CN106846251A - Orthography based on building roof vector inlays gauze network automatic selecting method - Google Patents

Orthography based on building roof vector inlays gauze network automatic selecting method Download PDF

Info

Publication number
CN106846251A
CN106846251A CN201710059088.8A CN201710059088A CN106846251A CN 106846251 A CN106846251 A CN 106846251A CN 201710059088 A CN201710059088 A CN 201710059088A CN 106846251 A CN106846251 A CN 106846251A
Authority
CN
China
Prior art keywords
building
imaging region
gauze network
line
inlays
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
Application number
CN201710059088.8A
Other languages
Chinese (zh)
Other versions
CN106846251B (en
Inventor
李朋龙
丁忆
徐永书
胡艳
张治清
金贤锋
李胜
卢涛
李静
李碧香
肖勇
罗鼎
蒲德祥
张少佳
段松江
吴凤敏
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Chongqing geographic information and Remote Sensing Application Center (Chongqing surveying and mapping product quality inspection and testing center)
Original Assignee
CHONGQING GEOGRAPHICAL INFORMATION CENTER
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by CHONGQING GEOGRAPHICAL INFORMATION CENTER filed Critical CHONGQING GEOGRAPHICAL INFORMATION CENTER
Priority to CN201710059088.8A priority Critical patent/CN106846251B/en
Publication of CN106846251A publication Critical patent/CN106846251A/en
Application granted granted Critical
Publication of CN106846251B publication Critical patent/CN106846251B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T3/00—Geometric image transformations in the plane of the image
    • G06T3/40—Scaling of whole images or parts thereof, e.g. expanding or contracting
    • G06T3/4038—Image mosaicing, e.g. composing plane images from plane sub-images
    • 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 using two or more images, e.g. averaging or subtraction
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10032—Satellite or aerial image; Remote sensing
    • 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

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Image Processing (AREA)

Abstract

Gauze network automatic selecting method is inlayed the invention discloses a kind of orthography based on building roof vector, including building roof vector combination digital complex demodulation automatically generates building simple model；Imaging region of the building on monolithic DOM is calculated using building simple model；Gauze network is inlayed using surveying all image plate nadir point positions in area and automatically generating the survey initial Voronoi diagram in area, and simplify treatment；Using imaging region of the building on monolithic DOM, all nodes are in optimized selection, obtain suboptimum and inlay gauze network；All lines of inlaying are in optimized selection in inlaying gauze network to suboptimum again, are entirely surveyed the overall optimal seam-line network that area gets around building imaging region.Its remarkable result is：Breach tradition and first carry out individual and just penetrating corrections, recycle monolithic DOM inlay the pattern of gauze network selection, solve the problems, such as to survey area inlays that gauze network cannot automatically bypass building imaging region and to cause texture dislocation and existing method to automatically select efficiency low.

