CN108305237A - Consider more stereopsis fusion drafting method of different illumination image-forming conditions - Google Patents
Consider more stereopsis fusion drafting method of different illumination image-forming conditions Download PDFInfo
- Publication number
- CN108305237A CN108305237A CN201810062375.9A CN201810062375A CN108305237A CN 108305237 A CN108305237 A CN 108305237A CN 201810062375 A CN201810062375 A CN 201810062375A CN 108305237 A CN108305237 A CN 108305237A
- Authority
- CN
- China
- Prior art keywords
- stereopsis
- image
- point
- dem
- adjustment
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration by the use of more than one image, e.g. averaging, subtraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T15/00—3D [Three Dimensional] image rendering
- G06T15/10—Geometric effects
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T15/00—3D [Three Dimensional] image rendering
- G06T15/50—Lighting effects
- G06T15/506—Illumination models
-
- 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
- G06T2207/10028—Range image; Depth image; 3D point clouds
-
- 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
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
Abstract
The present invention relates to a kind of more stereopsis considering different illumination image-forming conditions to merge drafting method, includes the following steps：1) it obtains survey region and has image, and select Multi folds coverage stereopsis pair；2) the imaging geometry model of remote sensing image is built；3) same place of remote sensing image is obtained to carrying out initial matching to stereopsis；Step 4) obtains imaging geometry model to step 2) and carries out stereopsis regional network bundle adjustment, the imaging geometry model after being refined；5) the intensive same place set of image space of each group stereopsis pair is respectively obtained to carrying out dense Stereo Matching to each group of stereopsis；6) according to the imaging geometry model after 4) refining, by it is rapid 5) in the intensive same place of obtained image space carry out forward intersection, the dense three-dimensional point cloud of object space is obtained, and then DEM is obtained by grid partition and interpolation, and DEM is merged to obtain without empty DEM；7) according to the imaging geometry model without empty DEM and after refining, generating DOM is：The image of geography information is carried after topographical correction.
Description
Technical field
The present invention relates to a kind of more stereopsis considering different illumination image-forming conditions to merge drafting method, is related to photographing
Measure neutral body photogrammetric technology field.
Background technology
The remote sensing mapping on planet (moon, Mars) surface is a basic work in planetary exploration mission, is planet
Scientific exploration mission planning and scientific research provide Geographic Reference basis, are the basic hands for obtaining planet pattern and tectonic information
Section.The current planetary orbit device remote sensing image the same area that charts is to make the region based on a pair of of stereopsis pair mostly
High-precision topographic map product.However single stereopsis, to being influenced by observation angle and illumination condition etc., the information provided is past
Toward limited, some special region surface hypsographies are larger, generate shade or the supersaturated area of imaging it is more in the case of, often
The missing of information when causing to chart.
Traditional charts to orbiter, orbital vehicle image based on single stereopsis, generally comprises following steps：First according to track
Posture and elements of interior orientation etc. establish the imaging model of stereopsis, then carry out relative orientation to two images or flux of light method is flat
Difference establishes accurate view stereoscopic to reduce the inconsistency brought due to track profile, location parameter equal error as far as possible
Imaging relations；Dense Stereo Matching finally is carried out to two images, it is preceding by accurate imaging model and the same place of dense Stereo Matching
Side's intersection obtains the three-dimensional coordinate of lunar surface point off density；Digital elevation mould is obtained finally by three-dimensional point cloud progress interpolation and editor
Type (Digital Elevation Model, DEM) is corrected to obtain number and just penetrated using imaging model and the DEM to image
Image (Digital Orthophoto Map, DOM).In above-mentioned steps, the effect of dense Stereo Matching will directly determine landform three
Tie up the effect rebuild, namely the quality of drawing product.However during dense Stereo Matching, due to lighting angle and landform etc. because
The influence of element, leads to the region of elevation big rise and fall on image, was susceptible to bright gray scale saturation region or excessively dark serious the moon
Shadow zone.These regions can not match to obtain same place since texture lacks, final so that there is sky on the landform product generated
Hole.If simply by the information of surrounding into row interpolation, though can solve the problems, such as empty, topographic details information is often resulted in
It loses so that landform Product Precision is difficult to ensure.Especially in hypsography large area, by way of repairing cavity interpolation
Large error can be brought, thus is difficult to meet the requirement of fine landform production.
Currently, with the extension of planetary orbit duty cycle, simultaneously because the acquisition of multitask orbiter, orbital vehicle data, in detection
It tends to have accumulated a large amount of repeated measures image on planet (such as moon, Mars) surface.Since planetary orbit is different from over the ground
The sun-synchronous orbit of observation satellite generally use, illumination consistency when can not often ensure to obtain Multi folds coverage image, because
Illumination condition is often different when this Multi folds coverage.These images are made full use of, its potential value is made greatly to be excavated,
Loss of learning when supplement list three-dimensional imaging.
Invention content
In view of the above-mentioned problems, considering that more stereopsis of different illumination image-forming conditions melt the object of the present invention is to provide a kind of
Drafting method is closed, the multiple stereopsis pair obtained under different illumination conditions can be utilized, carries out high-precision three-dimensional fusion drawing
Generate optimal landform product.
To achieve the above object, the present invention takes following technical scheme：It is a kind of to consider the mostly vertical of different illumination image-forming conditions
Body visual fusion drafting method, it is characterised in that include the following steps：
Step 1)：It obtains survey region and has image, and select Multi folds coverage stereopsis pair；
Step 2)：Build the imaging geometry model of remote sensing image；
Step 3)：To stereopsis to carrying out initial matching, the same place of remote sensing image is obtained；
Step 4)：Imaging geometry model is obtained to step 2) and carries out stereopsis regional network bundle adjustment, is refined
Imaging geometry model afterwards；
Step 5)：To each group of stereopsis to carrying out dense Stereo Matching, the image space for respectively obtaining each group stereopsis pair is close
Collect same place set；
Step 6)：Imaging geometry model after being refined according to step 4), by the intensive same place of the image space obtained in step 5)
Forward intersection is carried out, obtains the dense three-dimensional point cloud of object space, and then DEM is obtained by grid partition and interpolation, and carry out to DEM
Fusion is obtained without empty DEM；
Step 7)：According to the imaging geometry model without empty DEM and after refining, generating DOM is：The band after topographical correction
There is the image of geography information.
Further, the specific choice process of the step 1) is：According between image intersection angle, video imaging when the sun
Azimuth and elevation angle, the aspect ratio of image, image overlap range select multigroup stereopsis pair, form each group of stereopsis
To two images should meet three-dimensional intersection angle more than 10 °, imaging when solar azimuth and elevation angle it is essentially identical and
Image aspect ratio should have different sun altitudes or azimuth and mutually it close to 1 between multigroup stereopsis pair
Between there are overlapping ranges as big as possible.
