CN102222066B - Conflict shifting processing method for multi-source spatial data combination - Google Patents

Abstract

The invention relates to a conflict shifting processing method for multi-source spatial data combination. The method comprises the following steps of: (1) reading data, and constructing topological relationships among entities; (2) performing entity matching; (3) performing the regulation and combination of entity coordinates; (4) checking whether combination results comprise mismatching or not, correcting the mismatching and returning to the step (3) if the combination results comprise the mismatching, and executing the step (5) if the combination results do not comprise the mismatching; and (5) automatically detecting the existence or inexistence of a picture conflict, and if the picture conflict is automatically detected, performing conflict shifting by adopting a least square adjustment method. Compared with the prior art, the method has the advantages of saving manpower and material resources, eliminating differences among data sets, solving the spatial conflict, maintaining geometrical characteristics of the entities, and the like.

Description

The conflict shifting processing method that a kind of multi-source Spatial Data merges

Technical field

The present invention relates to the conflict shifting processing method that a kind of data merge, especially relate to the conflict shifting processing method that a kind of multi-source Spatial Data merges.

Background technology

Along with the development of Surveying and Mapping Industry and Geographic Information System (GIS), abundant GIS data resource has been accumulated by each department, and these data have various difference.Between data set, why producing difference is mainly due to the production division data from different, and, for self needs, there is different emphasis in each production division to data.As land administration department emphasizes the positional information in plot, thereby the database producing has higher spatial positional precision, and economic department is studied for space distribution information, more focuses on the attribute information of non-space, the database of its generation has abundant attribute data, but precision is poor.Wherein, of paramount importance a little species diversity is as follows:

1, the difference of data model.Existing GIS data model is divided into Vector Data Model and raster data model, even model of the same race, Method of Data Organization also can be different, as line entity may be used face list structure, also may use node segmental arc topological structure;

2, database structure difference, the field comprising as database is different, and same field data type is not equal;

3, data the level of detail is different, different mainly due to map scale, is subject to the impact of Map Generalization, for same entity information amount, varies in size.As, river under large scale than there is more multiple spot information under small scale, or due to data acquisition asynchronism(-nization), cause some entity on new figure not exist etc. on old figure;

4, the differences in spatial location of entity of the same name;

5, data precision difference, for example large scale topographical map is higher to the accuracy requirement of the volume coordinate position of atural object, and the cadastre of same ratio chi only requires that there is higher precision in ownership boundary line.

Due to above-mentioned kind of species diversity, realize data integrated and shared be difficulty comparatively, therefore most departments take the method for repeated acquisition, have wasted a large amount of manpower and materials, also caused the idle of a large amount of useful datas.Realize separate sources, have the integrated and shared important topic that has become present stage GIS development of the data of each species diversity.In order to utilize better these data, eliminate the difference of each data set to form the new data set that can meet certain requirement, therefore there is " spatial data merging " this concept.

Spatial data merging refers on the basis of spatial entities coupling of the same name, adjusts the geometric position of relevant atural object entity, realizes the integrated and information fusion of areal separate sources spatial data.In spatial data merging process, because the displacement of entity can cause and the conflicting of other entity.Be illustrated in fig. 1 shown below, in figure, entity comprises buildings entity and road entity, and in merging process, path layer carries out respectively the adjustment conversion of entity from buildings layer according to different maps, when two figure stacked with together with time, the too small Graphic Conflicts that causes of some buildings and road spacing.

Existing large quantity research attempts to solve the conflict displacement problem of figure.Lichtner (1976) has proposed the displacement formula of putting under perfect condition, and off-line entity is far away, and shift amount is less, and the method does not consider that a displacement is subject to the impact of many line entities, is only applicable to the displacement that single line entity causes.Many methods were all based on this thought afterwards, and expanded, for example, consider that road turnover is to the combined effect of every in displacement district, then according to corner, the floor area of building to whole building or turning point displacement.Mackness (1994) has set up and a kind ofly by target interphase interaction space exploration, has conflicted, thereby implements the attenuation type shifting function away from conflict center.Another kind method is that polar plot is converted into grid map, Li Zhilin (1996) has proposed the Map Generalization key element displacement model based on mathematical morphology, Fei Lifan (2004,2002) propose to adopt the grid one vector MIDAS mixed data structure owned building in district that will be shifted to classify, carry out the operations such as cartographic displacement and controlled-deformation.In additive method, Burghardt & Meier (1997) has proposed " snake " model, has considered that the external energy (conflict) of entity and internal energy (single order that entity is how much, second order reciprocal) are to reach energy minimum simultaneously.It is some triangles by whole Region Segmentation that Hojholt (1998) is used finite element method, to survey and manage conflict problem.Ai Tinghua (2004) sets up the subdivision structure that is similar to Voronoi figure in polygon cluster space, obtains " equidistant relation curve " for expressing the effect of a displacement power, by vector calculus, obtains direction of displacement and displacement.

About seldom mentioning the follow-up collision problem causing because merging in the research of spatial data, complete spatial data combination system should be able to detect conflict the displacement that automatically conflicts both at home and abroad.

The topological relation that must keep inter-entity while managing conflict displacement problem, the geometrical property of as far as possible preserving entity, can't bring new collision problem simultaneously.

Summary of the invention

The conflict shifting processing method that object of the present invention is exactly the difference that provides a kind of in order to overcome the defect that above-mentioned prior art exists and use manpower and material resources sparingly, eliminate each data set, solved space conflict, kept the multi-source Spatial Data of entity geometric properties to merge.

Object of the present invention can be achieved through the following technical solutions:

The conflict shifting processing method that multi-source Spatial Data merges, is characterized in that, comprises the following steps:

(1) reading out data, the topological relation of structure inter-entity;

(2) carry out entity coupling;

(3) carry out the adjustment merging of entity coordinate;

(4) whether check amalgamation result there is mistake coupling, and if yes, correct mistake coupling, and return to step (3), if NO, execution step (5);

(5) automatically whether detection there is Graphic Conflicts, if yes, adopts the least square adjustment displacement that manages conflict.