Description

Orthography based on building roof vector inlays gauze network automatic selecting method
Technical field
It is that one kind is sweared based on building roof specifically the present invention relates to aviation orthography embedding technique field The orthography of amount inlays gauze network automatic selecting method.
Background technology
Digital orthoimage (Digital Orthophoto Map abbreviation DOM) is that digital remote sensing image utilizes digital elevation Model (Digital Elevation Model vehicle economy M), corrects by individual image digital differential, then inlay melting Close, the digital image data for finally being cut according to certain map sheet.DOM is that digital photogrammetry is produced with the important 4D of remote sensing The significant data source that one of product, even more Fundamental Geographic Information Data are produced and updated, orthography is widely used in state The fields such as soil resource investigation, urban planning and construction, land use monitoring, digital city construction.With unmanned plane low-altitude remote sensing system The development of system, DOM also plays more and more important role in the quick the condition of a disaster assessment of natural calamity, emergency repair.
The production of DOM striographs mainly has individual image just penetrating correction, image light and color homogenization, image mosaic, image to cut Four steps, wherein image mosaic are links the most time-consuming the most complicated in DOM Imagery Maps, and are inlayed in image mosaic The selection of rule is the most important thing.So-called image mosaic refers to by two or multiple have the orthography unification of overlay region to together Under one coordinate system, a boundary line is then selected in image overlap area, the pixel on boundary line both sides selection different images, So as to multiple orthographies to be merged into the process of a width frame width bigger, orthography of texture seamless connection.And this edge circle Line is inlays line, by it is a plurality of inlay the network that line constitutes and turn into inlay gauze network.Most software is being carried out both at home and abroad at present The selection that gauze network is inlayed during image mosaic is required for being aided with man-machine interactively editor and could meet Production requirement, and this just hinders The process that the high-efficient automatic that DOM makes makes.
It is to entangle the raw video of central projection pixel-by-pixel by Differential rectification according to DEM terrain datas just to penetrate correction Just it is being orthography.Because DEM merely illustrates surface configuration, not including spatial surface target such as artificial structure's information, Therefore on the orthography after just penetrating and correcting still there is height displacement in building, building can deviate its correct position and There is inclined situation.Discounting for targets such as buildings, directly utilize and inlay gauze network by what geometrical relationship was automatically generated (such as Voronoi diagram inlays gauze network) carries out image mosaic, just occurs when line passing through building imaging region is inlayed and inlays Line both sides texture dislocation, the problems such as imperfect.Therefore want to realize the automation of image mosaic, it is necessary to by inlaying the automatic of line Detection makes it successfully bypass the targets such as building.
At present, inlaying automatically selecting for line mainly has two class methods, and the first kind is to build difference image based on left and right image, Then the minimum communication path of difference is searched in difference image.Such method is substantially Pixel-level, and differential expression is in list Individual pixel, not there will be the atural object of disparity as overall expression, thus inlay line unavoidably can be through the area such as building Domain, and algorithm is sufficiently complex, inefficiency；Equations of The Second Kind is that to inlay line automatic with the auxiliary such as DSM or building mapping data Search bypasses building.But the characteristics of distortion of projection that above method does not all account for image becomes big to image edge.Simultaneously Most methods are only optimized selection to the line of inlaying between adjacent image, do not provide the whole area that surveys and inlay the excellent of gauze network Change the scheme of selection.
The Voronoi diagram that position relationship based on sequential images is automatically generated inlays gauze network and can be good at selection every The minimum part of distortion of projection is used as effectively inlaying region on image, but cannot ensure that inlaying line bypasses building.And build The mapping data of thing are easy to be obtained in Fundamental Geographic Information Databases at different levels, thus using building mapping data and Voronoi diagram is inlayed gauze network and is effectively combined, automatically generate bypass construction zone inlay gauze network for solution inlay line The difficult problem of the intelligent selection of network is significant.
The content of the invention
In view of the shortcomings of the prior art, it is an object of the invention to provide a kind of orthography based on building roof vector Inlay gauze network automatic selecting method, the method initially inlays gauze network, then be based on using surveying all plate nadir points in area and automatically generate Building roof vector is in optimized selection to inlaying gauze network, and that is entirely surveyed that area bypasses building imaging region inlays line Network, small with amount of calculation, the degree of accuracy is high, the advantages of in hgher efficiency, and can effectively reduce the image side that central projection is caused The big influence of edge distortion of projection.
To reach above-mentioned purpose, the technical solution adopted by the present invention is as follows：
A kind of orthography based on building roof vector inlays gauze network automatic selecting method, its it is critical only that according to Following steps are carried out：
Step 1：Building roof vector combination digital complex demodulation automatically generates building simple model；
Step 2：Imaging region of the building on monolithic DOM is calculated using building simple model；
Step 3：Using survey all image plate nadir point positions in area automatically generate survey the initial Voronoi diagram in area inlay gauze network, And it is carried out simply to be inlayed gauze network after simplifying treatment；
Step 4：Imaging region of the building obtained using step 2 on monolithic DOM, the institute to simply inlaying gauze network There is node to be in optimized selection, obtain suboptimum and inlay gauze network；
Step 5：Imaging region of the building obtained using step 2 on monolithic DOM, the institute of gauze network is inlayed to suboptimum Have and inlay line and be in optimized selection, entirely surveyed the overall optimal seam-line network that area gets around building imaging region.
Further description is that the calculation procedure of imaging region of the building described in step 2 on monolithic DOM is as follows：