Further, the step 2) includes the structure of stringent imaging geometry model and general geometrical model, wherein general
Geometrical model is multinomial model, direct linear transformation's model or rational function model.
Further, the detailed process of the step 3) is：Step 3.1)：When direction of illumination is identical, in stereopsis pair
Overlapping region in, stereopsis pair identical for illumination condition, the same place of image is matched using automatic image in solid
Image obtains equally distributed same place to upper；Step 3.2)：When direction of illumination difference, in image by way of artificial interpretation
Between choose same place, then Mismatching point is rejected, retains the correct match point of each stereopsis pair.
Further, the detailed process of the step 4) is：
Step 4.1)：Build adjustment error equation：
Fx=samplei+a0+a1·samplei+a2·linei-xi=0
Fy=linei+b0+b1·samplei+b2·linei-yi=0
In formula, FxAnd FyIndicate x, y-coordinate error equation function, a0,a1,a2And b0,b1,b2To correct parameter, (sample,
Line coordinate) is measured for the image space after normalization, (x, y) is image space coordinate, and subscript i is a labelled notation；
Step 4.2)：Carry out error equation solution：First order Taylor series expansion is carried out to error equation, is linearized
Error equation：
In formula, lat is latitude, and lon is longitude, and h is elevation, Fx0, Fy0Respectively value of the error function at x0, y0,Respectively phase
Answer the first-order partial derivative of parameter, Δ a0、Δa1、Δa2、Δb0、Δb1、Δb2, Δ lat, Δ lon and Δ h be respectively relevant parameter
Corrected value；
Obtain indirect adjustment model：
V=AX+BY-L, P
In formula, X is adjustment parameter：
X=[Δ a0 Δa1 Δa2 Δb0 Δb1 Δb2]
A is the coefficient matrix of unknown number X：
Y is the ground coordinate of tie point：
Y=[Δ lat Δ lon Δs h]
B is the coefficient matrix of unknown number Y：
In formula, LINE_SCALE, SAMP_SCALE be image space coordinate normalized parameter, LAT_SCALE, LON_SCALE,
HEIGHT_SCALE is object coordinates normalized parameter；
For control point, B 0, L are constant term, can be calculated by initial value and error equation：
In formula, P is weight matrix, and for control point, unknown number is adjustment parameter (a0,a1,a2,b0,b1,b2)；And it is right
For tie point, unknown number includes adjustment parameter and corresponding object coordinates (lat, lon, h)；
When existing simultaneously control point and tie point, you can establish following error equation：
For the error equation of two class unknown numbers, the second class unknown number Y changed normal equations accordingly are eliminated：
The solution of unknown number X is completed by above formula；
Step 4.3)：Stereopsis obtains virtual controlling point to adjustment respectively
According to step 4.1) and 4.2), the adjustment at no control point is carried out respectively to every group to multigroup stereopsis, selection is flat
One group with high accuracy of difference, i.e., image space back projection error is minimum in adjustment one group utilize the rational function mould of this group of stereopsis
Type and adjustment corrected value and this on image more degree overlapping tie point this to the picpointed coordinate on image, forward intersection goes out more
The lunar surface three-dimensional coordinate of degree overlapping tie point, using these lunar surface three-dimensional coordinates as virtual controlling point coordinates；
Step 4.4)：Overall adjustment obtains all image corrections
By all stereopsis to carrying out adjustment with the method 4.2) provided using step 4.1), and using in step 4.3)
Virtual controlling point as control.
Further, the detailed process of the step 6)：
Step 6.1)：According to stereopsis to the object space range of overlay region, to every group of stereopsis to establishing same unit
The regular grid at interval；
Step 6.2)：Centered on each grid points, search radius is the point off density within the scope of R；
Step 6.3)：When search result is that points that are empty or searching are unsatisfactory for requiring, it is believed that the grid points are cavity
Point records its number and plane coordinates information (ID, X, Y)；Otherwise lattice are obtained according to the elevation value interpolation of the point off density searched
The height value of site；
Step 6.4)：It repeats the above steps, is corresponded to by each stereopsis to obtaining cavity position on a DEM and DEM
Number and coordinate information；
Step 6.5)：According to each stereopsis in step 4.2) adjustment result and each DEM on empty grid
The quantity of point, select error in adjustment checkpoint small and an empty DEM at least is as base map, using other DEM come to it
Carry out empty repairing, wherein the detailed process of cavity repairing is：
Step 6.5.1)：The plane coordinates for taking on DEM base maps each cavity point successively, on other DEM with n (n=1~
2) a mesh spacing is radius, searches the grid height value within the scope of this；
Step 6.5.2)：If on several DEM simultaneously search met the requirements as a result, if by stereopsis to putting down
Poor precision and with cavity put between at a distance from be weighted, be finally averaged to obtain the height value of the point；What specific weighting solved
Process is as follows：It is arbitrary to take in search result error RMS in the adjustment of i-th group of stereopsis pairiAnd j-th search result point with it is empty
The distance D of hole pointjOn the basis of, then the adjustment precision RMS of arbitrary kth group stereopsis pairkWith the ratio R MS of the benchmarki/RMSk
For the adjustment precision weights of kth group stereopsis pair, any one searches point h and cavity point distance DhWith the ratio of reference distance
Value Dj/DhIndicate as follows for the height value H apart from weights, final unknown point：
In formula, i, k=1,2 ..., m, m indicate the quantity of stereopsis pair；H, j=1,2 ..., Pn,PnIndicate search result
The quantity of point, HhIndicate h-th of height value for searching point position.
Step 6.5.3)：If only searched on a width DEM it is meeting the requirements as a result, if by with cavity put between at a distance from
It is weighted the height value for averagely acquiring cavity point as the following formula；
Step 6.5.4)：If the point is empty point on other all DEM corresponding positions, it is numbered simultaneously again
Store its plane coordinates；
Step 6.6)：The repairing of all cavity points is completed according to step 6.5).For being finally still recorded as cavity point
Position, the then method for using traditional interpolation, the height value of the position is obtained by grid points interpolation around the DEM；
Step 6.7)：Using abnormality value removing algorithm, mutated site is detected, and is obtained according to the pixel interpolating of surrounding
To the reasonable height value of mutated site, more stereopsis DEM fusion process are so far completed, obtain without cavity and retain topographic details
DEM.