Entity coupling in described step (2) comprises the coupling of an entity, the coupling of the coupling of line entity, face entity.

Employing least square adjustment in described step (5) manage conflict displacement as follows:

If point coordinate is (x in former figure _i ⁰, y _i ⁰) (i=1,2 ..., n) as parameter approximate value, the position coordinates after point is mobile as adjusted value, have

$\left\{\begin{array}{c}{\hat{x}}_i=x_i^0+{\mathrm{\Δx}}_i\\ {\hat{y}}_i=y_i^0+{\mathrm{\Δy}}_i\end{array}\right.(i=\mathrm{1,2},\·\·\·,n)---\left(1\right)$

(Δ x in formula _i, Δ y _i) be coordinate translation amount, be parameter to be solved is adjusted value correction, the adjustment Equations of use comprises as follows:

1) coordinate amount of movement error equation

The point that forms entity should remain on original position as far as possible, and coordinate amount of movement will, close to 0, obtain following error equation

$\left\{\begin{array}{c}{v}_{x_i}={\mathrm{\Δx}}_i\\ v_{y_i}={\mathrm{\Δy}}_i\end{array}\right.---\left(2\right)$

All corresponding two the coordinate amount of movement error equations of each point, therefore the number of whole coordinate amount of movement error equation is total number of the point of twice;

2) relative shift error equation

In order to keep the geometry of entity inside, should make the relative shift of consecutive point on entity close to 0, can obtain following error equation

$\left\{\begin{array}{c}{v}_{x_{i,j+1}}={\mathrm{\Δx}}_i-{\mathrm{\Δx}}_{i+1}\\ v_{y_{i,i+1}}={\mathrm{\Δy}}_i-{\mathrm{\Δy}}_{i+1}\end{array}\right.---\left(3\right)$

3) shape error equation

Keep the original shape of entity when making Coordinate Adjusting, need carry out shape restriction to entity to be adjusted, if keep the turning value of entity frontier point, can obtain error equation and be

$v_{{\mathrm{\β}}_i}=a_{\mathrm{ik}}{\mathrm{\Δx}}_k+b_{\mathrm{ik}}{\mathrm{\Δy}}_k-(a_{\mathrm{ik}}-a_{\mathrm{ij}}){\mathrm{\Δx}}_i-(b_{\mathrm{ik}}-b_{\mathrm{ij}}){\mathrm{\Δy}}_i-a_{\mathrm{ij}}{\mathrm{\Δx}}_j-b_{\mathrm{ij}}{\mathrm{\Δy}}_j-l_{{\mathrm{\β}}_i}---\left(4\right)$

In formula $a_{\mathrm{ij}}=-\frac{\sin {\mathrm{\α}}_{\mathrm{ij}}}{s_{\mathrm{ij}}};$ $b_{\mathrm{ij}}=\frac{\cos {\mathrm{\α}}_{\mathrm{ij}}}{s_{\mathrm{ij}}};$ $a_{\mathrm{ik}}=-\frac{\sin {\mathrm{\α}}_{\mathrm{ik}}}{s_{\mathrm{ik}}};$ $b_{\mathrm{ik}}=\frac{\cos {\mathrm{\α}}_{\mathrm{ik}}}{s_{\mathrm{ik}}}$

$s_{\mathrm{ij}}=\sqrt{{({x_j}^0-{x_i}^0)}^2+{({y_j}^0-{y_i}^0)}^2};$ $s_{\mathrm{ik}}=\sqrt{{({x_k}^0-{x_i}^0)}^2+{({y_k}^0-{y_i}^0)}^2}$

$l_{{\mathrm{\β}}_i}={\mathrm{\β}}_i-{\mathrm{\β}}_i^0;$ ${\mathrm{\β}}_i^0=\arctan \frac{{y_k}^0-{y_i}^0}{{x_k}^0-{x_i}^0}-\arctan \frac{{y_j}^0-{y_i}^0}{{x_j}^0-{x_i}^0}$

Wherein, α _ik, α _ijbe respectively the position angle of ik, ij direction, by the approximation calculation of coordinate adjustment value, obtained, the former angle value that β is corner point, the number of shape error equation is the number of entity corner point;

4) conflict shift error equation

Graphic Conflicts can be divided into a little and conflicting with face with line, line with face, line with line, point with point, point,

Point is tried to achieve with the range formula of point

$v_{D(P_i,P_j)}+D(P_i,P_j)=\sqrt{{({\hat{x}}_j-{\hat{x}}_i)}^2+{({\hat{y}}_j-{\hat{y}}_i)}^2}---\left(5\right)$

Wherein, the error of distance observation, D (P _i, P _j) be a P _i(x _i, y _i) and P _j(x _j, y _j) distance, with it is the coordinate after its displacement;

Linearization obtains its error equation:

v _dist＝-a _ijΔx _i+b _ijΔy _i+a _ijΔx _j-b _ijΔy _j-l _dist (6)

A wherein _ij=cos α _ij, b _ij=-sin α _ij, α _ijapproximation calculation by coordinate adjustment value obtains, l _distvalue be divided into two kinds of situations: if the distance D of point and point before mobile _point-pointbe less than minimum threshold of distance mindist, l _dist=min dist-D _point-point, wherein if D _point-pointbe greater than mindist and be less than l of 1.5*mindist _dist=0;

Point with the range equation of line segment is:

$v_{D(P_i,L_{P_j{P}_k})}+D_{(P_i,L_{P_j{P}_k})}=\frac{|({\hat{y}}_j-{\hat{y}}_k){\hat{x}}_i+({\hat{x}}_k-{\hat{x}}_j){\hat{y}}_i+({\hat{x}}_j{\hat{y}}_k-{\hat{x}}_k{\hat{y}}_j)|}{\sqrt{{({\hat{y}}_k-{\hat{y}}_j)}^2+{({\hat{x}}_k-{\hat{x}}_j)}^2}}---\left(7\right)$

Wherein, with respectively apart from observed reading and error, with two end points of the point after adjustment and line segment;

Linearization obtains its error equation:

$v_{D(P_i,L_{P_j{P}_k})}=a_{\mathrm{ijk}}{\mathrm{\δx}}_i+b_{\mathrm{ijk}}{\mathrm{\δy}}_i+c_{\mathrm{ijk}}{\mathrm{\δx}}_j+d_{\mathrm{ijk}}{\mathrm{\δy}}_j+e_{\mathrm{ijk}}{\mathrm{\δx}}_k+f_{\mathrm{ijk}}{\mathrm{\δy}}_k-l_{D(P_i,P_j{P}_k)}---\left(8\right)$

Wherein, (x _i ⁰, y _i ⁰), (x _j ⁰, y _j ⁰) and (x _k ⁰, y _k ⁰) be the approximate value of coordinate, and

$a_{\mathrm{ijk}}=\frac{y_j^0-y_k^0}{{({(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2)}^{1/2}};$

$b_{\mathrm{ijk}}=\frac{x_k^0-x_j^0}{{({(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2)}^{1/2}};$

$c_{\mathrm{ijk}}=\frac{(y_k^0-y_i^0)[{(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2]+(x_k^0-x_j^0)[(y_j^0-y_k^0)x_i^0-(x_k^0-x_j^0)y_i^0+(x_j^0{y}_k^0-x_k^0{y}_j^0)]}{{({(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2)}^{3/2}};$

$d_{\mathrm{ijk}}=\frac{(x_i^0-x_k^0)[{(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2]+(y_k^0-y_j^0)[(y_j^0-y_k^0)x_i^0-(x_k^0-x_j^0)y_i^0+(x_j^0{y}_k^0-x_k^0{y}_j^0)]}{{({(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2)}^{3/2}};$

$e_{\mathrm{ijk}}=\frac{(y_i^0-y_j^0)[{(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2]+(x_j^0-x_k^0)[(y_j^0-y_k^0)x_i^0-(x_k^0-x_j^0)y_i^0+(x_j^0{y}_k^0-x_k^0{y}_j^0)]}{{({(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2)}^{3/2}};$

$f_{\mathrm{ijk}}=\frac{(x_j^0-x_i^0)[{(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2]+(y_j^0-y_k^0)[(y_j^0-y_k^0)x_i^0-(x_k^0-x_j^0)y_i^0+(x_j^0{y}_k^0-x_k^0{y}_j^0)]}{{({(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2)}^{3/2}};$

$l_{D(P_i,L_{P_j{P}_k})}=D_{(P_i,L_{P_j{P}_k})}-\frac{|(y_j^0-y_k^0)x_i^0+(x_k^0-x_j^0)y_i^0+(x_j^0{y}_k^0-x_k^0{y}_j^0)|}{\sqrt{{(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2}}.$

L in formula _distvalue with point with point error equation identical, if D _point-line< mindist, l _dist=min dist-D _point-lineif, mindist < D _point-line< 1.5*mindist, l _dist=0;

Point with the distance definition of face is: this,, to the minor increment forming in the distance of all polylines of this face, thus, can be similar to formula (8) and set up the distance error equation that point arrives face;

Line to the distance definition of line is: for line segment L _iand L _j, its distance is

$D_{(L_i,L_j)}=\underset{P_i\&Element;L_i}{\min }\left\{\underset{P_j\&Element;L_j}{\min }\right\{D(P_i,P_j)\left\}\right\}---\left(9\right)$

Wherein, P _iline segment L _ion any point, P _jline segment L _jon any point, D (P _i, P _j) be a P _iand P _jdistance, according to formula (5), calculate;

Line to the distance definition of face is: for line segment L _iwith polygon A _j, its distance is

$D_{(L_i,A_j)}=\underset{P_i\&Element;L_i}{\min }\left\{\underset{P_j\&Element;A_j}{\min }\right\{D(P_i,P_j)\left\}\right\}---\left(10\right)$

Wherein, P _iline segment L _ion any point, P _jpolygon A _jin any point, line is line segment to the minimum value forming in the distance of this polygonal polyline to the distance definition of face,

Line to the distance definition of face is: for line segment A _iwith polygon A _j, its distance is

$D_{(A_i,A_j)}=\underset{P_i\&Element;A_i}{\min }\left\{\underset{P_j\&Element;A_j}{\min }\right\{D(P_i,P_j)\left\}\right\}---\left(11\right)$

Wherein, P _ipolygon A _iin any point, P _jpolygon A _jin any point, the various equations of simultaneous, are indirect adjustment problem, its universal model is:

v＝Ax-l (12)

In formula, x is parameter vector, x=(Δ x ₁Δ y ₁Δ x _nΔ y _n) ^t; A is factor arrays; L is constant vector; 1 is constant vector; V is residual vector, by the principle of least square, can obtain normal equation and is

N _Ax-U＝0 (13)

In formula: N _a=A ^tpA; U=A ^tpl, P is weight matrix;

Separating normal equation can obtain:

$x=N_A^{-1}\&CenterDot;U---\left(14\right)$

Parameter x substitution error equation by trying to achieve, obtains residual error v and adjusted value coordinate adjustment value is the coordinate after a movement;

The inverse of weight matrix computing formula of coordinate adjustment value is

$Q_{\hat{x}\hat{x}}{\left(A^T\mathrm{PA}\right)}^{-1}=N_A^{-1}---\left(15\right)$

The factor arrays that wherein A is error equation, p is power battle array, the accuracy formula of adjusted value is

${\widehat{\mathrm{\&sigma;}}}_{{\hat{x}}_i}=\&PlusMinus;{\widehat{\mathrm{\&sigma;}}}_0\sqrt{Q_{{\hat{x}}_i{\hat{x}}_i}}---\left(16\right)$

Wherein for the pivot of matrix, error in weight unit

Compared with prior art, the present invention have use manpower and material resources sparingly, eliminate each data set difference, solved space conflict, kept entity geometric properties.