Step 2-1：By on all tri patch transmission projections of building simple model to raw video, original shadow is tried to achieve As tri patch；
Step 2-2：Union is asked to raw video tri patch, imaging region of the building on raw video is obtained；
Step 2-3：Solved based on DEM iterative approach, building is obtained corresponding single by the imaging region on raw video Imaging region on piece DOM.
Further description is that the obtaining step that initial Voronoi diagram inlays gauze network described in step 3 is as follows：
Step 3-1：The principal point that all images in area will be surveyed projects to ground, obtains its corresponding plate nadir point；
Step 3-2：It is scatterplot collection to survey all plate nadir points in area, Delaunay Triangulation is carried out using pointwise interpolation method；
Step 3-3：According to triangulation result, generate initial Voronoi diagram and inlay gauze network.
Further description is that the simplified process step that initial Voronoi diagram inlays gauze network in step 3 is：
Step S1：Calculate the length that initial Voronoi diagram inlays the polygonal four edges most long of certain Voronoi in gauze network Degree average value；
Step S2：Obtain short side of the length less than average value a quarter in above-mentioned Voronoi polygons；
Step S3：The short side is substituted with the geometry midpoint of two end points of short side, the Voronoi obtained after simplifying treatment is polygon Shape；
Step S4：Circulation is carried out in return to step S1, until initial Voronoi diagram inlays all Voronoi in gauze network Polygon is disposed, and is simplified after treatment and simply inlays gauze network.
Further description is, as follows to simply inlaying the optimum choice process of all nodes of gauze network in step 4：
Step 4-1：Determine the maximum number that each node is had by effective coverage polygon in network, and acquisition is accordingly built Build the total imaging region scope on these monolithics DOM of thing；
Step 4-2：Whether decision node P returns more than two video imaging regions to have, and step 4-3 is carried out in this way, otherwise Carry out step 4-8；
Step 4-3：Centered on node P, search radius are the owned building B in the range of 50m1,B2,…,Bn；
Step 4-4：For each building, its imaging region on every DOM is calculated, and seek union as the building The overall imaging region S of thing1,S2,…,Sn；
Step 4-5：To imaging region S1,S2,…,SnCarry out asking friendship to judge, then merge if occuring simultaneously, while writing down Si With BiCorresponding relation, until remaining imaging region S1,S2,…,SmAny of (m≤n) is neither intersected, and wherein i=1~ n；
Step 4-6：Whether decision node P is in imaging region S one by one1,S2,…,Sm(m≤n) is internal, if being write down if The numbering B in the housei, while finding BiS after corresponding mergingiIf do not existed, node P enters step 4- without optimization 8；
Step 4-7：Obtain comprising imaging region SiMinimum outsourcing rectangle, then find out minimum apart from node P distances Used as position after the optimization of node P, all of node P is updated to a little one rectangle end points Q in then simply inlaying gauze network Q；
Step 4-8：Whether decision node P is last node, and if it is optimization is finished, if not then carrying out down One detection of node is updated, and return to step 4-2 circulations are carried out, until all nodes all optimize finishing, obtain all nodes Suboptimum not in building imaging region inlays gauze network.
Further description is that all lines of inlaying for inlaying gauze network in step 5 to the suboptimum are in optimized selection step Suddenly it is：
Step 5-1：Judge that suboptimum is inlayed in gauze network and inlay whether line L is that two effective polygons of image have, if It is then to enter step 5-2, if only one polygon possesses, this inlays line without optimization, is directly entered step 5-9；
Step 5-2：According to line L is inlayed the image sequence number Pic for inlaying line both sides is found with the corresponding relation of imagei、 Picj, and according to the corresponding relation between building and image, picked out the building on two images is appeared in simultaneously It is designated as B1,B2,…,Bn；
Step 5-3：For each building, it is calculated in PiciAnd PicjImaging region on two orthographies, and ask Union, as region, is designated as S as the overall city of the building1,S2,…,Sn；
Step 5-4：Judge each building imaging region SiWhether with inlay line L and intersect, intersecting polygon is left It is designated as：S1,S2,…,Sm(m≤n), this inlays line if all non-intersect need not automatically select, into step 5-9；
Step 5-5：To S1,S2,…,SmAsked friendship to judge two-by-two, then merged if occuring simultaneously, until being left S1, S2,…,SlBoth are non-intersect each other for (l≤m) any；
Step 5-6：By polygon S1,S2,…,Sl(l≤m) according to distance inlay the distance of line L starting points LB according to from it is small to Big sequence obtains sequences of polygons SS1,SS2,…,SSl(l≤m)；
Step 5-7：From first polygon SS1Rise, obtain and inlay line L and polygon SSiThe intersection point of (i≤l), wherein the One intersection point is INSf, last intersection point is INSe, then line L from point INS is inlayedfTo INSeThis section is along polygon SSi(i≤ L) profile bending is advanced, and selects the most short path of total distance to be inlayed in line L from point INS instead of originalfTo INSePart, directly Line L is inlayed to this bypass all of polygon SSi(i≤l)；
Step 5-8：Thread path is inlayed according to what step 5-7 was obtained, is updated to inlaying line L, after being updated Inlay line L'；
Step 5-9：Return to step 5-1 circulations are carried out, and are inlayed line optimization and are finished until all, entirely surveyed area it is automatic around Cross the overall optimal seam-line network of building imaging region.
This method first generates building simple model based on building roof vector, and is obtained using building simple model Its imaging region on raw video and correspondence DOM；Then using survey all image plate nadir point positions in area automatically generate survey area Initial Voronoi diagram inlays gauze network, and gauze network is simply inlayed in simplification；The building being calculated is being utilized in monolithic Imaging region on DOM, is in optimized selection to simply inlaying all of node of gauze network；Finally using the building being calculated Imaging region of the thing on monolithic DOM, all lines of inlaying that gauze network is inlayed to suboptimum are in optimized selection, and quickly obtain whole Survey area and get around the overall optimal seam-line network of building imaging region, solving survey area and inlay gauze network cannot automatically bypass building Imaging region and cause texture misplace and the existing problem for inlaying gauze network automatic selecting method inefficiency.
Remarkable result of the invention is：
1st, initially gauze network is inlayed using surveying all plate nadir points in area and automatically generate, then based on building roof vector to inlaying Gauze network is in optimized selection, and that is entirely surveyed that area bypasses building imaging region inlays gauze network, breaches tradition advanced Individual is just penetrating correction to row, and recycling is just penetrated correction result and inlay the pattern of line options；
