CN102637227B - Land resource assessment factor scope dividing method based on shortest path - Google Patents
Land resource assessment factor scope dividing method based on shortest path Download PDFInfo
- Publication number
- CN102637227B CN102637227B CN 201210087604 CN201210087604A CN102637227B CN 102637227 B CN102637227 B CN 102637227B CN 201210087604 CN201210087604 CN 201210087604 CN 201210087604 A CN201210087604 A CN 201210087604A CN 102637227 B CN102637227 B CN 102637227B
- Authority
- CN
- China
- Prior art keywords
- grid
- evaluation
- subarea
- factor
- network node
- 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.)
- Expired - Fee Related
Links
Images
Abstract
A land resource assessment factor scope dividing method based on a shortest path includes: dispersing an assessment space range of land resource; setting up a road network dataset based on a road network of the assessment space range, wherein the road network dataset comprises network nodes and network edges; dividing the assessment space range into a plurality of assessment subareas; searching the assessment subarea of each grid in a grid set G; inquiring the network node and the assessment factor contained by each obtained assessment subarea; searching the network nodes to centralize the closest assessment factor of each network node; searching the closest assessment factor of each grid in the grid set G; and combining the grids with same closest assessment factors to obtain a corresponding assessment factor scope.
Description
Technical field
The invention belongs to Evaluation for Soil Resources factor action scope partitioning technology field, particularly relate to a kind of Evaluation for Soil Resources factor action scope division methods based on shortest path.
Background technology
(1) Evaluation for Soil Resources technology
China is as populous nation, and land resource is relatively rare, how rationally to utilize the soil, realizes that the sustainable use of land resource is current problem in the urgent need to address.The technological means of employing science and method are estimated the quality of land resource, are to promote Rational Land utilization, raising soil to utilize the necessary ways of output.
Definition according to the FAO of FAO (Food and Agriculture Organization of the United Nation), Evaluation for Soil Resources refers to process that the soil is assessed for the effect of specifically utilizing mode to show, comprise that the aspect attributes such as form to the soil, soil, vegetation, weather carry out quality comprehensive and identify, thereby distinguish and land use pattern suitability degree that evaluation objective is shown relatively.Related documents: [1] FAO.Land Evaluation.Towards a revised framework.2007..On the basis of foreign advanced technology, China has formed that the soil that comprises for farming land and construction land grades, the land valuation system that meets current national conditions demand of deciding grade and level, appraisal, appraisal of land suitability etc.
When estimating the land resource quality, the land resource quality there are the objects such as the various factors entity of appreciable impact such as business service center, country fair, iirigation water source, road network, are defined as " the Evaluation for Soil Resources factor " (referred to as " evaluation factor ").Different according to the spacial influence feature of estimating the factor, mainly can be divided into " the point-like factor ", " the wire factor " and " the planar factor ".Wherein, " the planar factor " (edaphic condition, geologic condition etc.) only limit to the coverage of this evaluation factor to the spatial dimension of land resource quality influence, and its action scope is exactly the area of space of this evaluation factor in the Evaluation for Soil Resources scope; Then there are stronger relation in " the point-like factor ", " the wire factor " with space length to the impact of land resource quality, and main manifestations constantly weakens for the impact of the estimating factor pair land resource quality increase along with space length.Therefore, for a plurality of evaluation factors in the range of value, a certain soil always is subject to the impact apart from its nearest evaluation factor.This rule spatially shows as the different evaluation factor and has different space behavior territories (reach).Can distinguish analyse by space buffer for the impact of " the wire factor " (mainly comprising Traffic Net in the Urban Land Evaluation, the agricultural land irrigation canals and ditches in estimating etc.) and obtain its action scope.Land resource quality in its buffer zone scope is affected by it, otherwise not affected by it.And to " the point-like factor " in the land resource quality assessment (school in the Urban Land Evaluation, business service center, water source during agricultural land is estimated, country fair etc.) the division relative complex then of action scope, its main cause is: the impact of point-like factor pair land resource quality often depends on certain specific network structure (Traffic Net, irrigation canals and ditches network etc.), its action intensity is not along the air line distance decay, but decays along network path distance.Therefore, when dividing its action scope, usually need to adopt the dividing mode based on path distance.What the present invention relates to designs for " the point-like factor " action scope division methods in the land resource quality assessment, and " the evaluation factor " hereinafter mentioned mainly refers to " the point-like factor ".
(2) Evaluation for Soil Resources factor action scope partitioning technology
The method of dividing about " the point-like factor " action scope both at home and abroad at present mainly can be divided into following 3 classes:
1) according to the round structure of Ke Lesi Teller " center is geographical to be said ", the principle according to equilibrium covers adopts formula
Calculate the action scope of each point-like factor, wherein D is that the radius of influence, the n of the factor are the area of range of value for number, the S that estimates the factor.The method is calculated simple, is the method that is widely used at present, also is the method for recommending in the relevant Evaluation for Soil Resources rules of Present Domestic.Related documents: [2] State Administration for Quality Supervision and Inspection and Quarantine. Urban Land Classification rules (GB/T 18507-2001) [S] .2001; Related documents: [3] Ministry of Land and Resources. Farmland Grading rules (TD/T1005-2004) [S] .2003.Yet the method does not consider to estimate the spatial distribution state of the factor, can not response factor to the feature of land resource quality influence along with the path distance decay, computational accuracy is low, factor action scope is divided unreasonable.