Accompanying drawing explanation

Fig. 1 is process flow diagram of the present invention;

Fig. 2 is structure table precedence diagram of the present invention.

Embodiment

Below in conjunction with the drawings and specific embodiments, the present invention is described in detail.

Embodiment

As shown in Figure 1, the conflict shifting processing method that a kind of multi-source Spatial Data merges, comprises the following steps:

(1) reading out data, the topological relation of structure inter-entity;

(2) carry out entity coupling;

(3) carry out the adjustment merging of entity coordinate;

(4) whether check amalgamation result there is mistake coupling, and if yes, correct mistake coupling, and return to step (3), if NO, execution step (5);

(5) automatically whether detection there is Graphic Conflicts, if yes, adopts the least square adjustment displacement that manages conflict.

Entity coupling in described step (2) comprises the coupling of an entity, the coupling of the coupling of line entity, face entity.

Employing least square adjustment in described step (5) manage conflict displacement as follows:

If point coordinate is (x in former figure _i ⁰, y _i ⁰) (i=1,2 ..., n) as parameter approximate value, the position coordinates after point is mobile as adjusted value, have

$\left\{\begin{array}{c}{\hat{x}}_i=x_i^0+{\mathrm{\&Delta;x}}_i\\ {\hat{y}}_i=y_i^0+{\mathrm{\&Delta;y}}_i\end{array}\right.(i=\mathrm{1,2},\&CenterDot;\&CenterDot;\&CenterDot;,n)---\left(1\right)$

(Δ x in formula _i, Δ y _i) be coordinate translation amount, be parameter to be solved is adjusted value correction, the adjustment Equations of use comprises as follows:

1) coordinate amount of movement error equation

The point that forms entity should remain on original position as far as possible, and coordinate amount of movement will, close to 0, obtain following error equation

$\left\{\begin{array}{c}{v}_{x_i}={\mathrm{\&Delta;x}}_i\\ v_{y_i}={\mathrm{\&Delta;y}}_i\end{array}\right.---\left(2\right)$

All corresponding two the coordinate amount of movement error equations of each point, therefore the number of whole coordinate amount of movement error equation is total number of the point of twice;

2) relative shift error equation

In order to keep the geometry of entity inside, should make the relative shift of consecutive point on entity close to 0, can obtain following error equation

$\left\{\begin{array}{c}{v}_{x_{i,j+1}}={\mathrm{\&Delta;x}}_i-{\mathrm{\&Delta;x}}_{i+1}\\ v_{y_{i,i+1}}={\mathrm{\&Delta;y}}_i-{\mathrm{\&Delta;y}}_{i+1}\end{array}\right.---\left(3\right)$

3) shape error equation

Keep the original shape of entity when making Coordinate Adjusting, need carry out shape restriction to entity to be adjusted, if keep the turning value of entity frontier point, can obtain error equation and be

$v_{{\mathrm{\&beta;}}_i}=a_{\mathrm{ik}}{\mathrm{\&Delta;x}}_k+b_{\mathrm{ik}}{\mathrm{\&Delta;y}}_k-(a_{\mathrm{ik}}-a_{\mathrm{ij}}){\mathrm{\&Delta;x}}_i-(b_{\mathrm{ik}}-b_{\mathrm{ij}}){\mathrm{\&Delta;y}}_i-a_{\mathrm{ij}}{\mathrm{\&Delta;x}}_j-b_{\mathrm{ij}}{\mathrm{\&Delta;y}}_j-l_{{\mathrm{\&beta;}}_i}---\left(4\right)$

In formula $a_{\mathrm{ij}}=-\frac{\sin {\mathrm{\&alpha;}}_{\mathrm{ij}}}{s_{\mathrm{ij}}};$ $b_{\mathrm{ij}}=\frac{\cos {\mathrm{\&alpha;}}_{\mathrm{ij}}}{s_{\mathrm{ij}}};$ $a_{\mathrm{ik}}=-\frac{\sin {\mathrm{\&alpha;}}_{\mathrm{ik}}}{s_{\mathrm{ik}}};$ $b_{\mathrm{ik}}=\frac{\cos {\mathrm{\&alpha;}}_{\mathrm{ik}}}{s_{\mathrm{ik}}}$

$s_{\mathrm{ij}}=\sqrt{{({x_j}^0-{x_i}^0)}^2+{({y_j}^0-{y_i}^0)}^2};$ $s_{\mathrm{ik}}=\sqrt{{({x_k}^0-{x_i}^0)}^2+{({y_k}^0-{y_i}^0)}^2}$

$l_{{\mathrm{\&beta;}}_i}={\mathrm{\&beta;}}_i-{\mathrm{\&beta;}}_i^0;$ ${\mathrm{\&beta;}}_i^0=\arctan \frac{{y_k}^0-{y_i}^0}{{x_k}^0-{x_i}^0}-\arctan \frac{{y_j}^0-{y_i}^0}{{x_j}^0-{x_i}^0}$

Wherein, α _ik, α _ijbe respectively the position angle of ik, ij direction, by the approximation calculation of coordinate adjustment value, obtained, the former angle value that β is corner point, the number of shape error equation is the number of entity corner point;

4) conflict shift error equation

Graphic Conflicts can be divided into a little and conflicting with face with line, line with face, line with line, point with point, point,