2nd, inlay line optimum choice based on building roof vector, more existing big multi-method (such as ant group algorithm, Dijkstra's algorithm etc.), during inventive algorithm only a little, line, polygonal calculating, with amount of calculation is small, the degree of accuracy more High, in hgher efficiency the advantages of, quickly can entirely be surveyed that area bypasses building imaging region inlays gauze network, solves survey Area inlay gauze network cannot automatically bypass building imaging region cause texture misplace and existing method to automatically select efficiency low Problem；
3rd, building imaging region had both been bypassed, has been efficiently solved to inlay line and cause texture mistake etc. through building and is asked Topic, also remains the advantage that Voronoi diagram inlays gauze network very well, that is, DOM images each pixel after inlaying is projected in theory Deformation is minimum, effectively reduces the big influence of image edge distortion of projection that central projection causes.
Brief description of the drawings
Fig. 1 is flow chart of the method for the present invention；
Fig. 2 is the schematic diagram that vector combination DEM in building roof obtains simple model；
Fig. 3 is the schematic diagram of building roof vector and the simple model for obtaining；
Fig. 4 is the imaging region flow chart for obtaining building on monolithic DOM；
Fig. 5 is the method schematic of the imaging region for obtaining building on monolithic DOM；
Fig. 6 is the principle that its coordinate method on correspondence monolithic DOM of DEM iteratives is based on by point coordinates on raw video Figure；
The initial Voronoi diagram in Tu7Shi Ce areas inlays the schematic diagram of gauze network and thin portion；
Fig. 8 is to inlay local contrast figure after line network reduction before processing；
Fig. 9 is the schematic diagram that same building thing is imaged overall area on different monolithic DOM；
Figure 10 is before and after network node optimization is selected and building imaging region relation comparison diagram；
Figure 11 is the schematic diagram for inlaying line optimum choice；
Figure 12 is to inlay comparison diagram before and after line optimization；
Figure 13 is to inlay the overall optimal seam-line network and thin portion schematic diagram obtained after gauze network optimum choice；
Figure 14 is that DOM inlays result local contrast figure before and after inlaying the line network optimization.
Specific embodiment
Specific embodiment of the invention and operation principle are described in further detail below in conjunction with the accompanying drawings.
In this example, for 4, certain city air strips, totally 32 resolution ratio are 0.1m, endlap to the aviation image sample for using Degree about 60%, sidelapping degree is about the aviation image that 45%, width size is 8230*6168.
As shown in figure 1, a kind of orthography based on building roof vector inlays gauze network automatic selecting method, specifically Step is as follows：
Step 1：Building roof vector combination digital complex demodulation automatically generates building simple model, specially：
In the photogrammetric stereoplotting stage, then carry out stereoplotting by stereogram structure three-dimensional model to obtain Roof line of vector, they not only have two dimension XY coordinates, while also have elevation coordinate Z, be the three-dimensional description to roof. It is such as attached if roof vector polygon projected on survey area DEM vertically downward, you can obtain the easy threedimensional model in the house Shown in Fig. 2, ABCD be building roof vector polygon, point A, B, C, D are projected to respectively respectively obtained on DEM point a, b, c, D, then ABCD-abcd be the easy threedimensional model of the building (Digital Building Model, abbreviation DBM).It is three-dimensional The storage model of model has many kinds, such as parameter model, CAD model, CSG models and polyhedral structure model.Wherein polyhedron Structural model is the topological relation between the dough sheet using single or multiple connections builds the geometry on whole roof, so Afterwards complete three-dimensional model building is obtained in conjunction with vertical metope.These constitute the dough sheet of building then generally with tri patch Form constitute the triangulation network and represent whole building thing, therefore, the simple threedimensional model of the building that will be obtained in the present invention with The mode of polyhedral structure model is stored.
Step 2：Imaging region of the building on monolithic DOM is calculated using building simple model；
Do not include building information due to just penetrating the DEM models used in correcting, therefore by the image after correction Still there is building object location to incline, the phenomenon such as building wall is visible.DBM is the closure geometry knot being made up of tri patch Structure, according to topological relation principle of invariance, still connects after the triangle projection being connected, therefore owning whole DBM Tri patch projects to arbitrary plane, then seeks union, obtain be all that one complete, closing, the polygon without hole.Therefore it is Imaging region of the building on DOM accurately is obtained, one kind is proposed in this example and is combined to obtain building by DBM and DEM The method of imaging region vector polygon on orthography.DBM is projected to raw video by the method first, obtains building Imaging region polygon S on raw video, is then based on DEM iteratives algorithm and polygon S is mapped into orthography On, you can imaging region polygon S ' of the building on corresponding monolithic DOM is obtained, its flow is as shown in Figure 4.
Referring to accompanying drawing 4, the obtaining step of building imaging region on monolithic DOM is as follows：
Initially enter step 2-1：By on all tri patch transmission projections of building simple model to raw video, ask Obtain its all tri patch on raw video；
Step 2-2：Union is asked to all tri patch on raw video, imaging of the building on raw video is obtained Region, concrete principle such as Fig. 5, point O are photo centres, and cuboid ABCD-abcd is assumed to be simple construction thing model, is owned Tri patch projected on raw video according to formula (1), then ask union to obtain the building on raw video Imaging polygon A'B'C'cba (polygon end points is image space coordinate representation).
In formula：(X, Y, Z) is the object coordinates of point, (x, y) image space coordinate, (X in photo coordinate system for pointS,YS, ZS) be photo centre object coordinates, f is camera focus, a1,a2,a3,b1,b2,b3,c1,c2,c3It is photogrammetric spin matrix 9 rotation parameters.
Step 2-3：Solved based on DEM iterative approach, building is obtained in monolithic DOM by the imaging region on raw video On imaging region.
As shown in Figure 5, by each angle point of polygon A'B'C'cba on raw video, according to formula (2) and based on DEM Iterative approach is solved and obtains polygon corresponding region on orthography, i.e. building imaging region polygon on DOM A " B " C " D " c " b " a " (the polygon end points is represented by object coordinates).
