CN102074028A - Adaptive spatial interpolation method - Google Patents

Abstract

The invention relates to an adaptive spatial interpolation method, which belongs to the technical field of soil digital mapping. In order to overcome the difficulty of unreliability of spatial interpolation of soil property in the digital soil mapping, the invention provides an interpolation method, which comprises the following steps of: in a geographic space, separating a soil space from a landscape element in a vertical direction, wherein the bottom layer is a soil layer and the upper layer is a landscape element characteristic vector layer; gridding the soil layer according to a certain scale, extracting landscape element characteristic vectors corresponding to any grid units in the grid soil layer, and constructing a landscape element characteristic vector set; classifying the soil space according to the landscape element characteristic vector set so as to obtain a soil space classification map; and establishing an interpolation operator on the basis of the soil space classification map, and interpolating any unknown units in the soil space according to the interpolation operator. By using the technical scheme, the accuracy and the reliability of the space interpolation of the soil property content under a complicated landscape environment are improved and ensured.

Description

The adaptive space interpolation method

Technical field

The present invention relates to soil digital mapping technical field, relate in particular in a kind of being applicable to/large scale on the space interpolation of the attribute information of research objects such as soil, the hydrology and meteorology and the adaptive space interpolation method that numerical map is made under the complicated landscape conditions.

Background technology

Be subjected to man power and material's influence, always limited to the sampling observation of regional soil attribute information, how the face source distribution information of regional soil attribute to be described exactly by limited, discrete observation sampling point data, be the interested difficult point problems of people always.At present, the main method of the space distribution information of description soil attribute is geo-statistic method and soil-landscape model.

The prerequisite that the geo-statistic method is set up is that soil attribute exists spatial autocorrelation.In fact, because many landscape factors of soil attribute audient and artificial factor, soil attribute just has the better space autocorrelation usually on small scale, and with the increase of research yardstick, this spatial autocorrelation is difficult to set up.Therefore, the geo-statistic method can realize the reliable interpolation of soil attribute on the little zonule of soil landscape structure variation, but to the space interpolation of the baroque regional soil attribute of view, is undoubtedly rough, insecure.

Soil-landscape model is predicted soil attribute content by analyzing and extract the relevance of soil attribute and landscape environment variable by certain inference technology.But because the spatial variability of soil attribute is very big, and its relation with the landscape environment variable have highly non-linear, this make the soil landscape model be difficult to estimate exactly in/the regional soil attribute content that has complicated landscape structure feature on the large scale.

Summary of the invention

(1) technical matters that will solve

The technical problem to be solved in the present invention is: how to overcome deficiency of the prior art, provide a kind of guarantee the space interpolation accuracy and applicable in/large scale on the adaptive space interpolation method of complicated landscape conditions to solve the unreliable difficult problem of the space interpolation of soil attribute in the digital soil drawing.

(2) technical scheme

In order to solve the problems of the technologies described above, the invention provides a kind of adaptive space interpolation method, comprise the following steps:

S1: on geographical space, soil space and landscape element are separated in vertical direction;

Wherein, bottom is the pedological map layer, is the passive active layer that is subjected to, and the upper strata is a landscape element proper vector layer, is the factor combination that causes soil variation;

S2: with the gridding of described pedological map layer, extract the pairing landscape element proper vector in arbitrary mess unit in the described gridding pedological map layer by user-defined yardstick, and make up the landscape element set of eigenvectors;

S3: classify according to described landscape element proper vector set pair soil space, obtain the soil space classification chart;

S4: set up interpolation operator, any unknown elements in the soil space is carried out interpolation according to described interpolation operator based on described soil space classification chart.

Preferably, described step S2 specifically comprises:

S201:, the gridding equally spacedly of described pedological map layer is divided into the plurality of grids unit for given pedological map layer;

S202: the requirement of view element characteristic vector is come the eigenwert of view element characteristic vector is extracted according to the user;

S203: obtain the mapping of landscape element proper vector on each described grid cell according to described eigenwert;

S204:, make up the landscape element set of eigenvectors that obtains corresponding described pedological map layer according to the mapping of described landscape element proper vector on each described grid cell.