The invention adopts the above technical scheme, which has the following advantages：1, it is multiple to propose planetary surface by the present invention
Overlay area, using multiple stereopsis to merging the method charted to carry out high-precision, this method is by using the different sun
The complementary information that stereopsis pair is obtained under elevation angle and azimuth, on individual drawing product due to shade or it is highlighted caused by
Hole area is repaired, the high-precision landform product merged through the invention, not only ensure that the integrality of Global Information but also
The detailed information of landform is remained as much as possible.2, the present invention is not necessarily under the premise of ensureing cartographic accuracy and details integrality
Image is reacquired, is to be made full use of to existing image, compared to traditional slur as stereomapping, the present invention can be more
Fully excavate the potential value of Multi folds coverage area redundancy image.
Description of the drawings
Fig. 1 is the stereopsis fusion drafting method flow diagram of the different illumination image-forming conditions of the present invention；
Fig. 2 is different illumination conditions LRO NAC image contrasts of the present invention, and figure (a) is that left and right mirror images are (left：
M129133239RE, it is right：M162154734RE), figure (b) is that normal orientation image is (left：M150361817RE, it is right：
M150368601RE), direction of illumination is opposite with left and right mirror images.
Specific implementation mode
Come to carry out detailed description to the present invention below in conjunction with attached drawing.It should be appreciated, however, that attached drawing has been provided only more
Understand the present invention well, they should not be interpreted as limitation of the present invention.
As shown in Figure 1, more stereopsis high-precision fusion drafting method of difference illumination image-forming condition provided by the invention,
Include the following steps：
1, obtain survey region have image, select Multi folds coverage stereopsis to S1, S2 ... Sn-1, Sn
Stereopsis is to indicating from the image with stereopsis condition absorbed to areal.Obtain survey region
Existing image, according between image intersection angle, video imaging when solar azimuth and elevation angle, the aspect ratio of image, image
The information such as overlapping range select multigroup stereopsis pair, and three-dimensional friendship should be met by forming two images of each group of stereopsis pair
Can angle be more than 10 °, imaging when solar azimuth and elevation angle it is essentially identical, image aspect ratio is and multigroup vertical close to the conditions such as 1
Should have different sun altitudes or azimuth between body image pair and there is overlapping model as big as possible between each other
It encloses.
As shown in Fig. 2, with lunar reconnaissance orbiter, orbital vehicle (lunar reconnaissance orbiter, LRO) narrow angle camera
For (narrow angle camera, NAC), the Track desigh of LRO allows obtains normal orientation and a left side in the same area
The NAC images of right mirror image, there is the opposite features of direction of illumination for the two.Select multigroup solid with similar illumination feature
Image pair merges the landform product of generation the stereopsis with different illumination conditions, to eliminate due to texture
Landform product cavity problem caused by missing can not match.
2, remote sensing image imaging geometry model is built
Currently used planetary remote sensing video imaging geometrical model is broadly divided into stringent imaging geometry model and universal imaging
Geometrical model.Stringent imaging geometry model is the mathematical model for having tight theoretical foundation, it mainly based on collinearity equation,
The stringent geometrical relationship between image coordinate and ground point space coordinate can accurately be expressed.Universal imaging geometrical model then returns
The complex relationship for having kept away imaging process, the mutual pass being fitted between picpointed coordinate and object space point three-dimensional coordinate using mathematical model
System.Common model of fit has average polynomial, direct linear transformation's model and rational function model etc., wherein reasonable
Function model becomes due to the advantages that its fitting precision is high, versatility is good, using facilitating in the general geometrical model of remote sensing image
A kind of most widely used mathematical model.Any stringent imaging geometry model and universal imaging geometry mould may be used in the present invention
Type.
By taking the imaging geometry model construction of LRO NAC images as an example, the stringent imaging geometry model that the present invention will be described in detail
With the building process of general geometrical model.The structure of stringent imaging geometry model generally comprises two mistakes of interior orientation and outer orientation
Journey, and the structure of general geometrical model then needs based on the stringent imaging geometry model of structure, detailed process is：
1) structure of the stringent imaging geometry models of LRO NAC
1.1) LRO NAC interior orientations
The interior orientation parameter of NAC cameras is obtained from the IK secondary files of LRO for example：Focal length, ranks direction centre coordinate,
Pixel dimension and distortion parameter etc. carry out NAC cameras default then according to the distortion model (being shown below) of LRO NAC
To.
In formula, sample is row coordinate of the picture point in NAC EDR initial data, and BORESIGHT_SAMPLE is column direction
Centre coordinate, PIXEL_PITCH is the pixel dimension of column direction, and xd is the coordinate (measure coordinate) comprising photogrammetric distortion, and k1 is
Radial distortion parameter, r are distance of the picture point to principal point, and xc is to correct rear image point in the coordinate of focal plane, unit mm.Due to
NAC is CCD linear array scanning cameras, therefore similar the parameter yd=0, yc=0 of line direction.
1.2) it is oriented outside LRO NAC
1.2.1) establish collinearity equation
After the completion of interior orientation, coordinate of each pixel on focal plane after distortion correction can be obtained, outer orientation is
It establishes focal plane coordinate system and consolidates the relationship of coordinate system with star, stringent imaging geometry model can be expressed with collinearity equation：
In formula, (xc, yc) is the focal plane coordinate of picture point, and f is focal length, and (X, Y, Z) is that corresponding object space point is sat admittedly in star
The coordinate of system is marked, (Xs, Ys, Zs) is the coordinate that photo centre consolidates coordinate system in star, and the referred to as line element of elements of exterior orientation, λ is
One scale factor, R is the spin matrix that image space coordinate system consolidates coordinate system to star, by three exterior orientation angle elementsComposition.
1.2.2) the reading of initial elements of exterior orientation
Outer orientation is carried out to image, it is necessary first to obtain the elements of exterior orientation of imaging moment.Elements of exterior orientation is from track
It measures and is obtained in obtained position of aircraft and attitude data, these data measured are stored in LRO NAC as auxiliary data
In the SPICE kernel files of image, read so the elements of exterior orientation of every image can be corresponded to from it in SPICE kernel
It takes.
1.2.3) the elements of exterior orientation of every scan line of interpolation
For the orbiter, orbital vehicle image of push-broom type imaging, each scan line has corresponding elements of exterior orientation.But satellite
Orbit measurement time interval is more than each row image scan imaging time interval, to obtain the elements of exterior orientation needs of every scan line
By the way of interpolation.The general elements of exterior orientation function established using three rank multinomials relative to imaging time t, according to record
Often row CCD imaging times, the elements of exterior orientation of every scan line can be obtained with interpolation.