Point is tried to achieve with the range formula of point

$v_{D(P_i,P_j)}+D(P_i,P_j)=\sqrt{{({\hat{x}}_j-{\hat{x}}_i)}^2+{({\hat{y}}_j-{\hat{y}}_i)}^2}---\left(5\right)$

Wherein, the error of distance observation, D (P _i, P _j) be a P _i(x _i, y _i) and P _j(x _j, y _j) distance, with it is the coordinate after its displacement;

Linearization obtains its error equation:

v _dist＝-a _ijΔx _i+b _ijΔy _i+a _ijΔx _j-b _ijΔy _j-l _dist (6)

A wherein _ij=cos α _ij, b _ij=-sin α _ij, α _ijapproximation calculation by coordinate adjustment value obtains, l _distvalue be divided into two kinds of situations: if the distance D of point and point before mobile _point-pointbe less than minimum threshold of distance mindist, l _dist=mindist-D _point-point, wherein if D _point-pointbe greater than mindist and be less than l of 1.5*mindist _dist=0;

Point with the range equation of line segment is:

$v_{D(P_i,L_{P_j{P}_k})}+D_{(P_i,L_{P_j{P}_k})}=\frac{|({\hat{y}}_j-{\hat{y}}_k){\hat{x}}_i+({\hat{x}}_k-{\hat{x}}_j){\hat{y}}_i+({\hat{x}}_j{\hat{y}}_k-{\hat{x}}_k{\hat{y}}_j)|}{\sqrt{{({\hat{y}}_k-{\hat{y}}_j)}^2+{({\hat{x}}_k-{\hat{x}}_j)}^2}}---\left(7\right)$

Wherein, with respectively apart from observed reading and error, with two end points of the point after adjustment and line segment;

Linearization obtains its error equation:

$v_{D(P_i,L_{P_j{P}_k})}=a_{\mathrm{ijk}}{\mathrm{\&delta;x}}_i+b_{\mathrm{ijk}}{\mathrm{\&delta;y}}_i+c_{\mathrm{ijk}}{\mathrm{\&delta;x}}_j+d_{\mathrm{ijk}}{\mathrm{\&delta;y}}_j+e_{\mathrm{ijk}}{\mathrm{\&delta;x}}_k+f_{\mathrm{ijk}}{\mathrm{\&delta;y}}_k-l_{D(P_i,P_j{P}_k)}---\left(8\right)$

Wherein, (x _i ⁰, y _i ⁰, (x _j ⁰, y _j ⁰) and (x _k ⁰, y _k ⁰) be the approximate value of coordinate, and

$a_{\mathrm{ijk}}=\frac{y_j^0-y_k^0}{{({(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2)}^{1/2}};$

$b_{\mathrm{ijk}}=\frac{x_k^0-x_j^0}{{({(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2)}^{1/2}};$

$c_{\mathrm{ijk}}=\frac{(y_k^0-y_i^0)[{(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2]+(x_k^0-x_j^0)[(y_j^0-y_k^0)x_i^0-(x_k^0-x_j^0)y_i^0+(x_j^0{y}_k^0-x_k^0{y}_j^0)]}{{({(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2)}^{3/2}};$

$d_{\mathrm{ijk}}=\frac{(x_i^0-x_k^0)[{(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2]+(y_k^0-y_j^0)[(y_j^0-y_k^0)x_i^0-(x_k^0-x_j^0)y_i^0+(x_j^0{y}_k^0-x_k^0{y}_j^0)]}{{({(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2)}^{3/2}};$

$e_{\mathrm{ijk}}=\frac{(y_i^0-y_j^0)[{(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2]+(x_j^0-x_k^0)[(y_j^0-y_k^0)x_i^0-(x_k^0-x_j^0)y_i^0+(x_j^0{y}_k^0-x_k^0{y}_j^0)]}{{({(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2)}^{3/2}};$

$f_{\mathrm{ijk}}=\frac{(x_j^0-x_i^0)[{(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2]+(y_j^0-y_k^0)[(y_j^0-y_k^0)x_i^0-(x_k^0-x_j^0)y_i^0+(x_j^0{y}_k^0-x_k^0{y}_j^0)]}{{({(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2)}^{3/2}};$

$l_{D(P_i,L_{P_j{P}_k})}=D_{(P_i,L_{P_j{P}_k})}-\frac{|(y_j^0-y_k^0)x_i^0+(x_k^0-x_j^0)y_i^0+(x_j^0{y}_k^0-x_k^0{y}_j^0)|}{\sqrt{{(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2}}.$

L in formula _distvalue with point with point error equation identical, if D _point-line< mindist, l _dist=min dist-D _point-lineif, mindist < D _point-line< 1.5*mindist, l _dist=0;

Point with the distance definition of face is: this,, to the minor increment forming in the distance of all polylines of this face, thus, can be similar to formula (8) and set up the distance error equation that point arrives face;

Line to the distance definition of line is: for line segment L _iand L _j, its distance is

$D_{(L_i,L_j)}=\underset{P_i\&Element;L_i}{\min }\left\{\underset{P_j\&Element;L_j}{\min }\right\{D(P_i,P_j)\left\}\right\}---\left(9\right)$

Wherein, P _iline segment L _ion any point, P _jline segment L _jon any point, D (P _i, P _j) be a P _iand P _jdistance, according to formula (5), calculate;

Line to the distance definition of face is: for line segment L _iwith polygon A _j, its distance is

$D_{(L_i,A_j)}=\underset{P_i\&Element;L_i}{\min }\left\{\underset{P_j\&Element;A_j}{\min }\right\{D(P_i,P_j)\left\}\right\}---\left(10\right)$

Wherein, P _iline segment L _ion any point, P _jpolygon A _jin any point, line is line segment to the minimum value forming in the distance of this polygonal polyline to the distance definition of face,