2) based on the evaluation factor action scope division methods of Voronoi figure.The method is divided into several Voronoi polygons according to the spatial distribution state of estimating the factor with the land valuation spatial dimension, and the scope that the Voronoi polygon that each factor generates covers is exactly the action scope that changes the factor.Related documents [4]: Hu Shiyuan, the expansion of Liu Yao woods .Voronoi figure and the application [J] in the land grading factor radius of influence is determined thereof. China Land Science, 2004 (03). the method is calculated relative complex, but its computation process has considered to estimate the space distribution of the factor, more reasonable with respect to traditional simple algorithmic approach, can the whole evaluation region of complete covering.Yet, adopt the method to estimate the division of factor action scope and still do not consider the path factor, thereby can not reflect that factor pair land resource quality influence along with the feature of path distance decay, has certain limitation.
3) based on the Voronoi drawing method of path distance.The method is take network as the basis, carry out Voronoi when dividing with network on shortest path apart from replacing conventional space line Euclidean distance.With respect to front 2 kinds of methods, the method can the reflected appraisal Effects of Factors along the feature of path attenuation, also be the highest method of accuracy in the present practical application.Related documents [5]: Zhu Guorui, Tang Xu, Wang Ping. based on the point-like deciding grade and level factor analysis [J] of effect characteristics. Wuhan University Journal (information science version) 2004 (03). related documents [6]: thank to Shunping County, Feng Xuezhi, Lu Wei. the Voronoi face territory figure based on the road network analysis makes up algorithm [J]. the mapping journal .2010.39 (01) yet., the method need to adopt grid cell, raster cell or vector grid that the land valuation scope is dispersed usually, divides on this basis the action scope of estimating the factor.Because the Evaluation for Soil Resources scope can reach tens of square kilometres usually, and in order to obtain higher computational accuracy, when it was dispersed, the size of grid cell, raster cell or vector grid was generally 50-100 rice.Therefore calculate the pixel that relates to or meshes number and can reach hundreds thousand ofly, and need to repeatedly carry out path query, its calculated amount is huge.Single evaluation factor action scope partition process for 1 medium-sized city namely needs to calculate 3-4 hour, and the Evaluation for Soil Resources process is usually directed to 10-20 the evaluation factor, and as seen its computation process is very consuming time.Therefore, must give full play to the newest fruits of computer software and hardware development, improve counting yield.
(3) parallel computing
Parallel computation (being also referred to as high-performance calculation, supercomputing) is the calculating of carrying out under parallel computing platform.Adopt the fundamental purpose of parallel computation to be to improve speed and the scale that computing machine solves problem.Related documents [7]: cypress. the data directory technical research [D] on the parallel computing platform. the .2011. of China Science ﹠ Technology University is along with the fast development of computer hardware technique, the advanced persons such as grid, polycaryon processor, cluster, desktop supercomputer, cloud computing calculate facility and successively occur, for the speed of finding the solution and the efficient that improves the large-scale calculations problem provides important technical support.In addition, the CPU on the personal computer is also day by day towards the future development of multinucleation (from double-core, four nuclears, eight nuclears to more multinuclear number development), so that the hardware cost of parallel computation is cheaper.On the other hand, the appearance of the multiple programming standards such as OpenMP, MPICH, OpenMPI, MapReduce also greatly reduces the threshold of parallel software development.Therefore, how taking full advantage of the computing power of multi-core CPU, improve the speed of finding the solution and the efficient of large-scale calculations problem, is the subject matter that faces during following program development and science are calculated.
But in Evaluation for Soil Resources factor action scope partitioning technology field, not yet there is appropriate technical solution to occur.
Summary of the invention
For the poor efficiency that exists in the existing Evaluation for Soil Resources factor action scope division methods based on shortest path, the problem such as consuming time, the present invention improves existing division methods, and design meets the spatial data decomposition strategy of concurrent operation requirement, the parallelization that realization Evaluation for Soil Resources factor action scope is divided.Fully excavate on this basis the resource potential of multi-core computer, improve the efficient of Evaluation for Soil Resources factor action scope partition process, shorten working hours.
Technical scheme provided by the invention is a kind of Evaluation for Soil Resources factor action scope division methods based on shortest path, may further comprise the steps:
Step 7.1 according to the lookup result of step 4, obtains the numbering PID that the subarea is estimated at grid g place, the subarea is estimated at grid g place be designated as evaluation subarea P;
Step 7.2 according to the Query Result of step 5, obtains to estimate network node and the evaluation factor that subarea P comprises; Comprised the evaluation factor if estimate subarea P, execution in step 7.3 does not comprise the evaluation factor if estimate subarea P, execution in step 7.4,
Step 7.3 compares grid g and estimates the air line distance of respectively estimating the factor that subarea P comprises, and records the numbering FID of the nearest evaluation factor, and grid g processing is finished;
Step 7.4, traversal is estimated the all-network node that subarea P comprises, and the distance B of the nearest evaluation factor that records in computing grid g and each network node relatively and record numbering FID with the nearest evaluation factor of grid g, finishes grid g processing; The mode of calculating during traversal is, if the network node that traverses is J, computing grid g and network node J apart from d1, then according to the lookup result of step 6 obtain the nearest evaluation factor of network node J numbering FID and with the path distance d2=Dis of the nearest evaluation factor, calculate D=d1+d2;
And in the step 6, searching the nearest evaluation factor implementation of certain network node is that all estimate the path distance of the factors in employing dijkstra's algorithm computational grid node and the evaluation space scope, therefrom obtain the nearest evaluation factor.