In formula, Xs(t),Ys(t),Zs(t) indicate that t moment photo centre consolidates the coordinate in coordinate system, i.e. foreign side's bit line in star
Element；ω (t), κ (t) indicate that the attitude angle in coordinate system, i.e. exterior orientation angle element are consolidated in t moment focal plane in star；
a0...f3Indicate that least square method can be used to be solved according to orbital measurement data for the multinomial coefficient of corresponding parameter, these coefficients.
1.2.4) by collinearity equation and the elements of exterior orientation acquired, can will turn by the focal plane coordinate of distortion correction
It changes object coordinates into, completes the foundation of the stringent imaging geometry model of sensor.
2) foundation of LRO NAC rational function models
The foundation of LRO NAC rational function models needs to initially set up virtual controlling grid, further according to the virtual control of generation
System point solves rational function model parameter.
2.1) foundation of virtual controlling grid
When establishing virtual controlling grid, need the elevation of imagery zone to be divided into several elevation faces, in image space with certain
Spacing generate the grid point coordinates of image as image space virtual controlling point, then according to rigorous geometric model by grid points image
Object space virtual controlling point is obtained on coordinate projection to each elevation face.
2.2) solution of rational function model parameter
Rational polynominal model sets up arbitrary topocentric coordinates (lat, lon, h) and corresponding image by ratio multinomial
One-to-one relationship between coordinate (sample, line), expression-form are as follows：
In formula,
NumL(P, L, H)=a1+a2L+a3P+a4H+a5LP+a6LH+a7PH+a8L2+a9P2
+a10H2+a11PLH+a12L3+a13LP2+a14LH2+a15L2P+a16P3+a17PH2
+a18L2H+a19P2H+a20H3
DenL(P, L, H)=b1+b2L+b3P+b4H+b5LP+b6LH+b7PH+b8L2+b9P2
+b10H2+b11PLH+b12L3+b13LP2+b14LH2+b15L2P+b16P3+b17PH2
+b18L2H+b19P2H+b20H3
Nums(P, L, H)=c1+c2L+c3P+c4H+c5LP+c6LH+c7PH+c8L2+c9P2
+c10H2+c11PLH+c12L3+c13LP2+c14LH2+c15L2P+c16P3+c17PH2
+c18L2H+c19P2H+c20H3
Dens(P, L, H)=d1+d2L+d3P+d4H+d5LP+d6LH+d7PH+d8L2+d9P2
+d10H2+d11PLH+d12L3+d13LP2+d14LH2+d15L2P+d16P3+d17PH2
+d18L2H+d19P2H+d20H3
Wherein, ai,bi,ci,di(i=1~20) are rational function model parameter, b1And d1Usually 1, (P, L, H) is to return
One ground coordinate changed, (X, Y) are normalized image coordinate,
Normalization mode is as follows：
In formula, LINE_SCALE, SAMP_SCALE, SAMP_OFF and LINE_OFF are the normalized parameter of image space coordinate；
LAT_OFF, LON_OFF, HEIGHT_OFF, LAT_SCALE, LON_SCALE, HEIGHT_SCALE are object coordinates normalization ginseng
Number, lat is latitude, and lon is longitude, and h is elevation.
2.3) by the virtual controlling point obtained in 2.1), 78 rational function model parameters are solved by least square, by
Obtained rational function model parameter is solved, the rational function model of every image can be established.
3, the same place of remote sensing image is obtained to carrying out initial matching to stereopsis.
3.1) when the same place of image refers to that image photographs to areal, same atural object on different images it is corresponding at
Picture point.When direction of illumination is identical, pending stereopsis pair is obtained by step 1, in the overlapping region of stereopsis pair,
Obtain the same place being evenly distributed.The same place of image obtains by the way of Image Matching, identical vertical for illumination condition
Automatic image matching may be used in body image pair, obtains equally distributed same place to upper in stereopsis, wherein automatic shadow
As matching algorithm has：Correlation coefficient matching method, SIFT (Scale-invariant feature transform) or SURF
(Speeded Up Robust Feature) characteristic matching etc.；
3.2) when direction of illumination difference, same place can be chosen between image by way of artificial interpretation, is then used
Random sampling consistency (RANdomSAmple Consensus, RANSAC) algorithm or other elimination of rough difference algorithms are to Mismatching point
It is rejected, retains the correct match point of each stereopsis pair.
4, more stereopsis regional network bundle adjustments, the imaging geometry model after being refined
Due to the presence of the systematic errors such as satellite position, sensor attitude angle and camera lens distortion, pass through step 2 structure
The rigorous geometric model built and the rational function model solved by rigorous geometric model all include inevitably positioning
Error.It needs to carry out object space by adjustment or image space is compensated, to improve the positioning of remote sensing image target and cartographic accuracy.
In order to keep the stability of block adjustment process more preferable, in more image adjustments, it is still necessary to certain auxiliary information
Carry out fixed area net.By taking common rational function model adds affine Transform Model as an example, the sensor after refining is molded geometry
Model formulation is as follows：
In formula, (sample, line) is that the image space after normalization measures coordinate, and (x, y) is that the calculated image spaces of RFM are sat
Mark, (Δ x, Δ y) be respectively line direction and column direction picture point correction (being shown below), is corrected with this since image is original
Deviation between the imaging geometry model ground point back projection coordinate brought there are error and actual image point coordinate, form is such as
Shown in following formula：
Δ x=a0+a1·sample+a2·line+…
Δ y=b0+b1·sample+b2·line+…(8)
In formula, a0,a1,a2... and b0,b1,b2... it is imitative for common image space when including only first three items for correction parameter
Penetrate transformation model.
The present invention is correction parameter with image space affine Transform Model, is described in detail to solve in regional network bundle adjustment and correct
The detailed process of parameter:
4.1) adjustment error equation is built, to establishing error equation according to formula (7), (8)：
In formula, FxAnd FyIndicate x, y-coordinate error equation function, a0,a1,a2And b0,b1,b2To correct parameter, (sample,
Line coordinate) is measured for the image space after normalization, (x, y) is image space coordinate, and subscript i is a labelled notation.
4.2) error equation solution is carried out
First order Taylor series expansion, the error equation that can be linearized are carried out to (9)：
In formula, lat is latitude, and lon is longitude, and h is elevation, Fx0, Fy0Respectively value of the error function at x0, y0,Respectively phase
According to the first-order partial derivative of parameter, Δ a0、Δa1、Δa2、Δb0、Δb1、Δb2, Δ lat, Δ lon and Δ h be respectively relevant parameter
Corrected value.