Preferably, described proper vector comprises vegetation pattern, hyposographic features, land type key element and road network;

The granule size of the grid cell of described pedological map layer is consistent with the granularity of the minimum landscape element proper vector that can obtain.

Preferably, described step S3 specifically comprises:

S301: for given pedological map layer and by the resulting landscape element set of eigenvectors of step S2, regard any two different landscape element proper vectors as two different nodes, the similarity between described two different landscape element proper vectors is regarded the weights between described two different nodes as; Described weights are defined as Euclidean distance between described two different landscape element proper vectors;

S302: preset threshold values,, then will connect the limit between described two different nodes, thereby set up the similar diagram of described landscape element set of eigenvectors if described weights are no more than this threshold value;

S303: make up the object matrix and the adjacency matrix of similar diagram, thereby obtain Laplce's matrix of similar diagram;

S304: described Laplce's matrix is carried out svd, obtain proper value of matrix;

S305: described proper value of matrix is arranged from small to large, described similar diagram is divided into two, obtain the soil space classification chart according to the second little pairing landscape element proper vector of eigenwert in the described proper value of matrix.

Preferably, described step S305 is specially:

Obtain its pairing landscape element proper vector according to the second little eigenwert in the described proper value of matrix;

Sign according to each component in the described landscape element proper vector is divided into two pairing landscape element set of eigenvectors;

According to classification, and then obtain the soil space classification chart indirectly to described landscape element proper vector.

Preferably, described step S4 specifically comprises:

S401: preestablish the minimum monitoring sampling point number that participates in interpolation;

S402: utilize the soil space classification chart that interpolating unit is classified, the interpolating unit that will be in the identical soil space class is orientated same class as; Belong to of a sort interpolating unit and constitute an interpolating unit class; That is corresponding soil space class of each interpolating unit class;

S403: judge whether the monitoring sampling point number in a certain interpolating unit class monitors the sampling point number more than or equal to default minimum in the step 401;

If then the observed value structure according to all the monitoring sampling points in the described interpolating unit class obtains interpolation operator, and changes step S404 over to;

If not, then change described interpolating unit class repeated execution of steps S403;

S404: according to described interpolation operator the arbitrary unknown elements in the described interpolating unit class is carried out interpolation calculation, obtain the interpolation result of unknown elements.

Preferably, the mode classification among the described step S402 is a binary tree data directory frame mode, specifically comprises:

With described arbitrary interpolating unit class definition is a leaf node, this interpolating unit category information of storage on a leaf node;

The index information of two child nodes is pointed in storage on each father node;

Wherein, from bottom to up, the space homogeney between lower floor's child node is set to strong than between the father node of upper strata, and has the space homogeney of the contiguous leaf node of identical father node to be set to the strongest.

Preferably, the judged result among the described step S403 is under the situation not, and the process of changing described interpolating unit class repeating step S403 is specially:

If the monitoring sampling point number in the described estimative interpolating unit class is less than described default minimum monitoring sampling point number, then from be adjacent node, search monitoring sampling point is finished interpolation in another promptly adjacent interpolating unit class;

If the monitoring sampling point number sum in described two adjacent estimative interpolating unit classes, then dates back to the common father node place search monitoring sampling point that described two adjacent interpolating unit classes are had still less than described default minimum monitoring sampling point number and finishes interpolation.

Preferably, among the described step S404, the interpolation result Z (X of unknown elements ₀) be calculated as follows:

$Z\left(X_0\right)=\underset{i=1}{\overset{N\left(D_i\right)}{\mathrm{\Σ}}}{\mathrm{\λ}}_{i}Z\left(X_i\right),$

Wherein, Z={z (x ₁, y ₁), z (x ₂, y ₂) ..., z (x _k, y _k) for falling into soil space class D _iInterior monitoring sampling point collection, herein, soil space class D _iBe the interpolating unit class, N (D _i) refer to soil space class D _iIn the monitoring sampling point number that comprised, Z (X _i) refer to corresponding D _iInterior monitoring sampling point z (x _i, y _i) observed reading, λ _iBe Z (X _i) corresponding weights.