And the implementation of step 1 is, averages according to the minimum outsourcing rectangle in evaluation space scope place and cuts apart, and obtains some space partition zones, and each space partition zone is assigned to the grid that calculates separately create-rule on the different threads; Grid with each space partition zone merges at last, obtains the grid set G of whole evaluation space scope.
And, the implementation of step 4 is, search the evaluation subarea at each grid place among the grid set G, to estimate subarea set Ps and be divided into several evaluation subarea subsets, from grid set G, separate each and estimate the corresponding partition network grid of subarea subset collection, each is estimated the subarea subset, and execution is stacked separately to different threads with respective partition grid subset allocation, obtains the partition network grid and concentrates each grid place to estimate the numbering PID in subarea; At last each stacked result who estimates the subarea subset is merged, obtain the numbering PID in each evaluation subarea, grid place among the grid set G.
And, the implementation of step 5 is, to estimate subarea set Ps and be divided into several and estimate the subarea subset, and each be estimated subarea subset allocation to different threads, carry out separately space querying, and obtain to estimate and respectively estimate the network node that the subarea comprises in the subset of subarea and estimate the factor; At last each space querying result who estimates the subarea subset is merged, obtain estimating the network node that each evaluation subarea comprises among the subarea set Ps and estimating the factor.
And, the implementation of step 6 is, the network node ensemble average is divided into several network node subsets, and be assigned on the different threads separately the nearest evaluation factor of each network node in the Network Search node subset, then in network node, record the nearest evaluation factor numbering FID and with the path distance Dis of the nearest evaluation factor; Lookup result with each network node subset merges at last, obtains network node and concentrates the nearest evaluation factor of each network node.
And the implementation of step 7 is, grid gathered G on average be divided into several grid subsets, and be assigned on the different threads and separately the grid in the grid subset carried out substep 7.1～7.4, obtains the nearest evaluation factor of each grid in the grid subset; Execution result with each grid subset merges at last, obtains the nearest evaluation factor of each grid among the grid set G.
With respect to domestic and international existing evaluation factor action scope division methods based on shortest path, the present invention has redesigned and has estimated the flow process that factor action scope is divided according to the basic demand of new data structure and calculating parallelization.The present invention also provides the data structure that is beneficial to shortest path query, reduces dependence and correlation degree between the data, make it more appropriate to parallel processing, also can improve the efficient of shortest path query simultaneously.The present invention has also designed rational data decomposition strategy, realized the parallelization of Evaluation for Soil Resources factor action scope division methods, can give full play to the processing power of the parallel computers such as cluster of multinuclear personal computer, small-sized workstation, shared storage, increase work efficiency.Technical scheme of the present invention has simply, characteristics fast, can improve preferably the operational efficiency of point-like factor action scope partition process in the Evaluation for Soil Resources.
Description of drawings
The process flow diagram of Fig. 1 embodiment of the invention;
The discrete grid block synoptic diagram of Fig. 2 Evaluation for Soil Resources spatial dimension of the present invention;
The parallelization of the Evaluation for Soil Resources spatial dimension of Fig. 3 embodiment of the invention synoptic diagram that disperses;
The road network of Fig. 4 embodiment of the invention makes up synoptic diagram;
The evaluation space Region Segmentation synoptic diagram of Fig. 5 embodiment of the invention;
The network topology relation of Fig. 6 embodiment of the invention makes up synoptic diagram;
The evaluation subregion topology constructing synoptic diagram of Fig. 7 embodiment of the invention;
The nearest evaluation factor inquiry synoptic diagram of the network node of Fig. 8 embodiment of the invention;
The nearest evaluation factor querying method process flow diagram of the grid of Fig. 9 embodiment of the invention;
Trellis state design sketch after the nearest factor inquiry of the execution of Figure 10 embodiment of the invention;
The legend master data situation synoptic diagram of Figure 11 embodiment of the invention;
The legend range of value discretize result schematic diagram of Figure 12 embodiment of the invention;
The legend road network of Figure 13 embodiment of the invention makes up result schematic diagram;
The legend range of value of Figure 14 embodiment of the invention is cut apart and is estimated subarea structure result schematic diagram;
The legend of Figure 15 embodiment of the invention is estimated factor action scope and is divided as a result comparison diagram.
Embodiment
The division methods flow process based on the Evaluation for Soil Resources point-like factor action scope of shortest path of the present invention design is seen accompanying drawing 1, can adopt the computer software technology realization flow automatically to move by those skilled in the art.Wherein step (1)-(3) are the basic data pretreatment stage of whole flow process.Embodiment specific implementation process is as follows:
Grid adopts the vector rectangle to store.Grid is less, and then computational accuracy is higher, and calculated amount is larger; Otherwise computational accuracy is lower, and calculated amount is less.Usually when carrying out Evaluation for Soil Resources, adopt the grid of 50*50 or 100*100 rice to be dispersed in the evaluation space zone.Such as Fig. 2, the evaluation space scope of land resource is dispersed, obtain the grid set G of whole evaluation space scope.