In formula：(X, Y, Z) is the object coordinates put, and (x, y) is image space coordinate, (X in the photo coordinate system putS,YS, ZS) be photo centre object coordinates, f is camera focus, a1,a2,a3,b1,b2,b3,c1,c2,c3It is photogrammetric spin matrix 9 rotation parameters.
Object space polygon A on DOM is being solved by image space polygon A'B'C'cba " B " C " D " c " b " a " when, it is two-dimensional coordinate The ill-conditioned process of three-dimensional coordinate is solved, it is necessary to carry out the interative computation based on DEM.As shown in Figure 6, point O is photo centre, ground Millet cake M coincides with point m with the picture point of room angle point B, and the coordinate of ground point M is now solved by picpointed coordinate m.First, will to survey area minimum Elevation Zmin can obtain point M as initial elevation in substitution formula (2)0Coordinate；Secondly, according to M0Coordinate it is interior on DEM Insert elevational point T1(coordinate and M0Equally), then by T1Height value substitute into formula (2), obtain M1Coordinate；Then according to M1 The coordinate Interpolation spots T on DEM again2(coordinate and M1Equally), then substitute into formula (2) and obtain point M2.Iterative cycles, directly Meet qualifications with the difference of upper once interpolation height value to the height value currently in DEM interpolations, be then currently calculated Point MnThe point M of as required solution.
Into step 3：Using survey all image plate nadir point positions in area automatically generate survey the initial Voronoi diagram in area inlay gauze Network, and it being carried out simplify obtain after treatment and relatively simple inlay gauze network；
Wherein, initial Voronoi diagram inlays gauze network obtaining step and is：
Step 3-1：The principal point that all images in area will be surveyed projects to ground by formula (2), obtains its corresponding plate nadir point；
Step 3-2：It is scatterplot collection to survey all plate nadir points in area, Delaunay Triangulation is carried out using pointwise interpolation method；
Step 3-3：According to triangulation result, it is many that the corresponding Thiessen polygon of every image is then referred to as image effective coverage Side shape；Every topological relation inlayed between line and two side images is obtained simultaneously, the initial Voronoi diagram in as this survey area is inlayed Gauze network, as shown in Figure 7.
It is pure because traditional Voronoi diagram is the division to discrete point set " sphere of influence " in actual mechanical process Essence occurs very short side based on geometry site generation in the initial Voronoi diagram network of generation.Such as accompanying drawing 8 (a) institute Show, the two-end-point on these sides increases the difficulty to inlaying the line network optimization if in the region for fully falling in building imaging Degree.And for the Voronoi diagram that air strips rule, the rational sequence aviation image of degree of overlapping are generated, these short sides can regularly go out The boundary of Xian Yu centers image and non-surrounding image (referred to four images up and down).Therefore these short sides can be deleted Inlaying gauze network to initial Voronoi diagram carries out certain simplification, not only can largely ensure Voronoi diagram characteristic but also Can simplify and inlay the follow-up optimization difficulty of line.
Therefore, in the present embodiment, the initial Voronoi diagram that will be obtained inlays gauze network and follows the steps below simplified place Reason：
Step S1：Calculate the length that initial Voronoi diagram inlays the polygonal four edges most long of certain Voronoi in gauze network Degree average value；
Step S2：Obtain short side of the length less than average value a quarter in above-mentioned Voronoi polygons；
Step S3：The short side is substituted with the geometry midpoint of two end points of short side, the Voronoi obtained after simplifying treatment is polygon Shape；
Step S4：Circulation is carried out in return to step S1, until initial Voronoi diagram inlays all Voronoi in gauze network Polygon is disposed, and is simplified after treatment and simply inlays gauze network.
Voronoi diagram after simplification is inlayed shown in gauze network such as figure accompanying drawing 8 (b).As can be seen from the figure to the letter of short side Change operation, the layout and feature for not inlaying gauze network to surveying area have a very big change, the corresponding effective polygon of every width image Also it is basically identical；All become for quadrangle by the corresponding effective polygon of the every width image of simplification, each section in addition to fringe node Point at most returns four effective polygons to have, and inlays line for every and at most returns two effective polygons to have.
Into step 4：Imaging region of the building obtained using step 2 on monolithic DOM, to simply inlaying gauze network All nodes be in optimized selection, obtain suboptimum and inlay gauze network；
Step 4-1：Imaging overall area of the solitary building in object space is obtained, it is determined that by the video imaging region of total node Maximum number；
Because photo centre is different, a solitary building is also differed in the imaging region of every orthography, such as the institute of accompanying drawing 9 Show, O1,O2It is the photo centre of adjacent two images, prism ABCD-abcd is assumed to be building, then the building is just being penetrated Image O1Upper imaging region is polygon A'B'C'cba, and in orthography O2Upper imaging region is polygon B " C " D " dab.Then The imaging overall area of the building is polygon A'B'EB " C " D " da.From step 3-2, each node at most returns 4 images Effective polygon have, i.e., the maximum number in the described video imaging region by total node is 4.
Therefore to ensure node not on any image in four images in the imaging region of building, it is necessary to ensure Node is not on four orthographies in the union polygon of imaging region.
Step 4-2：Whether decision node P returns more than two video imaging regions to have, and step 4-3 is carried out in this way, otherwise Carry out step 4-8；
Step 4-3：Centered on node P, search radius are the owned building B in the range of 50m1,B2,…,Bn；
Step 4-4：For each building, its imaging region on every DOM is calculated, and seek union as the building The overall imaging region S of thing1,S2,…,Sn；
Step 4-5：To imaging region S1,S2,…,SnCarry out asking friendship to judge, then merge if occuring simultaneously, while writing down Si With BiCorresponding relation, until remaining imaging region S1,S2,…,SmAny of (m≤n) is neither intersected, and wherein i=1~ n；
Step 4-6：Whether decision node P is in imaging region S one by one1,S2,…,Sm(m≤n) is internal, if being write down if The numbering B in the housei, while finding BiS after corresponding mergingiIf do not existed, node P enters step 4- without optimization 8；
Step 4-7：Obtain comprising imaging region SiMinimum outsourcing rectangle, then find out minimum apart from node P distances Used as position after the optimization of node P, all of node P is updated to a little one rectangle end points Q in then simply inlaying gauze network Q；
Step 4-8：Whether decision node P is last node, and if it is optimization is finished, if not then carrying out down One detection of node is updated, and return to step 4-2 circulations are carried out, until all nodes all optimize finishing, obtain all nodes Suboptimum not in building imaging region inlays gauze network.
By after above-mentioned steps, you can obtain suboptimum of all nodes not in building imaging region and inlay gauze Network, before and after node optimization with the position relationship of building imaging region as shown in Figure 10.
Node of the gauze network decline in building imaging region will have been inlayed in step 4 to be all imaged out of building Outside polygon, it is ensured that every end points for inlaying line is beyond building imaging region.However, every is inlayed line by two shadows As effective polygon is total, therefore the optimization of building is automatically bypassed, it is necessary to consider that two images have simultaneously for inlaying line Building imaging contexts and optimization path choose.
Into step 5：Imaging region of the building obtained using step 2 on monolithic DOM, gauze network is inlayed to suboptimum All lines of inlaying be in optimized selection, entirely surveyed the overall optimal seam-line network that area gets around building imaging region.
Step 5-1：Judge that suboptimum is inlayed in gauze network and inlay whether line L is that two effective polygons of image have, if It is then to enter step 5-2, if only one polygon possesses, this inlays line without optimization, is directly entered step 5-9；
Step 5-2：According to line L is inlayed the image sequence number Pic for inlaying line both sides is found with the corresponding relation of imagei、 Picj, and according to the corresponding relation between building and image, picked out the building on two images is appeared in simultaneously It is designated as B1,B2,…,Bn；
Step 5-3：For each building, it is calculated in PiciAnd PicjImaging region on two orthographies, and ask Union, as region, is designated as S as the overall city of the building1,S2,…,Sn；
Step 5-4：Judge each building imaging region SiWhether with inlay line L and intersect, intersecting polygon is left It is designated as：S1,S2,…,Sm(m≤n), this inlays line if all non-intersect need not automatically select, into step 5-9；
Step 5-5：To S1,S2,…,SmAsked friendship to judge two-by-two, then merged if occuring simultaneously, until being left S1, S2,…,SlBoth are non-intersect each other for (l≤m) any；
Step 5-6：By polygon S1,S2,…,Sl(l≤m) according to distance inlay the distance of line L starting points LB according to from it is small to Big sequence obtains sequences of polygons SS1,SS2,…,SSl(l≤m)；
Step 5-7：From first polygon SS1Rise, obtain and inlay line L and polygon SSiThe intersection point of (i≤l), wherein the One intersection point is INSf, last intersection point is INSe, then line L from point INS is inlayedfTo INSeThis section is along polygon SSi(i≤ L) profile bending is advanced, and selects the most short path of total distance to be inlayed in line L from point INS instead of originalfTo INSePart, directly Line L is inlayed to this bypass all of polygon SSi(i≤l)；
As shown in Figure 11, polygon ABLBLE is an effective coverage polygon for image to the detailed process of step 5-7, Wherein LB and LE are respectively the beginning and end for inlaying line L, SiAnd SjIt is two imaging region polygons of building, point INSfi With point INSeiRespectively inlay line L and polygon SiFirst intersection point and last intersection point, point INSfjWith point INSejRespectively To inlay line L and polygon SjFirst intersection point and last intersection point.So inlay line L selection bypass polygon SiWhen There are two lines available, first is broken line LB-INS for left side routefi-Pi0-Pi7-INSei, Article 2 is the right route That is broken line LB-INSfi-Pi1-…-Pi6-INSei, now select the broken line of shortest path to bypass polygon Si, therefore selection is left Wing footpath.Similarly, inlay line L selection bypass polygon SjWhen, then selection the right route is broken line INSfj-Pj1-…-Pj6- INSej.If inlaying line L just through SiAnd Sj, then inlay the final optimum results of line L as broken line LB-INSfi-Pi0- Pi7-INSei-INSfj-Pj1-…-Pj6-INSej- LE, as shown in thick line in Figure 11.
Accompanying drawing 12 then give wall scroll inlay line optimization before and after comparison diagram, from Figure 12 (a) it can be seen that optimization before this Inlay line and have passed through 7 solitary buildings, comparison diagram 12 (b) understands to inlay line and successfully bypassed these by this after above-mentioned optimization to build The imaging region of thing is built, and be have selected and optimal bypassed path.
Step 5-8：Thread path is inlayed according to what step 5-7 was obtained, is updated to inlaying line L, after being updated Inlay line L'；
Step 5-9：Return to step 5-1 circulations are carried out, and are inlayed line optimization and are finished until all, entirely surveyed area it is automatic around The overall optimal seam-line network of building imaging region is crossed, the overall optimal seam-line network is as shown in figure 13.
Figure 14 is then given to be automatically selected using this method and inlays DOM before and after gauze network and inlay the local comparison diagram of result, Figure (a) does not bypass building imaging region to inlay line, line both sides is inlayed after inlaying and texture dislocation occurs, and figure (b) is by this Method automatically selects DOM after inlaying gauze network and inlays effect, it can be seen that inlays line and has automatically bypassed building imaging region, has The problems such as effect solves texture and misplaces.
This method first generates building simple model based on building roof vector, and is obtained using building simple model Its imaging region on raw video and correspondence DOM；Then using survey all image plate nadir point positions in area automatically generate survey area Voronoi diagram inlays gauze road, and gauze network is simply inlayed in simplification；The building that recycling is calculated is in monolithic DOM On imaging region, inlay all of node of gauze network and be in optimized selection to simple；Finally using the building being calculated Imaging region on monolithic DOM, all lines of inlaying that gauze network is inlayed to suboptimum are in optimized selection, and are quickly entirely surveyed Area gets around the overall optimal seam-line network of building imaging region.
This programme breaches tradition first to carry out individual and is just penetrating corrections, and recycling is just penetrated and corrects result and carry out inlaying line options Pattern, more existing big multi-method (such as ant group algorithm, dijkstra's algorithm etc.), during only a little, vector, polygonal meter Calculate, have the advantages that amount of calculation is small, the degree of accuracy is high, in hgher efficiency, solve survey area inlay gauze network cannot bypass building into As region causes texture dislocation and existing method to automatically generate the low problem of efficiency.