Obtain indirect adjustment model：
V=AX+BY-L, P (11)
In formula, X is adjustment parameter：
X=[Δ a0 Δa1 Δa2 Δb0 Δb1 Δb2](12)
A is the coefficient matrix of unknown number X：
Y is the ground coordinate of tie point：
Y=[Δ lat Δ lon Δs h] (14)
B is the coefficient matrix of unknown number Y：
For control point, B 0, L are constant term, can be calculated by initial value and formula (9)：
In formula, P is weight matrix.For control point, unknown number is adjustment parameter (a0,a1,a2,b0,b1,b2)；And it is right
For tie point, unknown number includes adjustment parameter and corresponding object coordinates (lat, lon, h).
When existing simultaneously control point and tie point, you can establish following error equation：
For the error equation of two class unknown numbers, the second class unknown number Y changed normal equations accordingly can be eliminated：
The solution of unknown number X is completed using formula (18).
4.3) stereopsis obtains virtual controlling point to adjustment respectively
According to step 4.1) and the method 4.2) provided, no control point is carried out respectively to every group to multigroup stereopsis respectively
Adjustment, select image space back projection error is minimum in adjustment one group namely adjustment with high accuracy one group.Nothing during the adjustment
The participation at palpus control point, since image is stereopsis pair two-by-two, adjustment process is stablized.Utilize the reasonable of this group of stereopsis
Function model and adjustment corrected value and this on image more degree overlapping tie point (the same point is in more than two shadows i.e. on ground
The match point extracted is matched on picture) this to the picpointed coordinate on image, forward intersection goes out more degree overlapping tie points
Lunar surface three-dimensional coordinate, using these lunar surface three-dimensional coordinates as virtual controlling point coordinates 4.4) overall adjustment obtains all images and changes
Positive number
By all stereopsis to carrying out adjustment with the method 4.2) provided using step 4.1), and using in step 4.3)
Virtual controlling point as control.
5, to each group of stereopsis to carrying out dense Stereo Matching, the image space for respectively obtaining each group stereopsis pair is intensive of the same name
Point set.
Dense Stereo Matching is to stereopsis to carrying out pixel-by-pixel or the matching of 3~5 pixels in interval, and when specific implementation can
Dense Stereo Matching is guided based on the correct same place obtained in by initial matching.
6, DEM is merged
High-precision imaging geometry model after being refined by what is obtained in step 4, the dense Stereo Matching obtained in step 5 is same
Famous cake carries out forward intersection, obtains the dense three-dimensional point cloud of object space.Regular grid division is carried out to survey region, by interpolation side
Method obtains the height value at grid using the point off density cloud level journey value interpolation around grid points.Each group of stereopsis is to all may be used
To obtain one group of object space point off density cloud by this method, and then DEM is obtained by grid partition and interpolation.
As previously described, because the reasons such as illumination and landform, can there are some shadow regions or highlight bar, these regions on image
It is often difficult to obtain correct match point in the dense Stereo Matching of step 5, to cause the sky of image space or even object space point off density cloud
It lacks.The present invention utilizes the characteristics of multigroup stereopsis under different illumination conditions to obtaining, and is carried out to its shadow region and highlight bar
Complementation, cavity caused by repair major part since it fails to match.It is repaiied by stereopsis to generating DEM and carrying out cavity
Benefit is as follows：
6.1) according to stereopsis to the object space range of overlay region, to every group of stereopsis to establishing same unit interval
Regular grid.The image resolution that general identical sensor obtains will not differ too many, by taking NAC cameras as an example, the moon of acquisition
Face image resolution is generally in the range of 0.5m~2.0m, it is possible to different stereopsis between taking identical unit
Every；
6.2) centered on each grid points, search radius is the point off density within the scope of R, and the selection of radius R should be able to be protected
Demonstrate,proving the points that each grid point search obtains should ensure to complete the required minimal point of interpolation, and be unlikely to the point for having excessive
Participate in operation；
6.3) when search result is that points that are empty or searching are unsatisfactory for requiring, it is believed that the grid points are empty point, note
Record its number and plane coordinates information (ID, X, Y)；Otherwise grid points are obtained according to the elevation value interpolation of the point off density searched
Height value；
6.4) it repeats the above steps, by each stereopsis to obtaining the corresponding number of cavity position on a DEM and DEM
And coordinate information；
6.5) according to each stereopsis to the number of the adjustment result in step 4.2) and empty grid points on each DEM
Amount, select error in adjustment checkpoint small and an empty DEM at least is as base map, and sky is carried out to it using other DEM
It repairs in hole；
Since the block adjustment of more images has been carried out, eliminated as far as possible since what systematic error was brought differs
Cause property, so no longer needing to be registrated in DEM fusion process, the detailed process of cavity repairing is：
6.5.1 the plane coordinates for) taking each cavity point on DEM base maps successively, with n (n=1~2) a lattice on other DEM
It is divided into radius between net, searches the grid height value within the scope of this；
6.5.2) if on several DEM simultaneously search met the requirements as a result, if by stereopsis pair adjustment essence
Degree and with cavity put between at a distance from be weighted, be finally averaged to obtain the height value of the point；The process that specific fixed power solves
It is as follows：It is arbitrary to take in search result error RMS in the adjustment of i-th group of stereopsis pairiAnd j-th search result point and empty point
Distance DjOn the basis of, then the adjustment precision RMS of arbitrary kth group stereopsis pairkWith the ratio R MS of the benchmarki/RMSkFor kth
The adjustment precision weights of group stereopsis pair, any one searches point h and cavity point distance DhWith the ratio D of reference distancej/
DhIndicate as follows for the height value H apart from weights, final unknown point：
In formula, i, k=1,2 ..., m, m indicate the quantity of stereopsis pair；H, j=1,2 ..., Pn,PnIndicate search result
The quantity of point, HhIndicate h-th of height value for searching point position.
6.5.3) if only searched on a width DEM it is meeting the requirements as a result, if by with cavity put between at a distance from carry out
Weighted average acquires the height value of cavity point as the following formula；
6.5.4) if the point is empty point on other all DEM corresponding positions, it is numbered and is stored again
Its plane coordinates；
6.6) repairing of all cavity points is completed according to step 6.5).For being finally still recorded as the position of cavity point,
The method for then using traditional interpolation, the height value of the position is obtained by grid points interpolation around the DEM；
6.7) since the fusion of DEM carries out pixel-by-pixel, there may be mutation by the DEM after fusion.It is picked using exceptional value
Except algorithm, mutated site is detected, and the reasonable height value of mutated site is obtained according to the pixel interpolating of surrounding.So far complete
At more stereopsis DEM fusion process, the high accuracy DEM of topographic details is obtained without cavity and retained.
