTFC remote sensing image restoration method 
CN102116626A (en) *  20091231  20110706  北京控制工程研究所  Prediction and correction method of node of star point track image 
WO2013019180A1 (en) *  20110729  20130207  Raytheon Company  Method and system for vicarious spatial characterization of a remote image sensor 
CN102779333A (en) *  20120710  20121114  武汉大学  Optical image restoration method based on Kalman filter 
CN105092156A (en) *  20150921  20151125  上海卫星工程研究所  Triaxial air bearing table highprecision mass center balance adjusting device and method 
CN105572692A (en) *  20151216  20160511  上海卫星工程研究所  Satellite image navigation and registering fullphysical testing device and testing method 
CN105758400A (en) *  20160215  20160713  上海卫星工程研究所  Extracting method for fixed star sensitive east and west parameters of stationary satellite imaging navigation and registration 
NonPatent Citations (3)
Title 

点扩散函数的改进倾斜刃边重建的高精度估计;范冲等;《测绘学报》;20151130;第44卷(第11期);第12191226页 
点扩散函数高斯拟合估计与遥感图像恢复;杨利红等;《中国光学》;20120430;第5卷(第2期);第181188页 
遥感相机在轨点扩散函数高精度测量方法;郭玲玲等;《测绘学报》;20140331;第43卷(第3期);第284289页 
Also Published As
Publication number  Publication date 

CN106528498A (en)  20170322 
Similar Documents
Publication  Publication Date  Title 

James et al.  3‐D uncertainty‐based topographic change detection with structure‐from‐motion photogrammetry: precision maps for ground control and directly georeferenced surveys  
Eltner et al.  Analysis of different methods for 3D reconstruction of natural surfaces from parallel‐axes UAV images  
Mohammed et al.  Positional accuracy testing of Google Earth  
CA2592293C (en)  Method for processing images using automatic georeferencing of images derived from a pair of images captured in the same focal plane  
Tang et al.  Triple lineararray image geometry model of ZiYuan3 surveying satellite and its validation  
EP2111530B1 (en)  Automatic stereo measurement of a point of interest in a scene  
Grodecki  IKONOS stereo feature extraction–RPC approach  
Hu et al.  Understanding the rational function model: methods and applications  
Leprince et al.  Inflight CCD distortion calibration for pushbroom satellites based on subpixel correlation  
US8452123B2 (en)  Distortion calibration for optical sensors  
KR101295747B1 (en)  System of automatic geometric correction using rational polynomial cofficient and method thereof  
Tong et al.  Framework of jitter detection and compensation for high resolution satellites  
Fraser et al.  Three‐dimensional geopositioning accuracy of IKONOS imagery  
Ayoub et al.  Coregistration and correlation of aerial photographs for ground deformation measurements  
US8098958B2 (en)  Processing architecture for automatic image registration  
EP3060880A2 (en)  Sky polarization and sun sensor system and method  
CN105823416B (en)  The method and apparatus of polyphaser measurement object  
Li et al.  Rigorous photogrammetric processing of HiRISE stereo imagery for Mars topographic mapping  
Kääb et al.  Coseismic displacements of the 14 November 2016 Mw 7.8 Kaikoura, New Zealand, earthquake using the Planet optical cubesat constellation  
US20120257792A1 (en)  Method for GeoReferencing An Imaged Area  
CN103822615B (en)  A kind of multicontrol point extracts and the unmanned aerial vehicle target realtime location method be polymerized automatically  
JP2008506167A (en)  Method and apparatus for determining a location associated with an image  
CN102968631B (en)  The automatic geometric of mountain area multispectral remote sensing satellite image is corrected and orthorectification method  
Jiang et al.  Geometric calibration and accuracy assessment of ZiYuan3 multispectral images  
KR20110120317A (en)  Registration of 3d point cloud data to 2d electrooptical image data 
Legal Events
Date  Code  Title  Description 

C06  Publication  
PB01  Publication  
SE01  Entry into force of request for substantive examination  
SE01  Entry into force of request for substantive examination  
GR01  Patent grant  
GR01  Patent grant 