Because number of grid can reach hundreds thousand of in this step, calculated amount is larger, calculation procedure more consuming time in the Evaluation for Soil Resources process normally, therefore the evaluation space scope can be divided into several space partition zones by coordinate, again it is assigned to the upper independent execution discrete operations of different thread (process), this data decomposition parallel processing has improved efficient, the discretize result of each thread (process) merges the most at last, just obtain the grid set G of whole evaluation space scope, such as Fig. 3: average according to the minimum outsourcing rectangle in evaluation space scope place and to cut apart, obtain some space partition zones, and each space partition zone is assigned to the grid that calculates separately create-rule on the different threads; Grid with each space partition zone merges at last, obtains the grid set G of whole evaluation space scope.Embodiment is equidistantly cut apart by the horizontal direction of minimum outsourcing rectangle, also can adopt other schemes during implementation, for example equidistantly cuts apart by vertical direction.
Embodiment according to the road network in the evaluation space scope of land resource, makes up the road network data collection in this step.The road network data collection can be used for calculating the interior any point of evaluation space scope to the shortest path distance between the evaluation factor, mainly formed by network node and network edge, stored adjacent with it node information, the length on limit, the information such as direction of network on the network edge.Network node has been stored adjacent with it network edge information.The efficient of network struction is higher, makes up for the road network of a large size city usually also to be no more than 1 minute and can to finish.So step is as the basic data preliminary work of estimating the division of factor action scope, the Interface realization that mainly provides by existing software (such as ArcGIS) is not as focal point of the present invention.Such as Fig. 4, from the road network that network edge E1, E2, E3, E4, E5 consist of, the road network data collection of structure has also comprised network node N1, N2, N3, N4, N5, N6.
Embodiment is according to the polygon of road network and evaluation space scope, the evaluation space scope of land resource is divided into several independently subregions, for the parallel basic data of carrying out of the data decomposition of subsequent step is prepared.Because this step is as the basic data preliminary work of estimating the division of factor action scope, and computation process is shorter, the Interface realization that mainly provides by existing software (such as ArcGIS) is not as focal point of the present invention.Its ultimate principle is seen Fig. 5: according to road, the evaluation space scope is cut apart.
Embodiment carries out network topology in this step and makes up.Spatial Overlap is carried out in the evaluation subarea that obtains in the discrete grid that obtains and the step 3 in the step 1, obtain the numbering that the subarea is estimated at each grid element center place.Such as Fig. 6: the subarea number of dividing based on road graticule is 1,2,3,4,5,6,7,8,9,10,11,12, stacked by subregion decomposition and executed in parallel and step 1 gained original mesh (being grid set G) merges the grid that obtains after stacked at last.
During implementation, " PID " field can be set in gridding information, this numbering be stored in " PID " field of grid.
Because step 3 has become the evaluation space Range-partition a plurality of evaluations subarea.Therefore the mode that this step also can be by data decomposition to be to estimate the subarea as unit, is assigned to respectively that different thread (process) is upper to be carried out separately.Embodiment searches the evaluation subarea at each grid place among the grid set G, to estimate subarea set Ps and be divided into several evaluation subarea subsets, from grid set G, separate each and estimate the corresponding partition network grid of subarea subset collection, each is estimated the subarea subset, and execution is stacked separately to different threads with respective partition grid subset allocation, obtains the partition network grid and concentrates each grid place to estimate the numbering PID in subarea; At last each stacked result who estimates the subarea subset is merged, obtain the numbering PID in each evaluation subarea, grid place among the grid set G.
Embodiment estimates the subarea topology constructing in this step.The evaluation subarea that step 3 is obtained respectively with step 2 in network node and estimate the factor and carry out Spatial Overlap, the network node that the query evaluation subarea comprises and estimate the factor.Such as Fig. 7, establish and estimate subarea 5, estimate subarea 8 behind parallel execution query count, estimating on the subarea 5 has network node 1,2,3,4; Estimating on the subarea 8 has network node 3,4,5,6,7,8, and contains the evaluation factor 1.
During implementation, can in the data structure of storage evaluation subarea information, " NIDs " field and " FIDs " field be set, network node be stored in " NIDs " field, estimate the factor and be stored in " FIDs " field.If it is a plurality of that the network node in certain evaluation subarea comprises, the mode that then adopts ", " to cut apart between the numbering NID of each network node forms character string and stores.If it is a plurality of that the evaluation factor also comprises, can adopt in the same way and store.
This step can adopt the paralleling tactic identical with step 4, will estimate subarea set Ps and be divided into several subsets, and each data centralization comprises one or several evaluation subareas.Be that a subset distributes a thread, space querying is carried out at thread (process) one by one in the evaluation subarea that antithetical phrase is concentrated.For example, when 200 are estimated subareas and decompose 10 thread execution, have 10 subsets with regard to one, every subset comprises 20 polygons.Embodiment will estimate subarea set Ps and be divided into several and estimate the subarea subset, each be estimated subarea subset allocation carry out separately space querying to different threads, obtain to estimate respectively to estimate the network node that the subarea comprises in the subset of subarea and estimate the factor; At last each space querying result who estimates the subarea subset is merged, obtain estimating the network node that each evaluation subarea comprises among the subarea set Ps and estimating the factor.
Embodiment inquires about the nearest evaluation factor apart from each network node, and will estimate recently at this node place record the numbering FID of the factor.During implementation, can travel through each network node, adopt existing dijkstra's algorithm (Di Jiesitela algorithm) to calculate the path distance of the current network node that traverses and all evaluation factors in the ergodic process, therefrom obtain the evaluation factor nearest apart from this network node.Such as Fig. 8, to the nearest factor query script of a certain network node on the road network of having simplified be, the factor 1,2,3 of evaluation is arranged around the network node 1, by dijkstra's algorithm, it is to estimate the factor 1 that inquiry obtains wherein nearest, and distance is 31.8.