7, DOM is generated
The video imaging geometrical model without empty DEM and after refining obtained by fusion, by each ground point edge on DEM
Light back projection photograph to image space, and gray scale resampling is carried out according to raw video, obtains after topographical correction with geography
The image of information, i.e. DOM products.The generation of DOM products is generally divided into positive solution and anti-solution, is described by taking anti-solution as an example here
Specific implementation process：
7.1) topocentric coordinates are calculated
If the picpointed coordinate of any point pixel center P is (x on pre-generatmg DOMdom,ydom), by the starting point ground of DOM
Coordinate (X0,Y0) (namely starting point ground coordinate of DEM) ground coordinate corresponding with DOM scale denominators M calculating P points (X,
Y)；
X=X0+M*xdom
Y=Y0+M*xdom
7.2) picpointed coordinate is calculated
Using the video imaging geometrical model established and after refining, corresponding picpointed coordinate p (x, y) on original image is calculated；
X=f1 (X, Y, Z)
Y=f2 (X, Y, Z)
In formula, f1 and f2 represent the video imaging geometrical model after refining, and Z indicates the elevation of P points, obtained by DEM interpolations.
7.3) gray scale interpolation
It, thus must be into since obtained picpointed coordinate p (x, y) not necessarily falls in the pixel center of raw video
Row gray scale interpolation, generally can be used bilinear interpolation, acquire the gray value g (x, y) of picture point p；
7.4) gray scale assignment
The gray value of picture point p is assigned to the P points on pixel namely DOM after correcting；
7.5) above-mentioned operation is completed to each pixel on DOM successively, the DOM products by topographical correction can be obtained.
The various embodiments described above are merely to illustrate the present invention, and the implementation steps of wherein method may be changed, all
The equivalents carried out based on the technical solution of the present invention and improvement, should not exclude protection scope of the present invention it
Outside.
Claims (6)
1. a kind of more stereopsis considering different illumination image-forming conditions merge drafting method, it is characterised in that including following step
Suddenly：
Step 1)：It obtains survey region and has image, and select Multi folds coverage stereopsis pair；
Step 2)：Build the imaging geometry model of remote sensing image；
Step 3)：To stereopsis to carrying out initial matching, the same place of remote sensing image is obtained；
Step 4)：Imaging geometry model is obtained to step 2) and carries out stereopsis regional network bundle adjustment, after being refined
Imaging geometry model；
Step 5)：To each group of stereopsis to carrying out dense Stereo Matching, the image space for respectively obtaining each group stereopsis pair is intensive same
Name point set；
Step 6)：Imaging geometry model after being refined according to step 4) carries out the intensive same place of the image space obtained in step 5)
Forward intersection obtains the dense three-dimensional point cloud of object space, and then obtains DEM by grid partition and interpolation, and is merged to DEM
It obtains without empty DEM；
Step 7)：According to the imaging geometry model without empty DEM and after refining, generating DOM is：With ground after topographical correction
Manage the image of information.
2. considering more stereopsis fusion drafting method of different illumination image-forming conditions as described in claim 1, feature exists
In the specific choice process of the step 1) is：According between image intersection angle, video imaging when solar azimuth and height
Angle, the aspect ratio of image, image overlap range select multigroup stereopsis pair, form two images of each group of stereopsis pair
Solar azimuth when three-dimensional intersection angle is more than 10 °, is imaged should be met and elevation angle is essentially identical and image aspect ratio connects
Nearly 1, and should have different sun altitudes or azimuth between multigroup stereopsis pair and exist between each other as far as possible
Big overlapping range.
3. considering more stereopsis fusion drafting method of different illumination image-forming conditions as described in claim 1, feature exists
In the step 2) includes the structure of stringent imaging geometry model and general geometrical model, wherein general geometrical model is multinomial
Formula model, direct linear transformation's model or rational function model.
4. as claims 1 to 3 any one of them considers more stereopsis fusion drafting method of different illumination image-forming conditions,
It is characterized in that, the detailed process of the step 3) is：
Step 3.1)：When direction of illumination is identical, in the overlapping region of stereopsis pair, three-dimensional shadow identical for illumination condition
As right, the same place of image is matched using automatic image obtains equally distributed same place in stereopsis to upper；
Step 3.2)：When direction of illumination difference, same place is chosen between image by way of artificial interpretation, then to error hiding
Point is rejected, and the correct match point of each stereopsis pair is retained.
5. considering more stereopsis fusion drafting method of different illumination image-forming conditions as claimed in claim 2, feature exists
In the detailed process of the step 4) is：
Step 4.1)：Build adjustment error equation：
Fx=samplei+a0+a1·samplei+a2·linei-xi=0
Fy=linei+b0+b1·samplei+b2·linei-yi=0
In formula, FxAnd FyIndicate x, y-coordinate error equation function, a0,a1,a2And b0,b1,b2To correct parameter, (sample,
Line coordinate) is measured for the image space after normalization, (x, y) is image space coordinate, and subscript i is a labelled notation；
Step 4.2)：Carry out error equation solution：First order Taylor series expansion, the error linearized are carried out to error equation
Equation：
In formula, lat is latitude, and lon is longitude, and h is elevation, Fx0, Fy0Respectively value of the error function at x0, y0,Respectively phase
Answer the first-order partial derivative of parameter, Δ a0、Δa1、Δa2、Δb0、Δb1、Δb2, Δ lat, Δ lon and Δ h be respectively relevant parameter
Corrected value；
Obtain indirect adjustment model：
V=AX+BY-L, P
In formula, X is adjustment parameter：
X=[Δ a0 Δa1 Δa2 Δb0 Δb1 Δb2]
A is the coefficient matrix of unknown number X：
Y is the ground coordinate of tie point：
Y=[Δ lat Δ lon Δs h]
B is the coefficient matrix of unknown number Y：
In formula, LINE_SCALE, SAMP_SCALE be image space coordinate normalized parameter, LAT_SCALE, LON_SCALE,
HEIGHT_SCALE is object coordinates normalized parameter；
For control point, B 0, L are constant term, can be calculated by initial value and error equation：
In formula, P is weight matrix, and for control point, unknown number is adjustment parameter (a0,a1,a2,b0,b1,b2)；And to connection
For point, unknown number includes adjustment parameter and corresponding object coordinates (lat, lon, h)；
When existing simultaneously control point and tie point, you can establish following error equation：
For the error equation of two class unknown numbers, the second class unknown number Y changed normal equations accordingly are eliminated：
The solution of unknown number X is completed by above formula；
Step 4.3)：Stereopsis obtains virtual controlling point to adjustment respectively
According to step 4.1) and 4.2), the adjustment at no control point is carried out respectively to every group to multigroup stereopsis, select adjustment essence
Spend high one group, i.e., image space back projection error is minimum in adjustment one group, using this group of stereopsis rational function model and
Adjustment correction value and this to more degree overlapping tie point on image this to the picpointed coordinate on image, forward intersection goes out more degree weights
The lunar surface three-dimensional coordinate of folded tie point, using these lunar surface three-dimensional coordinates as virtual controlling point coordinates；
Step 4.4)：Overall adjustment obtains all image corrections
By all stereopsis to carrying out adjustment with the method 4.2) provided using step 4.1), and utilize the void in step 4.3)
Quasi- control point is as control.