Line to the distance definition of face is: for line segment A _iwith polygon A _j, its distance is

$D_{(A_i,A_j)}=\underset{P_i\&Element;A_i}{\min }\left\{\underset{P_j\&Element;A_j}{\min }\right\{D(P_i,P_j)\left\}\right\}---\left(11\right)$

Wherein, P _ipolygon A _iin any point, P _jpolygon A _jin any point, the various equations of simultaneous, are indirect adjustment problem, its universal model is:

v＝Ax-l (12)

In formula, x is parameter vector, x=(Δ x ₁Δ y ₁Δ x _nΔ y _n) ^t; A is factor arrays; L is constant vector; 1 is constant vector; V is residual vector, by the principle of least square, can obtain normal equation and is

N _Ax-U＝0 (13)

In formula: N _a=A ^tpA; U=A ^tpl, P is weight matrix;

Separating normal equation can obtain:

$x=N_A^{-1}\&CenterDot;U---\left(14\right)$

Parameter x substitution error equation by trying to achieve, obtains residual error v and adjusted value coordinate adjustment value is the coordinate after a movement;

The inverse of weight matrix computing formula of coordinate adjustment value is

$Q_{\hat{x}\hat{x}}{\left(A^T\mathrm{PA}\right)}^{-1}=N_A^{-1}---\left(15\right)$

The factor arrays that wherein A is error equation, p is power battle array, the accuracy formula of adjusted value is

${\widehat{\mathrm{\&sigma;}}}_{{\hat{x}}_i}=\&PlusMinus;{\widehat{\mathrm{\&sigma;}}}_0\sqrt{Q_{{\hat{x}}_i{\hat{x}}_i}}---\left(16\right)$

Wherein for the pivot of matrix, error in weight unit

Realize hardware system of the present invention and comprise that Organization of Data module, pattern process module, entity matching module, adjustment merge module, conflict shift module 5 modules altogether.The function of each module is as follows:

1, Organization of Data module

This module is mainly to read GIS data file, generates the topological relation of a some entity, line entity, face entity, sets up the data structure on summit after vertex data structure, polyline data structure, polygon data structure, adjustment, store graphics data.

2, pattern process module

Pattern process module comprises the basic operation of GIS polar plot, as amplified, dwindle, the control of translation, figure layer etc., also be included in the function to graphics process in map conflation process, as manual matching feature, remove matching relationship, the graphic result after being combined adds up etc.

3, entity matching module

Entity matching module can carry out respectively according to matching algorithm used herein the coupling of an entity, line entity, face entity, and can know the matching relationship between display entity.

4, adjust and merge module

The adjustment of entity merges module and not only can realize and adjust the Indirect Adjustment Method merging, and clearly illustrates the result of adjusting after merging, and can also realize some other classical algorithm, so as with method comparison herein.

5, conflict shift module

Conflict shift module is mainly that the Graphic Conflicts problem in the contingency question that map conflation is caused is carried out shifting processing, and can demonstrate the result after displacement.

System has adopted point-line or point-face form, and only storage forms the point coordinate of line or face, so after having read file data, also needs to build the topological relation of inter-entity.Build topological relation mainly for line entity and face entity.In building process, set up a some table, line segment table, line segment-line string list, line string list, at this, the initial end points of line string must be node or the end end points that is not connected with other any line, and face entity also needs to generate line string-face table and face table.The point table coordinate of main memory point, the coupling period of the degree of point, this point, the line segment number that connects at this point etc.; The starting point of line segment table storage line segment and termination period; What line segment-line string list played is the effect of a connecting line segment table and line string list, each section of line segment number comprising according to line string of end to end sequential storage; Line string list is stored this position of line string initial segment in line segment-line string list, and actual is an index value, also comprises in addition initial period, stops period, the matched line string of left and right polygon number, this line string number etc.; Line string-face table has been also the effect of a connecting line string list and face table, mainly stores the line string number that a face entity comprises; Face table has been stored the start line string number in online string-face table.The order of the information of each table and structure table is as shown in table 1, Fig. 2.

Table 1 builds the information of all kinds of tables of topological relation

Point table	Line segment table	Line segment-line string list
			The coupling period of this point of line segment number that the degree of the coordinate points of point connects at this point	Initial period terminating point number	Line segment number
Line string list	Line string-face table	Face table
			The initial period terminating point of the index value left and right polygon number of initial segment	Line string number	The index value of start line string

Claims (2)

1. the conflict shifting processing method that multi-source Spatial Data merges, is characterized in that, comprises the following steps:

(1) reading out data, the topological relation of structure inter-entity;

(2) carry out entity coupling;

(3) carry out the adjustment merging of entity coordinate;

(4) whether check amalgamation result there is mistake coupling, and if yes, correct mistake coupling, and return to step (3), if NO, execution step (5);

(5) automatically whether detection there is Graphic Conflicts, if yes, adopts the least square adjustment displacement that manages conflict;

Employing least square adjustment in described step (5) manage conflict displacement as follows:

If point coordinate is in former figure as parameter approximate value, the position coordinates after point is mobile as adjusted value, have

$\left\{\begin{array}{cc}{\hat{x}}_i=x_i^0+{\mathrm{\&Delta;x}}_i& \\ {\hat{y}}_i=y_i^0+{\mathrm{\&Delta;y}}_i& (i=\mathrm{1,2},\&CenterDot;\&CenterDot;\&CenterDot;,n)\end{array}\right.---\left(1\right)$

(Δ x in formula _i, Δ y _i) be coordinate translation amount, be parameter to be solved is adjusted value modified value, the adjustment Equations of use comprises as follows:

1) coordinate amount of movement error equation

The point that forms entity should remain on original position as far as possible, and coordinate amount of movement will, close to 0, obtain following error equation