During implementation, field " FID " and " Dis " can be set in the data structure of storage networking node information, in the attribute field " FID " of network node, record the numbering (FID) of the evaluation factor nearest apart from it, storage networking node and the path distance (Dis) of estimating the factor in " Dis " field.
For a medium-sized city, its road network node scale is usually about 1000, and the process of the nearest evaluation factor inquiry of each network node is a relatively independent process, is fit to very much parallel processing.Therefore can adopt the parallel method of data decomposition, the network node collection is resolved into some subsets, and be assigned to the upper executed in parallel of different threads (process).This step needn't be divided according to estimating the subarea.When embodiment searches the network node ensemble average is divided into several network node subsets, and be assigned on the different threads separately the nearest evaluation factor of each network node in the Network Search node subset, then in network node, record the nearest evaluation factor numbering FID and with the path distance Dis of the nearest evaluation factor; Lookup result with each network node subset merges at last, obtains network node and concentrates the nearest evaluation factor of each network node.
Embodiment travels through grid set G.If the arbitrary grid that traverses is designated as g, carry out following substep:
Step 7.1 according to the lookup result of step 4, obtains the numbering PID that the subarea is estimated at grid g place, the subarea is estimated at grid g place be designated as evaluation subarea P;
Step 7.2 according to the Query Result of step 5, obtains to estimate network node and the evaluation factor that subarea P comprises; Comprised the evaluation factor if estimate subarea P, execution in step 7.3 does not comprise the evaluation factor if estimate subarea P, execution in step 7.4,
Step 7.3 compares grid g and estimates the air line distance of respectively estimating the factor that subarea P comprises, and records the numbering FID of the nearest evaluation factor, and grid g processing is finished;
Step 7.4, traversal is estimated the all-network node that subarea P comprises, and the distance B of the nearest evaluation factor that records in computing grid g and each network node relatively and record numbering FID with the nearest evaluation factor of grid g, finishes grid g processing; The mode of calculating during traversal is, if the network node that traverses is J, computing grid g and network node J apart from d1, then according to the lookup result of step 6 obtain the nearest evaluation factor of network node J numbering FID and with the path distance d2=Dis of the nearest evaluation factor, calculate D=d1+d2.
During implementation, field " FID " and " Dis " can be set in the data structure of save mesh, the numbering FID that estimates recently the factor is stored in the attribute field " FID " of this grid, and with grid and estimate recently distance between the factor be stored in " Dis ' in the field.When searching the nearest evaluation factor of each grid among the grid set G, be stored in the information in the data structure before also can utilizing, corresponding flow process such as Fig. 9:
The current grid g that traverses is begun to process, the grid numbering GID that provides this grid given is provided;
Obtain the evaluation subarea P at grid g place according to " PID " property value of grid g;
According to estimating subarea P's " NIDs " and " FIDs " property value, obtain to estimate network node collection N and the evaluation factor set F that subarea P comprises;
Judge whether estimate factor set F is empty; Be then to carry out 2., otherwise carry out 1..
1. for respectively estimating the air line distance of the factor among computing grid g and the evaluation factor set F, take out the evaluation factor nearest with grid g, be designated as f.At last grid g and the air line distance of estimating factor f are deposited in grid " Dis " field, the FID property value of estimating factor f is stored into the g of grid " FID " field.As among Fig. 9 for example, estimating the subarea has network node 3,4,5,7..., the factor 1 of evaluation is wherein arranged, grid 1 is nearest with the air line distance of estimating the factor 1, deposits grid 1 in " Dis " field.
2. respectively the air line distance d1 between each network node among computing grid g and the network node collection N, obtain each network node and it estimates the distance B=d1+d2 apart from d2, the computing grid g Distance evaluation factor of the factor recently.Therefrom find out the network node n on the shortest path, with network node n's " FID " the numbering FID of the nearest evaluation factor of storing in the attribute is as the nearest evaluation factor f of grid g.Deposit grid g in grid g with the distance of estimating factor f at last " Dis " field, the numbering FID that estimates factor f is stored into grid g's " FID " field.As giving an example among Fig. 9, estimate the subarea network node 1 is arranged, 2,3,4, wherein do not estimate the factor, grid is to network node 1,2,3,4 distance is designated as respectively d1 (1), d1 (2), d1 (3), d1 (4), network node 1 is designated as d2 (1) to the distance of the evaluation factor nearest with it, network node 2 is designated as d2 (2) to the distance of the evaluation factor nearest with it, network node 3 is designated as d2 (3) to the distance of the evaluation factor nearest with it, network node 4 is designated as d2 (4) to the distance of the evaluation factor nearest with it, find out d1 (1)+d2 (1), d1 (2)+d2 (2), d1 (3)+d2 (3), reckling among d1 (4)+d2 (4) also deposits grid in " Dis " field, the corresponding network node " FID " factor numbering of storing in the attribute is as current nearest object f.
Calculate effect and see Figure 10: establishing has the factor 1,2 of evaluation in the road network under the original state, through after the above query processing, in the trellis state that calculates, the nearest evaluation factor is all stored numbering 1 for the grid of estimating the factor 1, and the nearest evaluation factor is all stored numbering 2 for the grid of estimating the factor 2.
Because number of grid can reach hundreds thousand of usually, serial is calculated one by one shortest path and is carried out path distance relatively according to the conventional method, efficient is very low, and the time that this computation process consumes in Evaluation for Soil Resources factor action scope partition process can account for more than 80% of whole computation process usually.Because it also is relatively independent that each grid is searched the process of the nearest evaluation factor, therefore can be divided into several subsets to grid being gathered G equally, and be assigned on the different threads and carry out separately above substep and decompose, to improve counting yield, shorten computing time.This step needn't be divided according to estimating the subarea.Embodiment gathers G with grid and on average is divided into several grid subsets, and is assigned on the different threads separately the grid in the grid subset is carried out substep 7.1～7.4, obtains the nearest evaluation factor of each grid in the grid subset; Execution result with each grid subset merges at last, obtains the nearest evaluation factor of each grid among the grid set G.
Embodiment inquires about according to the numbering FID of the nearest evaluation factor of each grid among the step 7 gained grid set G, and the grid that the FID property value is identical merges, and can respectively be estimated factor action scope polygon, and whole computation process finishes.Because during implementation, the inquiry in this step can be inquired about based on relational database fully, thereby efficient is higher, can finish so need not to carry out the parallelization operation.
The general data object that relates among the present invention has road network data collection (comprising network node and network edge), grid, evaluation subarea.For the sake of ease of implementation, the data structure of the present invention's suggestion is: 1) the road network data collection is expanded the data structure of network node on this basis take Network Dataset (Network data set) data model of ESRI (company of U.S. environment system research institute) as the basis; 2) mesh object is to the discrete rectangular node that obtains of Evaluation for Soil Resources spatial dimension according to fixing distance; 3) Evaluation for Soil Resources subarea makes up the polygon of sealing according to Evaluation for Soil Resources scope and road network, thereby whole evaluation region spatially is divided into several independently subareas, thereby prepares for the parallel data decomposition of carrying out of algorithm.
Embodiment has selected respectively accordingly the simplest decomposing scheme for step that can executed in parallel, also can adopt other decomposing schemes during implementation.The parallel number of concrete number, the active thread that decomposes subset and actual operation ability etc. are relevant, and the user also can arrange as the case may be.
In order to improve flow performing efficient, to reduce the dependence between the data, after the embodiment of the invention is expanded the data structure of 3 class objects (network node, grid, evaluation subarea), as shown in the table:
For the purpose of explanation invention effect, the present invention with the action scope partition process of the point-like factor " country fair " in certain urban land resource evaluation as legend, by writing computer program and finishing parallel computation at the personal computer of one 4 nuclear CPU.Its basic data has: traffic route network, the position, country fair in Evaluation for Soil Resources scope, evaluation district.As shown in figure 11.Carry out successively according to process step provided by the present invention:
(1) according to the Evaluation for Soil Resources scope, big or small generating mesh according to 100*100 rice, and generate corresponding attribute field structure, the results are shown in Figure 12: the attribute field structure comprises GID, Shape, PID, FID, Dis, the GID field is used for the numbering of save mesh, the Shape field is for the X of the geometric object of having stored grid, Y coordinate.Geometric object is Polygon (polygon) model of ESRI.
(2) adopt the Network dataset Network data set model construction road network data collection of ESRI, and according to design of the present invention, the data structure of network node is expanded, add " FID " " Dis " field; The results are shown in Figure 13: the attribute field structure comprises NID, SHAPE, PID, FID, Dis, the numbering ID of NID field store current network node, and the Shape field is used for the X of storage networking node, the Y coordinate, point adopts point (Point) model of ESRI.
(3) carry out area of space according to road network and range of value and cut apart, and according to design of the present invention, add " FIDs " " NIDs " field; The results are shown in Figure 14: the attribute field structure comprises PID, SHAPE, FID, NIDs, FIDs, and the PID field store is estimated the subarea numbering, and the X of subarea all-network node, Y coordinate are estimated in the SHAPE storage.
(4) will estimate subarea set Ps and be divided into several evaluation subarea subsets, grid will be carried out data decomposition according to the evaluation subarea at its place, and be assigned to separately execution on the different threads, the PID numbering of estimating the subarea will be assigned to the mesh object that it comprises.
(5) will estimate subarea set Ps and be divided into several evaluation subarea subsets, and be assigned to separately execution on the different threads, and calculate respectively and respectively estimate the network node that comprises in the subarea and estimate the factor.
(6) the network node collection is decomposed, and be assigned to separately execution on the different threads, calculate the evaluation factor nearest apart from each network node.Wherein the search of shortest path adopts dijkstra's algorithm to realize.
(7) set is decomposed to grid, is assigned on the different threads and carries out separately.According to grid and estimate topological relation between subarea, network node, the evaluation factor, search the evaluation factor the shortest with its path distance, and record its FID.
(8) " FID " according to mesh object inquires about, and the mesh object that will have identical FID property value merges, and finally obtains the evaluation factor action scope based on the shortest path distance.The results are shown among Figure 15 shown in (a) part.
Wherein (4), (5) all are to decompose according to estimating the subarea.But the object that calculates is different, and the object data set of decomposition is different, and the foundation of decomposition all is the evaluation subarea according to the object place: what (4) operated is mesh object, and what (5) operated is to estimate the subarea.
Adopt prior art to divide the result shown in (b) part among Figure 15 based on the evaluation factor of air line distance.Divide the result with respect to the evaluation factor based on air line distance, result obtained by the method for the present invention obviously meets real rule more.And through the parallel efficiency test, the method for the present invention's design is with respect to traditional serial processing method, and parallel efficiency can reach 70-80%, can amplitude shorten computing time under multi-core environment, increases work efficiency.