Claims (6)

1. the orthography based on building roof vector inlays gauze network automatic selecting method, it is characterised in that according to following step Suddenly carry out：
Step 1：Building roof vector combination digital complex demodulation automatically generates building simple model；
Step 2：Imaging region of the building on monolithic DOM is calculated using building simple model；
Step 3：Gauze network is inlayed using surveying all image plate nadir point positions in area and automatically generating the survey initial Voronoi diagram in area, it is and right It carries out simply being inlayed gauze network after simplifying treatment；
Step 4：Imaging region of the building obtained using step 2 on monolithic DOM, all sections to simply inlaying gauze network Point is in optimized selection, and obtains suboptimum and inlays gauze network；
Step 5：Imaging region of the building obtained using step 2 on monolithic DOM, all edges of gauze network are inlayed to suboptimum Rule is in optimized selection, and is entirely surveyed the overall optimal seam-line network that area gets around building imaging region.
2. the orthography based on building roof vector according to claim 1 inlays gauze network automatic selecting method, It is characterized in that：The calculation procedure of imaging region of the building described in step 2 on monolithic DOM is as follows：
Step 2-1：By on all tri patch transmission projections of building simple model to raw video, raw video three is tried to achieve Edged surface piece；
Step 2-2：Union is asked to raw video tri patch, imaging region of the building on raw video is obtained；
Step 2-3：Solved based on DEM iterative approach, building is obtained in corresponding monolithic DOM by the imaging region on raw video On imaging region.
3. the orthography based on building roof vector according to claim 1 inlays gauze network automatic selecting method, It is characterized in that：The obtaining step that initial Voronoi diagram inlays gauze network described in step 3 is as follows：
Step 3-1：The principal point that all images in area will be surveyed projects to ground, obtains its corresponding plate nadir point；
Step 3-2：It is scatterplot collection to survey all plate nadir points in area, Delaunay Triangulation is carried out using pointwise interpolation method；
Step 3-3：According to triangulation result, generate initial Voronoi diagram and inlay gauze network.
4. the orthography based on building roof vector according to claim 1 inlays gauze network automatic selecting method, It is characterized in that：The simplified process step that initial Voronoi diagram inlays gauze network in step 3 is：
Step S1：Calculate initial Voronoi diagram inlay the polygonal four edges most long of certain Voronoi in gauze network length put down Average；
Step S2：Obtain short side of the length less than average value a quarter in above-mentioned Voronoi polygons；
Step S3：The short side is substituted with the geometry midpoint of two end points of short side, the Voronoi polygons after simplifying treatment are obtained；
Step S4：Circulation is carried out in return to step S1, until all Voronoi are polygon during initial Voronoi diagram inlays gauze network Shape is disposed, and is simplified after treatment and simply inlays gauze network.
5. the orthography based on building roof vector according to claim 1 inlays gauze network automatic selecting method, It is characterized in that：It is as follows to simply inlaying the optimum choice process of all nodes of gauze network in step 4：
Step 4-1：Determine the maximum number that each node is had by effective coverage polygon in network, and obtain corresponding building Total imaging region scope on these monolithics DOM；
Step 4-2：Whether decision node P returns more than two video imaging regions to have, and step 4-3 is carried out in this way, otherwise carries out Step 4-8；
Step 4-3：Centered on node P, search radius are the owned building B in the range of 50m1,B2,…,Bn；
Step 4-4：For each building, its imaging region on every DOM is calculated, and seek union as the building Overall imaging region S1,S2,…,Sn；
Step 4-5：To imaging region S1,S2,…,SnCarry out asking friendship to judge, then merge if occuring simultaneously, while writing down SiWith Bi's Corresponding relation, until remaining imaging region S1,S2,…,SmAny of (m≤n) is neither intersected, wherein i=1~n；
Step 4-6：Whether decision node P is in imaging region S one by one1,S2,…,Sm(m≤n) is internal, if writing down the room if The numbering B in roomi, while finding BiS after corresponding mergingiIf do not existed, node P enters step 4-8 without optimization；
Step 4-7：Obtain comprising imaging region SiMinimum outsourcing rectangle, then find out apart from a minimum square of node P distances Used as position after the optimization of node P, all of node P is updated to point Q to shape end points Q in then simply inlaying gauze network；
Step 4-8：Whether decision node P is last node, and if it is optimization is finished, next if not then carrying out The detection of node is updated, and return to step 4-2 circulations are carried out, until all nodes all optimize finishing, obtain all nodes not Suboptimum in building imaging region inlays gauze network.
6. the orthography based on building roof vector according to claim 1 inlays gauze network automatic selecting method, It is characterized in that：The suboptimum inlayed in step 5 gauze network it is all inlay line and be in optimized selection step be：
Step 5-1：Judge that suboptimum is inlayed in gauze network and inlay whether line L is that two effective polygons of image have, if then Into step 5-2, if only one polygon possesses, this inlays line without optimization, is directly entered step 5-9；
Step 5-2：According to line L is inlayed the image sequence number Pic for inlaying line both sides is found with the corresponding relation of imagei、Picj, and According to the corresponding relation between building and image, the building for appearing on two images simultaneously is picked out and is designated as B1, B2,…,Bn；
Step 5-3：For each building, it is calculated in PiciAnd PicjImaging region on two orthographies, and seek union As the overall city of the building as region, S is designated as1,S2,…,Sn；
Step 5-4：Judge each building imaging region SiWhether with inlay line L and intersect, intersecting polygon is left and is designated as： S1,S2,…,Sm(m≤n), this inlays line if all non-intersect need not automatically select, into step 5-9；
Step 5-5：To S1,S2,…,SmAsked friendship to judge two-by-two, then merged if occuring simultaneously, until being left S1,S2,…,Sl Both are non-intersect each other for (l≤m) any；
Step 5-6：By polygon S1,S2,…,Sl(l≤m) inlays the distance of line L starting points LB according to arranging from small to large according to distance Sequence obtains sequences of polygons SS1,SS2,…,SSl(l≤m)；
Step 5-7：From first polygon SS1Rise, obtain and inlay line L and polygon SSiThe intersection point of (i≤l), wherein first Intersection point is INSf, last intersection point is INSe, then line L from point INS is inlayedfTo INSeThis section is along polygon SSi(i≤l's) Profile bending is advanced, and selects the most short path of total distance to be inlayed in line L from point INS instead of originalfTo INSePart, until this Bar inlays line L and bypasses all of polygon SSi(i≤l)；
Step 5-8：Thread path is inlayed according to what step 5-7 was obtained, is updated to inlaying line L, inlaying after being updated Line L'；
Step 5-9：Return to step 5-1 circulations are carried out, and are inlayed line optimization and are finished until all, are entirely surveyed area and automatically bypassed and built Build the overall optimal seam-line network of thing imaging region.
CN201710059088.8A 2017-01-24 2017-01-24 Orthography based on building roof vector inlays gauze network automatic selecting method Active CN106846251B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710059088.8A CN106846251B (en) 2017-01-24 2017-01-24 Orthography based on building roof vector inlays gauze network automatic selecting method