(3) beneficial effect

Technical scheme provided by the present invention is to drive with the pass between soil space unit and the landscape element proper vector, by extraction landscape element set of eigenvectors and to its classification, produce the different soils spatial class, foundation is based on the interpolation operator of soil space classification chart, interpolation operator is the best interpolation problem on the local homogeneity space adaptively with the space interpolation problem reduction on the heterogeneous space of the overall situation, estimation to the soil attribute of unknown elements will be carried out in the soil space class of homogeneity, thereby improve and guaranteed the accuracy and the reliability of the space interpolation of soil attribute content under the complicated landscape environment, and this invention technology also is applicable to ecology, the space interpolation of other field such as the hydrology and meteorology and numerical map are made.

Description of drawings

Fig. 1 is the process flow diagram of the related adaptive space interpolation method of the specific embodiment of the invention;

Fig. 2 is the structure synoptic diagram of the related binary tree index structure of the specific embodiment of the invention;

Fig. 3 is the related interpolating unit search synoptic diagram based on the binary tree index structure of the specific embodiment of the invention.

Embodiment

For making purpose of the present invention, content and advantage clearer,, the specific embodiment of the present invention is described in further detail below in conjunction with drawings and Examples.

At first, for disambiguation, content related in the following content is done following explanation:

Interpolating unit refers to constitute the basic grid unit of soil space class, sees shown in Figure 2;

Unknown elements refers to not comprise the interpolating unit of monitoring sampling point;

Interpolation operator only carries out interpolation calculation at unknown elements, and the observed reading of interpolating unit is with monitoring sampling point data characterization that this interpolating unit comprised in the interpolation process.

Below, as shown in Figure 1, the adaptive space interpolation method that technical solution of the present invention is related comprises the following steps:

S1: on geographical space, separate soil space is vertical with landscape element; Wherein, bottom is the pedological map layer, is the passive active layer that is subjected to, and the upper strata is a landscape element proper vector layer, is the factor combination that causes soil variation;

S2: with the gridding of described pedological map layer, extract the pairing landscape element proper vector in arbitrary mess unit in the described gridding pedological map layer by user-defined yardstick, and make up the landscape element set of eigenvectors;

Described step S2 specifically comprises:

S201:, described pedological map layer S gridding equally spacedly is divided into several, such as n grid cell, i.e. S={S for given pedological map layer S ₁, S ₂..., S _i..., S _n;

At this moment, suppose landscape element set of eigenvectors F={F ₁, F ₂..., F _i..., F _n, F wherein _iFor the landscape element vector at soil grid cell S _iOn mapping;

Suppose proper vector F arbitrarily _i=(f ₁, f ₂..., f _j..., f _k), f wherein _iFor with grid cell S _iThe eigenwert of corresponding landscape element vector;

S202: the actual requirement of view element characteristic vector is come view element characteristic vector F according to the user _iEigenwert f _iCalculate extraction;

S203: according to described eigenwert f _iObtain the landscape element proper vector at each described soil grid cell S _iOn mapping F _i

S204: according to the mapping F of described landscape element proper vector on each described soil grid cell _i, and then make up the landscape element set of eigenvectors F that obtains corresponding described pedological map layer S.

Described proper vector comprises vegetation pattern, hyposographic features, land type key element and road network; For vegetation pattern and hyposographic features, represent with normalized differential vegetation index (Normalized Difference Vegetation Index-NDVI) and height value; For the land type key element, available categories is represented; For road network, with S _iContained road network distribution density is represented (promptly to cross S _iRoad network length compare S _iArea).In addition, the granule size of the grid cell of described pedological map layer is consistent with the granularity of the minimum landscape element proper vector that can obtain.

For convenience of explanation, suppose that the landscape element that influences soil attribute that can effectively obtain is vegetation pattern and elevation, characterize vegetation pattern and elevation with normalized differential vegetation index and normalization height value respectively.Definition

As follows:

$f_{\mathrm{NDVI}}^i=\frac{{\mathrm{NIR}}_i-{\mathrm{RED}}_i}{{\mathrm{NIR}}_i+{\mathrm{RED}}_i}$

Here, NIR _iAnd RED _iRefer to cell S respectively _iNear-infrared band and the reflectivity of red wave band.

$f_{\mathrm{elevation}}^i=\frac{\mathrm{elevatio}{n}_i-\mathrm{Min}\_\mathrm{elevation}}{\mathrm{Max}\_\mathrm{elevation}-\mathrm{Min}\_\mathrm{elevation}}$

Wherein, elevation _i, Min_elevation and Max_elevation refer to cell S respectively _iHeight value, the elevation minimum value and the maximal value of region S.

S3: classify according to described landscape element proper vector set pair soil space, obtain the soil space classification chart;

Be subjected to landscape environment factor affecting such as vegetation, landform, weather, the space distribution of soil attribute is discontinuous on overall geographical space, but on local geographical space, there is continuity, that is to say that the soil attribute that is positioned at the space homogeneous region is more approximate than the soil attribute that is positioned at the non-space homogeneous region.Therefore, the landscape element that influences soil attribute is classified, can obtain the soil space class indirectly.Like this, the space interpolation problem of soil attribute changes into best interpolation on the local homogeneity subspace on the overall heterogeneous space.

To the view element category is an iterative process repeatedly, needs to determine earlier the classification number of landscape element, so that guarantee the convergence of classification effectively.Normal conditions are to know the classification number of landscape environment key element in advance, therefore need the automatic convergent method of design to realize the landscape element tagsort.

For this reason, this embodiment has been developed a spectrum dividing method and has been realized the soil space classification, and the spectrum dividing method obtains the soil space class indirectly by to the classification of view element characteristic;

Specifically comprise the steps:

S301: for given pedological map layer S={S ₁, S ₂..., S _i..., S _nAnd by the resulting landscape element set of eigenvectors of step S2 F={F ₁, F ₂..., F _i..., F _n, the spectrum dividing method is regarded the landscape element proper vector as node V, and in view of the above, any two different landscape element proper vectors can be seen two different node V as _iWith V _j, the similarity between described two different landscape element proper vectors is regarded two different node V as _iWith node V _jBetween weights W _IjWith described weights W _IjBe defined as W _Ij=‖ F _i-F _j‖ ², the Euclidean distance between promptly described two different landscape element proper vectors;

S302: preset reservation threshold δ, if described weights W _IjBe no more than this predetermined threshold δ, then with described two different node V _iWith node V _jBetween connect limit E, thereby set up described landscape element set of eigenvectors similar diagram G (V, E);

S303: make up object matrix D (G) and the adjacency matrix A (G) of similar diagram G, and satisfy:

Wherein, m is the adjacent node number of node i,

Thereby obtain Laplce's matrix L (G)=D (G)-A (G) of similar diagram G, satisfy:

S304: described Laplce's matrix L (G) is carried out svd (Singular Value Decomposition-SVD), obtain proper value of matrix;

S305: described proper value of matrix is arranged from small to large, according to the second little eigenvalue in the described proper value of matrix ₂Pairing landscape element proper vector V ₂Described similar diagram G is divided into two, obtains the soil space classification chart, specifically comprise:

According to the second little eigenvalue in the described proper value of matrix ₂Obtain its pairing landscape element proper vector V ₂

According to described landscape element proper vector V ₂In the sign of each component, pairing landscape element set of eigenvectors is divided into two;

According to classification, and then obtain the soil space classification chart indirectly to view element characteristic vector.

S4:, obtain soil space category set D={D by view element characteristic vector set is composed cutting operation ₁, D ₂..., D _i..., D _k, satisfy

Satisfy that soil attribute has the space homogeney in the same soil space class, and the soil attribute expressive space is heterogeneous between the different soils spatial class.Space interpolation to unknown elements will be finished in the soil space homogeneous region, can think that soil attribute satisfies spatial autocorrelation in the soil space homogeneity district;

To the estimation of unknown elements, set up interpolation operator based on described soil space classification chart, according to described interpolation operator any unknown elements in the soil space is carried out interpolation.

Described step S4 specifically comprises:

S401: preestablish minimum monitoring sampling point (observation sampling point) the number M that participates in interpolation;

S402: utilize the soil space classification chart that interpolating unit is classified, will be in identical soil space class interpolate value cell location is same class; Belong to of a sort interpolating unit and constitute an interpolating unit class; That is corresponding soil space class of each interpolating unit class;

Mode classification among the described step S402 is a binary tree data directory frame mode, and its classification results binary tree data structure storage specifically comprises:

With described arbitrary interpolating unit class definition is a leaf node, this interpolating unit category information of storage on a leaf node;

The index information of two child nodes is pointed in storage on each father node;

Wherein, from bottom to up, strong than between the father node of upper strata of the space homogeney between lower floor's child node, and have the space homogeney of contiguous leaf node of identical father node the strongest.

Content to above-mentioned steps S402 is carried out following specific explanations:

When carrying out interpolation operation, at first search for M interpolating unit at least at every turn, be used for the property value of approximate estimation unknown elements.The direct search interpolating unit is more time-consuming, and this embodiment is set up the binary tree index structure.

In essence, the building process of binary tree index as shown in Figure 2.Last figure among Fig. 2 has represented a soil space is carried out 9 different subspace classes (representing respectively to nine capitalizations of I with A) that 5 iteration formed after two minutes.Two minutes of soil space is top down to be divided into two secondary subclasses by upper level soil space class two, carried out 5 times one by one after, formed 9 sub spaces classes.Each subclass constitutes a leaf node, leaf node storage subclass information.Such as, after the G class constitutes leaf node among Fig. 2, need the information of the storage G class G class is arranged 18 unit informations that comprised, the monitoring sampling point information that falls into the G class.In addition, each grid cell is represented an interpolating unit among the last figure of Fig. 2, falls into a monitoring of grid cell circle symbolic representation sampling point.Figure below of Fig. 2 is the binary tree structural characterization to 9 sub spaces classes among the last figure.Binary tree data directory structure image has embodied soil space iteration two minutes and last convergent process, and the binary tree index structure that builds up has following characteristic:

● each leaf node characterizes the inseparable soil space class of minimum, the interpolating unit information that this node storage soil space class and soil space class are comprised.

● the index information of two child nodes is pointed in each father node storage.

● from bottom to up, strong than between the father node of upper strata of the space homogeney between lower floor's child node, and have the space homogeney of contiguous leaf node of identical father node the strongest.

In this embodiment, interpolation operation always carries out in the soil space class, just finishes on leaf node.So the search to interpolating unit at first needs to add up the interpolating unit number that leaf node comprises.Given k the soil space class D that the soil grid cell is formed _i, fall into D _iIn interpolating unit collection Z _i, corresponding leaf node V _iStorage S _iAnd Z _iBecause the observation sampling point is limited, leaf node V _iThe interpolating unit number that comprises | Z _i| difference can cause three kinds of different search patterns:

(1) the monitoring number of samples that comprises of leaf node | Z _i| 〉=M (minimum monitoring number of samples), this moment, search need not to carry out, and the monitoring sampling point data that interpolation operator directly utilizes leaf node to comprise are finished the interpolation to unknown elements.

(2) the interpolating unit number that comprises of leaf node | Z _i|＜M, from contiguous brotgher of node search monitoring sampling point.

Monitoring total sample＜M that (3) two contiguous leaf nodes comprise dates back search monitoring sampling point on other leaf node with common ancestor.

As shown in Figure 3, three different search patterns have been shown intuitively.Given minimum monitoring number of samples is 5, comprises 5 monitoring sampling points in the G class, satisfy condition, so to unknown elements (be with question mark "? " the grid cell of symbol) during interpolation, interpolation is directly carried out, i.e. search pattern O; 3 monitoring sampling points are arranged, less than minimum monitoring number of samples, so to monitoring sampling point, i.e. search pattern P from adjacent node (H class) search during the unknown elements interpolation in the I class in the I class; To in the H class during unknown elements interpolation, owing to do not have known monitoring sampling point in the H class, search monitoring sampling point in the I class earlier, however search monitoring number of samples does not still satisfy condition in the I class, so to turning to upper layer node search, i.e. search pattern Q.

Continue to describe further concrete operation steps below:

S403: judge whether the monitoring sampling point number in a certain interpolating unit class monitors the sampling point number more than or equal to default minimum in the step 401;

If then the observed value structure according to all the monitoring sampling points in the described interpolating unit class obtains interpolation operator, and changes step S404 over to;

If not, then change described interpolating unit class repeated execution of steps S403;

Judged result among the described step S403 is under the situation not, and the process of changing described interpolating unit class repeating step S403 is specially:

If the monitoring sampling point number in the described estimative interpolating unit class is less than described default minimum monitoring sampling point number, then from be adjacent node, search monitoring sampling point is finished interpolation in another promptly adjacent interpolating unit class;

Have common father node search monitoring sampling point and finish interpolation if the monitoring sampling point number sum in described two adjacent estimative interpolating unit classes, then dates back to described two adjacent interpolating unit classes still less than described default minimum monitoring sampling point number.

S404: according to interpolation operator the arbitrary unknown elements in the described interpolating unit class is carried out interpolation calculation, obtain the interpolation result of unknown elements.

Among the described step S404, be calculated as follows the interpolation result Z (X of unknown elements ₀):

$Z\left(X_0\right)=\underset{i=1}{\overset{N\left(D_i\right)}{\mathrm{\Σ}}}{\mathrm{\λ}}_{i}Z\left(X_i\right),$

Wherein, Z={z (x ₁, y ₁), z (x ₂, y ₂) ..., z (x _k, y _k), 1≤i≤k＜n is for falling into soil space class D _iInterior monitoring sampling point collection, herein, soil space class D _iBe the interpolating unit class, unknown site z (x ₀, y ₀) ∈ D _i, N (D _i) refer to soil space class D _iIn the monitoring sampling point number that comprised, Z (X _i) refer to corresponding D _iInterior monitoring sampling point z (x _i, y _i) observed reading, λ _iBe Z (X _i) corresponding weights.

The above only is a preferred implementation of the present invention; should be pointed out that for those skilled in the art, under the prerequisite that does not break away from the technology of the present invention principle; can also make some improvement and distortion, these improvement and distortion also should be considered as protection scope of the present invention.

Claims (9)

1. an adaptive space interpolation method is characterized in that, comprises the following steps:

S1: on geographical space, soil space and landscape element are separated in vertical direction;

Wherein, bottom is the pedological map layer, is the passive active layer that is subjected to, and the upper strata is a landscape element proper vector layer, is the factor combination that causes soil variation;

S2: with the gridding of described pedological map layer, extract the pairing landscape element proper vector in arbitrary mess unit in the described gridding pedological map layer by user-defined yardstick, and make up the landscape element set of eigenvectors;

S3: classify according to described landscape element proper vector set pair soil space, obtain the soil space classification chart;

S4: set up interpolation operator, any unknown elements in the soil space is carried out interpolation according to described interpolation operator based on described soil space classification chart.

2. adaptive space interpolation method as claimed in claim 1 is characterized in that, described step S2 specifically comprises:

S201:, the gridding equally spacedly of described pedological map layer is divided into the plurality of grids unit for given pedological map layer;

S202: the requirement of view element characteristic vector is come the eigenwert of view element characteristic vector is extracted according to the user;

S203: obtain the mapping of landscape element proper vector on each described grid cell according to described eigenwert;

S204:, make up the landscape element set of eigenvectors that obtains corresponding described pedological map layer according to the mapping of described landscape element proper vector on each described grid cell.

3. adaptive space interpolation method as claimed in claim 2 is characterized in that, described proper vector comprises vegetation pattern, hyposographic features, land type key element and road network;

The granule size of the grid cell of described pedological map layer is consistent with the granularity of the minimum landscape element proper vector that can obtain.

4. adaptive space interpolation method as claimed in claim 1 is characterized in that, described step S3 specifically comprises:

S301: for given pedological map layer and by the resulting landscape element set of eigenvectors of step S2, regard any two different landscape element proper vectors as two different nodes, the similarity between described two different landscape element proper vectors is regarded the weights between described two different nodes as; Described weights are defined as Euclidean distance between described two different landscape element proper vectors;

S302: preset threshold values,, then will connect the limit between described two different nodes, thereby set up the similar diagram of described landscape element set of eigenvectors if described weights are no more than this threshold value;

S303: make up the object matrix and the adjacency matrix of similar diagram, thereby obtain Laplce's matrix of similar diagram;

S304: described Laplce's matrix is carried out svd, obtain proper value of matrix;

S305: described proper value of matrix is arranged from small to large, described similar diagram is divided into two, obtain the soil space classification chart according to the second little pairing landscape element proper vector of eigenwert in the described proper value of matrix.

5. adaptive space interpolation method as claimed in claim 4 is characterized in that, described step S305 is specially:

Obtain its pairing landscape element proper vector according to the second little eigenwert in the described proper value of matrix;

Sign according to each component in the described landscape element proper vector is divided into two pairing landscape element set of eigenvectors;

According to classification, and then obtain the soil space classification chart indirectly to described landscape element proper vector.

6. adaptive space interpolation method as claimed in claim 1 is characterized in that, described step S4 specifically comprises:

S401: preestablish the minimum monitoring sampling point number that participates in interpolation;

S402: utilize the soil space classification chart that interpolating unit is classified, the interpolating unit that will be in the identical soil space class is orientated same class as; Belong to of a sort interpolating unit and constitute an interpolating unit class; That is corresponding soil space class of each interpolating unit class;

S403: judge whether the monitoring sampling point number in a certain interpolating unit class monitors the sampling point number more than or equal to default minimum in the step 401;

If then the observed value structure according to all the monitoring sampling points in the described interpolating unit class obtains interpolation operator, and changes step S404 over to;

If not, then change described interpolating unit class repeated execution of steps S403;

S404: according to described interpolation operator the arbitrary unknown elements in the described interpolating unit class is carried out interpolation calculation, obtain the interpolation result of unknown elements.

7. adaptive space interpolation method as claimed in claim 6 is characterized in that, the mode classification among the described step S402 is a binary tree data directory frame mode, specifically comprises:

With described arbitrary interpolating unit class definition is a leaf node, this interpolating unit category information of storage on a leaf node;

The index information of two child nodes is pointed in storage on each father node;

Wherein, from bottom to up, the space homogeney between lower floor's child node is set to strong than between the father node of upper strata, and has the space homogeney of the contiguous leaf node of identical father node to be set to the strongest.

8. adaptive space interpolation method as claimed in claim 7 is characterized in that, the judged result among the described step S403 is under the situation not, and the process of changing described interpolating unit class repeating step S403 is specially:

If the monitoring sampling point number in the described estimative interpolating unit class is less than described default minimum monitoring sampling point number, then from be adjacent node, search monitoring sampling point is finished interpolation in another promptly adjacent interpolating unit class;

If the monitoring sampling point number sum in described two adjacent estimative interpolating unit classes, then dates back to the common father node place search monitoring sampling point that described two adjacent interpolating unit classes are had still less than described default minimum monitoring sampling point number and finishes interpolation.

9. adaptive space interpolation method as claimed in claim 6 is characterized in that, among the described step S404, and the interpolation result Z (X of unknown elements ₀) be calculated as follows:

$Z\left(X_0\right)=\underset{i=1}{\overset{N\left(D_i\right)}{\mathrm{\Σ}}}{\mathrm{\λ}}_{i}Z\left(X_i\right),$

Wherein, Z={z (x ₁, y ₁), z (x ₂, y ₂) ..., z (x _k, y _k) for falling into soil space class D _iInterior monitoring sampling point collection, herein, soil space class D _iBe the interpolating unit class, N (D _i) refer to soil space class D _iIn the monitoring sampling point number that comprised, Z (X _i) refer to corresponding D _iInterior monitoring sampling point z (x _i, y _i) observed reading, λ _iBe Z (X _i) corresponding weights.

Images