Specific embodiment described herein only is to the explanation for example of the present invention's spirit.Those skilled in the art can make various modifications or replenish or adopt similar mode to substitute described specific embodiment, but can't depart from spirit of the present invention or surmount the defined scope of appended claims.
Claims (2)
1. the Evaluation for Soil Resources factor action scope division methods based on shortest path is characterized in that, may further comprise the steps:
Step 1 disperses to the evaluation space scope of land resource, obtains the grid set G of whole evaluation space scope; The implementation of step 1 is, averages according to the minimum outsourcing rectangle in evaluation space scope place and cuts apart, and obtains some space partition zones, and each space partition zone is assigned to the grid that calculates separately create-rule on the different threads; Grid with each space partition zone merges at last, obtains the grid set G of whole evaluation space scope;
Step 2, based on the road network structure road network data collection of evaluation space scope, the road network data collection comprises network node and network edge, road network data concentrates the all-network node to consist of the network node collection;
Step 3 based on evaluation space scope and road network, is divided into several with the evaluation space scope and estimates the subarea, and the evaluation subarea set of formation is designated as Ps;
Step 4 is searched the evaluation subarea at each grid place among the grid set G, comprises grid set G and estimates subarea set Ps and carry out stackedly, obtains the numbering PID that the subarea is estimated at each grid place; The implementation of step 4 is, search the evaluation subarea at each grid place among the grid set G, to estimate subarea set Ps and be divided into several evaluation subarea subsets, from grid set G, separate each and estimate the corresponding partition network grid of subarea subset collection, each is estimated the subarea subset, and execution is stacked separately to different threads with respective partition grid subset allocation, obtains the partition network grid and concentrates each grid place to estimate the numbering PID in subarea; At last each stacked result who estimates the subarea subset is merged, obtain the numbering PID in each evaluation subarea, grid place among the grid set G;
Step 5 is carried out respectively space querying for each evaluation subarea among the evaluation subarea set Ps, obtains respectively to estimate network node and the evaluation factor that the subarea comprises; The implementation of step 5 is, to estimate subarea set Ps and be divided into several evaluation subarea subsets, each is estimated subarea subset allocation to different threads, carry out separately space querying, obtain to estimate and respectively estimate the network node that the subarea comprises in the subset of subarea and estimate the factor; At last each space querying result who estimates the subarea subset is merged, obtain estimating the network node that each evaluation subarea comprises among the subarea set Ps and estimating the factor;
Step 6, the nearest evaluation factor of each network node in the Network Search nodal set, then in network node, record the nearest evaluation factor numbering FID and with the path distance Dis of the nearest evaluation factor; The implementation of step 6 is, the network node ensemble average is divided into several network node subsets, and be assigned on the different threads separately the nearest evaluation factor of each network node in the Network Search node subset, then in network node, record the nearest evaluation factor numbering FID and with the path distance Dis of the nearest evaluation factor; Lookup result with each network node subset merges at last, obtains network node and concentrates the nearest evaluation factor of each network node;
Step 7 is searched the nearest evaluation factor of each grid among the grid set G, comprises traversal grid set G, for the following substep of any grid g execution wherein,
Step 7.1 according to the lookup result of step 4, obtains the numbering PID that the subarea is estimated at grid g place, the subarea is estimated at grid g place be designated as evaluation subarea P;
Step 7.2 according to the Query Result of step 5, obtains to estimate network node and the evaluation factor that subarea P comprises; Comprised the evaluation factor if estimate subarea P, execution in step 7.3 does not comprise the evaluation factor if estimate subarea P, execution in step 7.4,
Step 7.3 compares grid g and estimates the air line distance of respectively estimating the factor that subarea P comprises, and records the numbering FID of the nearest evaluation factor, and grid g processing is finished;
Step 7.4, traversal is estimated the all-network node that subarea P comprises, and the distance B of the nearest evaluation factor that records in computing grid g and each network node relatively and record numbering FID with the nearest evaluation factor of grid g, finishes grid g processing; The mode of calculating during traversal is, if the network node that traverses is J, computing grid g and network node J apart from d1, then according to the lookup result of step 6 obtain the nearest evaluation factor of network node J numbering FID and with the path distance d2=Dis of the nearest evaluation factor, calculate D=d1+d2;
The implementation of step 7 is, grid gathered G on average be divided into several grid subsets, and be assigned on the different threads and separately the grid in the grid subset carried out substep 7.1～7.4, obtains the nearest evaluation factor of each grid in the grid subset; Execution result with each grid subset merges at last, obtains the nearest evaluation factor of each grid among the grid set G;
Step 8, according to the nearest evaluation factor of each grid among the step 7 gained grid set G, the grid that the nearest evaluation factor is identical merges, and is estimated accordingly factor action scope.