$\left\{\begin{array}{c}{v}_{x_i}={\mathrm{\&Delta;x}}_i\\ v_{y_i}={\mathrm{\&Delta;y}}_i\end{array}\right.---\left(2\right)$

All corresponding two the coordinate amount of movement error equations of each point, therefore the number of whole coordinate amount of movement error equation is total number of the point of twice;

2) relative shift error equation

In order to keep the geometry of entity inside, should make the relative shift of consecutive point on entity close to 0, can obtain following error equation

$\left\{\begin{array}{c}{v}_{x_{i,i+1}}={\mathrm{\&Delta;x}}_i-{\mathrm{\&Delta;x}}_{i+1}\\ v_{y_{i,i+1}}={\mathrm{\&Delta;y}}_i-{\mathrm{\&Delta;y}}_{i+1}\end{array}\right.---\left(3\right)$

3) shape error equation

Keep the original shape of entity when making Coordinate Adjusting, need carry out shape restriction to entity to be adjusted, if keep the turning value of entity frontier point, can obtain error equation and be

$v_{{\mathrm{\&beta;}}_i}=a_{\mathrm{ik}}{\mathrm{\&Delta;x}}_k+b_{\mathrm{ik}}{\mathrm{\&Delta;y}}_k-(a_{\mathrm{ik}}-a_{\mathrm{ij}}){\mathrm{\&Delta;x}}_i-(b_{\mathrm{ik}}-b_{\mathrm{ij}}){\mathrm{\&Delta;y}}_i-a_{\mathrm{ij}}{\mathrm{\&Delta;x}}_j-b_{\mathrm{ij}}{\mathrm{\&Delta;y}}_j-{l_{\mathrm{\&beta;}}}_i---\left(4\right)$

In formula $a_{\mathrm{ij}}=-\frac{\sin {\mathrm{\&alpha;}}_{\mathrm{ij}}}{s_{\mathrm{ij}}};$ $b_{\mathrm{ij}}=\frac{\cos {\mathrm{\&alpha;}}_{\mathrm{ij}}}{s_{\mathrm{ij}}};$ $a_{\mathrm{ik}}=-\frac{\sin {\mathrm{\&alpha;}}_{\mathrm{ik}}}{s_{\mathrm{ik}}};$ $b_{\mathrm{ik}}=\frac{\cos {\mathrm{\&alpha;}}_{\mathrm{ik}}}{s_{\mathrm{ik}}}$

$s_{\mathrm{ij}}=\sqrt{{({x_j}^0-{x_i}^0)}^2+{({y_j}^0-{y_i}^0)}^2};$ $s_{\mathrm{ik}}=\sqrt{{({x_k}^0-{x_i}^0)}^2+{({y_k}^0-{y_i}^0)}^2}$

${l_{\mathrm{\&beta;}}}_i={\mathrm{\&beta;}}_i-{\mathrm{\&beta;}}_i^0;$ ${\mathrm{\&beta;}}_i^0=\arctan \frac{{y_k}^0-{y_i}^0}{{x_k}^0-{x_i}^0}-\arctan \frac{{y_j}^0-{y_i}^0}{{x_j}^0-{x_i}^0}$

Wherein, α _ik, α _ijbe respectively the position angle of ik, ij direction, by the approximation calculation of coordinate adjustment value, obtained, the former angle value that β is corner point, the number of shape error equation is the number of entity corner point;

4) conflict shift error equation

Graphic Conflicts can be divided into a little and conflicting with face with line, line with face, line with line, point with point, point,

Point is tried to achieve with the range formula of point

$v_{D(P_i,P_j)}+D(P_i,P_j)=\sqrt{{\left({\hat{x}}_j-{\hat{x}}_i\right)}^2+{\left({\hat{y}}_j-{\hat{y}}_i\right)}^2}---\left(5\right)$

Wherein, v _{d (Pi, Pj)}the error of distance observation, D (P _i, P _j) be a P _i(x _i, y _i) and P _j(x _j, y _j) distance, with it is the coordinate after its displacement;

Linearization obtains its error equation:

$v_{\mathrm{dist}}={-a}_{\mathrm{ij}}{\mathrm{\&Delta;x}}_i+b_{\mathrm{ij}}{\mathrm{\&Delta;y}}_i+a_{\mathrm{ij}}{\mathrm{\&Delta;x}}_j-b_{\mathrm{ij}}{\mathrm{\&Delta;y}}_j-l_{\mathrm{dist}}---\left(6\right)$

A wherein _ij=cos α _ij, b _ij=-sin α _ij, α _ijapproximation calculation by coordinate adjustment value obtains, l _distvalue be divided into two kinds of situations: if the distance D of point and point before mobile _point-pointbe less than minimum threshold of distance mindist, l _dist=mindist – D _point-line, wherein if D _point-pointbe greater than mindist and be less than l of 1.5*mindist _dist=0;

Point with the range equation of line segment is:

$v_{D(P_i,L_{P_j{P}_k})}+D_{(P_i,L_{P_j{P}_k})}=\frac{|({\hat{y}}_j-{\hat{y}}_k){\hat{x}}_i+({\hat{x}}_k-{\hat{x}}_j){\hat{y}}_i+({\hat{x}}_j{\hat{y}}_k-{\hat{x}}_k{\hat{y}}_j)|}{\sqrt{{({\hat{y}}_k-{\hat{y}}_j)}^2+{({\hat{x}}_k-{\hat{x}}_j)}^2}}---\left(7\right)$

Wherein, with respectively apart from observed reading and error, with two end points of the point after adjustment and line segment;

Linearization obtains its error equation:

$v_{D(P_i,L_{P_j{P}_k})}=a_{\mathrm{ijk}}{\mathrm{\&delta;x}}_i+b_{\mathrm{ijk}}{\mathrm{\&delta;y}}_i+c_{\mathrm{ijk}}{\mathrm{\&delta;x}}_j+d_{\mathrm{ijk}}{\mathrm{\&delta;y}}_j+e_{\mathrm{ijk}}{\mathrm{\&delta;x}}_k+f_{\mathrm{ijk}}{\mathrm{\&delta;y}}_k-l_{D(P_i,P_j{P}_k)}---\left(8\right)$

Wherein, with the approximate value of coordinate, and

$a_{\mathrm{ijk}}=\frac{y_j^0-y_k^0}{{({(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2)}^{1/2}};$

$b_{\mathrm{ijk}}=\frac{x_k^0-x_j^0}{{({(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2)}^{1/2}};$

$c_{\mathrm{ijk}}=\frac{(y_k^0-y_i^0)[{(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2]+(x_k^0-x_j^0)[(y_j^0-y_k^0)x_i^0-(x_k^0-x_j^0)y_i^0+(x_j^0{y}_k^0-x_k^0{y}_j^0)]}{{({(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2)}^{3/2}};$

$d_{\mathrm{ijk}}=\frac{(x_i^0-x_k^0)[{(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2]+(y_k^0-y_j^0)[(y_j^0-y_k^0)x_i^0-(x_k^0-x_j^0)y_i^0+(x_j^0{y}_k^0-x_k^0{y}_j^0)]}{{({(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2)}^{3/2}};$

$e_{\mathrm{ijk}}=\frac{(y_i^0-y_j^0)[{(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2]+(x_j^0-x_k^0)[(y_j^0-y_k^0)x_i^0-(x_k^0-x_j^0)y_i^0+(x_j^0{y}_k^0-x_k^0{y}_j^0)]}{{({(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2)}^{3/2}};$

$f_{\mathrm{ijk}}=\frac{(x_j^0-x_i^0)[{(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2]+(y_j^0-y_k^0)[(y_j^0-y_k^0)x_i^0-(x_k^0-x_j^0)y_i^0+(x_j^0{y}_k^0-x_k^0{y}_j^0)]}{{({(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2)}^{3/2}};$

$l_{D(P_i,L_{P_j{P}_k})}=D_{(P_i,L_{P_j{P}_k})}-\frac{|(y_j^0-y_k^0)x_i^0+(x_k^0-x_j^0)y_i^0+(x_j^0{y}_k^0-x_k^0{y}_j^0)|}{\sqrt{{(y_k^0-y_j^0)}^2+{(x_k^0-x_j^0)}^2}}.$

L in formula _distvalue with point with point error equation identical, if D _point-line<mindist,

L _dist=mindist – D _point-lineif, mindist<D _point-line<1.5*mindist, l _dist=0;

Point with the distance definition of face is: this is to the minor increment forming in the distance of all polylines of this face, and thus, in like manner available formula (8) is set up the distance error equation that point arrives face;

Line to the distance definition of line is: for line segment L _iand L _j, its distance is

$D_{(L_i,L_j)}=\underset{P_i\&Element;L_i}{\min }\left\{\underset{P_j\&Element;L_j}{\min }\right\{D(P_i,P_j)\left\}\right\}---\left(9\right)$

Wherein, P _iline segment L _ion any point, P _jline segment L _jon any point, D (P _i, P _j) be a P _iand P _jdistance, according to formula (5), calculate;

Line to the distance definition of face is: for line segment L _iwith polygon A _j, its distance is

$D_{(L_i,A_j)}=\underset{P_i\&Element;L_i}{\min }\left\{\underset{P_j\&Element;A_j}{\min }\right\{D(P_i,P_j)\left\}\right\}---\left(10\right)$

Wherein, P _iline segment L _ion any point, P _jpolygon A _jin any point, line is line segment to the minimum value forming in the distance of this polygonal polyline to the distance definition of face,

Line to the distance definition of face is: for line segment A _iwith polygon A _j, its distance is

$D_{(A_i,A_j)}=\underset{P_i\&Element;A_i}{\min }\left\{\underset{P_j\&Element;A_j}{\min }\right\{D(P_i,P_j)\left\}\right\}---\left(11\right)$

Wherein, P _ipolygon A _iin any point, P _jpolygon A _jin any point,

Simultaneous error equation (2), (3), (4), (5), (6), (7), (8), be indirect adjustment problem, and its model is:

v＝Ax-l （12）

In formula, x is parameter vector, x=(Δ x ₁Δ y ₁Δ x _nΔ y _n) ^t; A is factor arrays; L is constant vector; V is residual vector, by the principle of least square, can obtain normal equation and is

N _Ax-U＝0 （13）

In formula: N _a=A ^tpA; U=A ^tpl, P is weight matrix;

Separating normal equation can obtain:

$x=N_A^{-1}\&CenterDot;U---\left(14\right)$

Parameter x substitution error equation by trying to achieve, obtains residual error v and adjusted value coordinate adjustment value is the coordinate after a movement;

The inverse of weight matrix computing formula of coordinate adjustment value is

$Q_{\hat{x}\hat{x}}={\left(A^T\mathrm{PA}\right)}^{-1}=N_A^{-1}---\left(15\right)$

The factor arrays that wherein A is error equation, p is power battle array, the accuracy formula of adjusted value is

${\widehat{\mathrm{\&sigma;}}}_{{\hat{x}}_i}=\&PlusMinus;{\widehat{\mathrm{\&sigma;}}}_0\sqrt{Q_{{\hat{x}}_i{\hat{x}}_i}}---\left(16\right)$

Wherein for the pivot of matrix, error in weight unit

2. the conflict shifting processing method that a kind of multi-source Spatial Data according to claim 1 merges, is characterized in that, the entity coupling in described step (2) comprises the coupling of an entity, the coupling of the coupling of line entity, face entity.