6. considering more stereopsis fusion drafting method of different illumination image-forming conditions as claimed in claim 5, feature exists
In the detailed process of the step 6)：
Step 6.1)：According to stereopsis to the object space range of overlay region, to every group of stereopsis to establishing same unit interval
Regular grid；
Step 6.2)：Centered on each grid points, search radius is the point off density within the scope of R；
Step 6.3)：When search result is that points that are empty or searching are unsatisfactory for requiring, it is believed that the grid points are empty point, note
Record its number and plane coordinates information (ID, X, Y)；Otherwise grid points are obtained according to the elevation value interpolation of the point off density searched
Height value；
Step 6.4)：It repeats the above steps, by each stereopsis to obtaining the corresponding volume of cavity position on a DEM and DEM
Number and coordinate information；
Step 6.5)：According to each stereopsis to the adjustment result in step 4.2) and empty grid points on each DEM
Quantity, selects error in adjustment checkpoint small and an empty minimum DEM is as base map, is carried out to it using other DEM
Cavity repairing, wherein cavity repairing detailed process be：
Step 6.5.1)：The plane coordinates of each cavity point on DEM base maps is taken successively, it is a with n (n=1~2) on other DEM
Mesh spacing is radius, searches the grid height value within the scope of this；
Step 6.5.2)：If on several DEM simultaneously search met the requirements as a result, if by stereopsis pair adjustment essence
Degree and with cavity put between at a distance from be weighted, be finally averaged to obtain the height value of the point；The process that specific weighting solves
It is as follows：It is arbitrary to take in search result error RMS in the adjustment of i-th group of stereopsis pairiAnd j-th search result point and empty point
Distance DjOn the basis of, then the adjustment precision RMS of arbitrary kth group stereopsis pairkWith the ratio R MS of the benchmarki/RMSkFor kth
The adjustment precision weights of group stereopsis pair, any one searches point h and cavity point distance DhWith the ratio D of reference distancej/
DhIndicate as follows for the height value H apart from weights, final unknown point：
In formula, i, k=1,2 ..., m, m indicate the quantity of stereopsis pair；H, j=1,2 ..., Pn,PnIndicate search result point
Quantity, HhIndicate h-th of height value for searching point position.
Step 6.5.3)：If only searched on a width DEM it is meeting the requirements as a result, if by with cavity put between at a distance from carry out
Weighted average acquires the height value of cavity point as the following formula；
Step 6.5.4)：If the point is empty point on other all DEM corresponding positions, it is numbered and is stored again
Its plane coordinates；
Step 6.6)：The repairing of all cavity points is completed according to step 6.5).For being finally still recorded as the position of cavity point,
The method for then using traditional interpolation, the height value of the position is obtained by grid points interpolation around the DEM；
Step 6.7)：Using abnormality value removing algorithm, mutated site is detected, and is dashed forward according to the pixel interpolating of surrounding
The reasonable height value for becoming position, so far completes more stereopsis DEM fusion process, obtains without cavity and retains topographic details
DEM。
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810062375.9A CN108305237B (en) | 2018-01-23 | 2018-01-23 | Multi-stereo image fusion drawing method considering different illumination imaging conditions |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810062375.9A CN108305237B (en) | 2018-01-23 | 2018-01-23 | Multi-stereo image fusion drawing method considering different illumination imaging conditions |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108305237A true CN108305237A (en) | 2018-07-20 |
| CN108305237B CN108305237B (en) | 2021-09-21 |
Family
ID=62865910
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810062375.9A Active CN108305237B (en) | 2018-01-23 | 2018-01-23 | Multi-stereo image fusion drawing method considering different illumination imaging conditions |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108305237B (en) |
Cited By (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109579796A (en) * | 2018-12-24 | 2019-04-05 | 中国科学院遥感与数字地球研究所 | A kind of block adjustment method of image after projection |
| CN109903352A (en) * | 2018-12-24 | 2019-06-18 | 中国科学院遥感与数字地球研究所 | A kind of seamless orthography production method in the big region of satellite remote-sensing image |
| CN109934788A (en) * | 2019-03-22 | 2019-06-25 | 鲁东大学 | A kind of remote sensing images missing data restorative procedure based on standard remote sensing images |
| CN110189283A (en) * | 2019-05-21 | 2019-08-30 | 西安电子科技大学 | Remote sensing images DSM fusion method based on semantic segmentation figure |
| CN110322557A (en) * | 2019-06-18 | 2019-10-11 | 中南林业科技大学 | A kind of travelling error compensation method that combination GPS is merged with SRTM |
| CN110388898A (en) * | 2019-06-27 | 2019-10-29 | 中国科学院遥感与数字地球研究所 | Construct the multiple coverage remote sensing image error compensation method of multi-source of virtual controlling point constraint |
| CN110741409A (en) * | 2018-09-25 | 2020-01-31 | 深圳市大疆创新科技有限公司 | Image-related data processing method, unmanned aerial vehicle, system and storage medium |
| CN111156969A (en) * | 2020-02-07 | 2020-05-15 | 黄文超 | Wide remote sensing image stereo mapping method and system |
| CN111754458A (en) * | 2020-05-18 | 2020-10-09 | 北京吉威空间信息股份有限公司 | Satellite image three-dimensional space reference frame construction method oriented to geometric precision processing |
| CN112082576A (en) * | 2020-09-09 | 2020-12-15 | 桂林理工大学 | Satellite image orthorectification method under 'three-condition-free' condition |
| CN112146629A (en) * | 2020-09-24 | 2020-12-29 | 武汉大学 | Multi-angle close-up photography track and attitude planning method |
| CN113358091A (en) * | 2021-06-02 | 2021-09-07 | 自然资源部国土卫星遥感应用中心 | Method for producing digital elevation model by using three-linear array three-dimensional satellite image |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CA3089200A1 (en) * | 2018-01-25 | 2019-08-01 | Geomni, Inc. | Systems and methods for rapid alignment of digital imagery datasets to models of structures |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100328682A1 (en) * | 2009-06-24 | 2010-12-30 | Canon Kabushiki Kaisha | Three-dimensional measurement apparatus, measurement method therefor, and computer-readable storage medium |