2. as claimed in claim 1 based on the Evaluation for Soil Resources factor action scope division methods of shortest path, it is characterized in that: in the step 6, searching the nearest evaluation factor implementation of certain network node is, adopt the path distance of all evaluation factors in dijkstra's algorithm computational grid node and the evaluation space scope, therefrom obtain the nearest evaluation factor.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201210087604 CN102637227B (en) | 2012-03-29 | 2012-03-29 | Land resource assessment factor scope dividing method based on shortest path |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201210087604 CN102637227B (en) | 2012-03-29 | 2012-03-29 | Land resource assessment factor scope dividing method based on shortest path |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102637227A CN102637227A (en) | 2012-08-15 |
| CN102637227B true CN102637227B (en) | 2013-10-23 |
Family
ID=46621621
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 201210087604 Expired - Fee Related CN102637227B (en) | 2012-03-29 | 2012-03-29 | Land resource assessment factor scope dividing method based on shortest path |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102637227B (en) |
Families Citing this family (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103336825A (en) * | 2013-07-04 | 2013-10-02 | 上海交通大学 | Violence searching method and system for obtaining single reverse furthest neighbor in road network |
| CN104200045B (en) * | 2014-09-17 | 2016-01-13 | 武汉大学 | The parallel calculating method of a kind of basin large scale water system sediments formula hydrodynamic model |
| CN107978219B (en) * | 2016-10-25 | 2020-05-01 | 武汉四维图新科技有限公司 | Method and device for constructing road network of digital map |
| CN112199368B (en) * | 2017-05-26 | 2022-06-03 | 中国重汽集团福建海西汽车有限公司 | Body-in-white quality management method and system |
| CN113096425B (en) * | 2021-03-29 | 2021-12-31 | 紫清智行科技(北京)有限公司 | Dispatching method and system for automatic driving patrol car applied to large scene |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5561790A (en) * | 1992-03-24 | 1996-10-01 | International Business Machines Corporation | Shortest path determination processes for use in modeling systems and communications networks |
| US7571411B2 (en) * | 2006-01-12 | 2009-08-04 | International Business Machines Corporation | Methods and apparatus for providing flexible timing-driven routing trees |
| CN101751491A (en) * | 2008-11-28 | 2010-06-23 | 上海电机学院 | Searching method of fuzzy shortest path |
| CN101957876A (en) * | 2010-09-15 | 2011-01-26 | 清华大学 | Multilayer wiring method based on uneven grids in consideration of through holes |
| CN102054355A (en) * | 2011-01-07 | 2011-05-11 | 同济大学 | Virtual vehicle routing method applicable to large-scale traffic flow simulation |
-
2012
- 2012-03-29 CN CN 201210087604 patent/CN102637227B/en not_active Expired - Fee Related
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5561790A (en) * | 1992-03-24 | 1996-10-01 | International Business Machines Corporation | Shortest path determination processes for use in modeling systems and communications networks |
| US7571411B2 (en) * | 2006-01-12 | 2009-08-04 | International Business Machines Corporation | Methods and apparatus for providing flexible timing-driven routing trees |
| CN101751491A (en) * | 2008-11-28 | 2010-06-23 | 上海电机学院 | Searching method of fuzzy shortest path |
| CN101957876A (en) * | 2010-09-15 | 2011-01-26 | 清华大学 | Multilayer wiring method based on uneven grids in consideration of through holes |
| CN102054355A (en) * | 2011-01-07 | 2011-05-11 | 同济大学 | Virtual vehicle routing method applicable to large-scale traffic flow simulation |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102637227A (en) | 2012-08-15 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Fleischmann et al. | Morphological tessellation as a way of partitioning space: Improving consistency in urban morphology at the plot scale | |
| Huang et al. | Explorations of the implementation of a parallel IDW interpolation algorithm in a Linux cluster-based parallel GIS | |
| CN102637227B (en) | Land resource assessment factor scope dividing method based on shortest path | |
| CN109146204A (en) | A kind of wind power plant booster stations automatic addressing method of comprehensiveestimation | |
| CN107180450A (en) | A kind of algorithm of the river valley transverse shape based on DEM | |
| Verhagen et al. | The long and winding road: combining least cost paths and network analysis techniques for settlement location analysis and predictive modelling | |
| KR101675778B1 (en) | Decision making system corresponding to volcanic disaster | |
| CN110781267A (en) | Multi-scale space analysis and evaluation method and system based on geographical national conditions | |
| CN108364331A (en) | A kind of isopleth generation method, system and storage medium | |
| CN104820684A (en) | Fast on-line analysis processing method based on spatial position | |
| Song et al. | Parallel viewshed analysis on a PC cluster system using triple-based irregular partition scheme | |
| Ghosh et al. | Traj-cloud: a trajectory cloud for enabling efficient mobility services | |
| CN105138607B (en) | A kind of KNN querying methods based on combination grain distributed memory grid index | |
| CN106372127B (en) | The diversity figure sort method of large-scale graph data based on Spark | |
| He et al. | Footprint-based generalization of 3D building groups at medium level of detail for multi-scale urban visualization | |
| Hu et al. | An effective selecting approach for social media big data analysis—Taking commercial hotspot exploration with Weibo check-in data as an example | |
| Huang et al. | Grid interpolation algorithm based on nearest neighbor fast search | |
| Cao et al. | Research on searching algorithms for unstructured grid remapping based on KD tree | |
| Qiu et al. | Inferring road maps from sparsely-sampled GPS traces | |
| Xing et al. | Continuous monitoring of nearest neighbors on land surface | |
| CN102693161B (en) | Concurrent land resource quality evaluation factor space quantifying method | |
| García-García et al. | Efficient distributed algorithms for distance join queries in spark-based spatial analytics systems | |
| Zhang et al. | Study of Huizhou architecture component point cloud in surface reconstruction | |
| Guo et al. | A mapreduce algorithm for polygon retrieval in geospatial analysis | |
| WO2019024345A1 (en) | Uniform-speed search expansion algorithm based on trajectory query with sequential interest regions |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20131023 Termination date: 20190329 |