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710059088.8A CN106846251B (en) 2017-01-24 2017-01-24 Orthography based on building roof vector inlays gauze network automatic selecting method

Publications (2)

Publication Number Publication Date
CN106846251A true CN106846251A (en) 2017-06-13
CN106846251B CN106846251B (en) 2018-02-06

Family

ID=59121457

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710059088.8A Active CN106846251B (en) 2017-01-24 2017-01-24 Orthography based on building roof vector inlays gauze network automatic selecting method

Country Status (1)

Country Link
CN (1) CN106846251B (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108257130A (en) * 2018-02-08 2018-07-06 重庆市地理信息中心 A kind of aviation orthography panorama sketch garland region rapid detection method
CN109063711A (en) * 2018-07-06 2018-12-21 航天星图科技（北京）有限公司 A kind of satellite image based on LLTS frame just penetrates correct algorithm
CN111311622A (en) * 2020-01-30 2020-06-19 桂林理工大学 Ant colony algorithm-based optimization selection method for multiple image mosaic wire nets
CN112465986A (en) * 2020-11-27 2021-03-09 航天恒星科技有限公司 Method and device for inlaying satellite remote sensing image
CN117455815A (en) * 2023-10-18 2024-01-26 二十一世纪空间技术应用股份有限公司 Method and related equipment for correcting top-bottom offset of flat-top building based on satellite image
CN118135149A (en) * 2024-05-10 2024-06-04 北京数慧时空信息技术有限公司 Method for generating remote sensing orthographic image embedded line

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103871072A (en) * 2014-04-04 2014-06-18 武汉大学 Method for automatic extraction of orthoimage embedding line based on projection digital elevation models
CN106339985A (en) * 2016-08-29 2017-01-18 中国科学院地理科学与资源研究所 Method for aviation image mosaic by selecting mosaic lines from vector building data

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103871072A (en) * 2014-04-04 2014-06-18 武汉大学 Method for automatic extraction of orthoimage embedding line based on projection digital elevation models
CN106339985A (en) * 2016-08-29 2017-01-18 中国科学院地理科学与资源研究所 Method for aviation image mosaic by selecting mosaic lines from vector building data

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
WAN Y等: "《Automatic determination of seamlines for aerial image mosaicking based on vector roads alone》", 《ISPRS JOURNAL OF PHOTOGRAMMETRY AND REMOTE SENSING》 *
周清华等: "《遥感图像镶嵌接缝线自动生成方法综述》", 《国土资源遥感》 *
潘俊等: "《接缝线网络的自动生成及优化方法》", 《测绘学报》 *

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108257130A (en) * 2018-02-08 2018-07-06 重庆市地理信息中心 A kind of aviation orthography panorama sketch garland region rapid detection method
CN109063711A (en) * 2018-07-06 2018-12-21 航天星图科技（北京）有限公司 A kind of satellite image based on LLTS frame just penetrates correct algorithm
CN109063711B (en) * 2018-07-06 2021-10-29 中科星图股份有限公司 Satellite image orthorectification algorithm based on LLTS framework
CN111311622A (en) * 2020-01-30 2020-06-19 桂林理工大学 Ant colony algorithm-based optimization selection method for multiple image mosaic wire nets
CN112465986A (en) * 2020-11-27 2021-03-09 航天恒星科技有限公司 Method and device for inlaying satellite remote sensing image
CN117455815A (en) * 2023-10-18 2024-01-26 二十一世纪空间技术应用股份有限公司 Method and related equipment for correcting top-bottom offset of flat-top building based on satellite image
CN117455815B (en) * 2023-10-18 2024-07-09 二十一世纪空间技术应用股份有限公司 Method and related equipment for correcting top-bottom offset of flat-top building based on satellite image
CN118135149A (en) * 2024-05-10 2024-06-04 北京数慧时空信息技术有限公司 Method for generating remote sensing orthographic image embedded line

Also Published As

Publication number Publication date
CN106846251B (en) 2018-02-06

Similar Documents

Publication Publication Date Title
CN106846251B (en) Orthography based on building roof vector inlays gauze network automatic selecting method
US20230074265A1 (en) Virtual scenario generation method and apparatus, computer device and storage medium
CN107833273B (en) Oblique photography three-dimensional model objectification application method based on three-dimensional simulation model
CN112465976B (en) Storage yard three-dimensional map establishing method, inventory management method, equipment and medium
CN105976426B (en) A kind of quick three-dimensional atural object model building method
CN111737790B (en) Method and equipment for constructing simulated city model
CN105069843A (en) Rapid extraction method for dense point cloud oriented toward city three-dimensional modeling
CN101154313A (en) Three-dimensional simulation digital information navigation channel system and its implementing method
CN104835202A (en) Quick three-dimensional virtual scene constructing method
CN101154295A (en) Three-dimensional simulation electronic chart of navigation channel
CN103954970B (en) A kind of topographic(al) feature acquisition method
CN111199066A (en) Construction site virtual construction restoration method based on BIM + GIS
CN108921943A (en) A kind of road threedimensional model modeling method based on lane grade high-precision map
CN114283070B (en) Method for manufacturing terrain section by fusing unmanned aerial vehicle image and laser point cloud
CN109903382A (en) The fusion method and device of point cloud data
CN104751479A (en) Building extraction method and device based on TIN data
CN106875330B (en) Method for rotating plane model into spherical model
CN105184854A (en) Quick modeling method for cloud achievement data of underground space scanning point
Borisov et al. An automated process of creating 3D city model for monitoring urban infrastructures
CN111982077B (en) Electronic map drawing method and system and electronic equipment
CN111241228B (en) Comprehensive drawing method based on vector data and graphic processing technology
Gu et al. Surveying and mapping of large-scale 3D digital topographic map based on oblique photography technology
Li et al. A review of true orthophoto rectification algorithms
CN116662473A (en) Space database establishment method, device, equipment and storage medium
CN102830960A (en) Three-dimensional simulated electronic drawing for sea channels

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
CP03 Change of name, title or address

Address after: 400020 Jiangbei District, Chongqing electric measuring Village No. 231

Patentee after: Chongqing geographic information and Remote Sensing Application Center (Chongqing surveying and mapping product quality inspection and testing center)

Address before: 401121 Chongqing new high tech area in northern Yubei District science and Technology Building 8 floor D block Neptune

Patentee before: Chongqing Geographical Information Center

CP03 Change of name, title or address