| CN103310487A (en) * | 2013-06-21 | 2013-09-18 | 中国科学院遥感与数字地球研究所 | Generating method for universal time variable based imaging geometric model |
-
2018
- 2018-01-23 CN CN201810062375.9A patent/CN108305237B/en active Active
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20100328682A1 (en) * | 2009-06-24 | 2010-12-30 | Canon Kabushiki Kaisha | Three-dimensional measurement apparatus, measurement method therefor, and computer-readable storage medium |
| CN103310487A (en) * | 2013-06-21 | 2013-09-18 | 中国科学院遥感与数字地球研究所 | Generating method for universal time variable based imaging geometric model |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110741409A (en) * | 2018-09-25 | 2020-01-31 | 深圳市大疆创新科技有限公司 | Image-related data processing method, unmanned aerial vehicle, system and storage medium |
| WO2020061772A1 (en) * | 2018-09-25 | 2020-04-02 | 深圳市大疆创新科技有限公司 | Data processing method regarding image, and unmanned aerial vehicle, system and storage medium |
| CN109579796B (en) * | 2018-12-24 | 2020-11-10 | 中国科学院遥感与数字地球研究所 | Area network adjustment method for projected image |
| CN109903352A (en) * | 2018-12-24 | 2019-06-18 | 中国科学院遥感与数字地球研究所 | A kind of seamless orthography production method in the big region of satellite remote-sensing image |
| CN109579796A (en) * | 2018-12-24 | 2019-04-05 | 中国科学院遥感与数字地球研究所 | A kind of block adjustment method of image after projection |
| CN109934788A (en) * | 2019-03-22 | 2019-06-25 | 鲁东大学 | A kind of remote sensing images missing data restorative procedure based on standard remote sensing images |
| CN110189283A (en) * | 2019-05-21 | 2019-08-30 | 西安电子科技大学 | Remote sensing images DSM fusion method based on semantic segmentation figure |
| CN110189283B (en) * | 2019-05-21 | 2021-10-29 | 西安电子科技大学 | Remote sensing image DSM fusion method based on semantic segmentation graph |
| CN110322557A (en) * | 2019-06-18 | 2019-10-11 | 中南林业科技大学 | A kind of travelling error compensation method that combination GPS is merged with SRTM |
| CN110388898A (en) * | 2019-06-27 | 2019-10-29 | 中国科学院遥感与数字地球研究所 | Construct the multiple coverage remote sensing image error compensation method of multi-source of virtual controlling point constraint |
| CN111156969A (en) * | 2020-02-07 | 2020-05-15 | 黄文超 | Wide remote sensing image stereo mapping method and system |
| CN111754458A (en) * | 2020-05-18 | 2020-10-09 | 北京吉威空间信息股份有限公司 | Satellite image three-dimensional space reference frame construction method oriented to geometric precision processing |
| CN111754458B (en) * | 2020-05-18 | 2023-09-15 | 北京吉威空间信息股份有限公司 | Satellite image three-dimensional space reference frame construction method for geometric fine processing |
| CN112082576A (en) * | 2020-09-09 | 2020-12-15 | 桂林理工大学 | Satellite image orthorectification method under 'three-condition-free' condition |
| CN112146629A (en) * | 2020-09-24 | 2020-12-29 | 武汉大学 | Multi-angle close-up photography track and attitude planning method |
| CN113358091A (en) * | 2021-06-02 | 2021-09-07 | 自然资源部国土卫星遥感应用中心 | Method for producing digital elevation model by using three-linear array three-dimensional satellite image |
| CN113358091B (en) * | 2021-06-02 | 2021-11-16 | 自然资源部国土卫星遥感应用中心 | Method for producing digital elevation model DEM (digital elevation model) by using three-linear array three-dimensional satellite image |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108305237B (en) | 2021-09-21 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108305237A (en) | Consider more stereopsis fusion drafting method of different illumination image-forming conditions | |
| CN109903352B (en) | Method for making large-area seamless orthoimage of satellite remote sensing image | |
| CN110388898B (en) | Multisource multiple coverage remote sensing image adjustment method for constructing virtual control point constraint | |
| Henriksen et al. | Extracting accurate and precise topography from LROC narrow angle camera stereo observations | |
| Li et al. | Rigorous photogrammetric processing of HiRISE stereo imagery for Mars topographic mapping | |
| Wang et al. | Epipolar resampling of linear pushbroom satellite imagery by a new epipolarity model | |
| CN111126148B (en) | DSM (digital communication system) generation method based on video satellite images | |
| CN107492069B (en) | Image fusion method based on multi-lens sensor | |
| CN104299228B (en) | A kind of remote sensing image dense Stereo Matching method based on Accurate Points position prediction model | |
| CN108830889A (en) | The matching process of remote sensing image and reference images based on global geometrical constraint | |
| CN107705329A (en) | High-resolution optical satellite based on geometrical constraint stares Image registration method | |
| Albertz et al. | HRSC on Mars Express–Photogrammetric and cartographic research | |
| CN106780321A (en) | A kind of overall tight orientation of the satellite HR sensors images of CBERS 02 and correction joining method | |
| CN113358091B (en) | Method for producing digital elevation model DEM (digital elevation model) by using three-linear array three-dimensional satellite image | |
| CN110111274A (en) | A kind of spaceborne push-broom type optical sensor elements of exterior orientation calibrating method | |
| Gong et al. | Point cloud and digital surface model generation from high resolution multiple view stereo satellite imagery | |
| CN110986888A (en) | Aerial photography integrated method | |
| CN113739767A (en) | Method for producing orthoimage aiming at image acquired by domestic area array swinging imaging system | |
| CN109579796B (en) | Area network adjustment method for projected image | |
| Savopol et al. | Modelling of the IRS-1C satellite pan stereo-imagery using the DLT approach | |
| Zhou et al. | DEM generation using Ziyuan-3 mapping satellite imagery without ground control points | |
| Reinartz et al. | Accuracy analysis for DEM and orthoimages derived from SPOT HRS stereo data without using GCP | |
| CN112950763A (en) | Live-action modeling method in transformer substation engineering | |
| Lohse et al. | Multi-image shape-from-shading: Derivation of planetary digital terrain models using clementine images | |
| Rosiek et al. | Utility of Viking Orbiter images and products for Mars mapping |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |