CN104750707A - Decision support unit, land type identifying and verifying system - Google Patents
Decision support unit, land type identifying and verifying system Download PDFInfo
- Publication number
- CN104750707A CN104750707A CN201310740391.6A CN201310740391A CN104750707A CN 104750707 A CN104750707 A CN 104750707A CN 201310740391 A CN201310740391 A CN 201310740391A CN 104750707 A CN104750707 A CN 104750707A
- Authority
- CN
- China
- Prior art keywords
- remotely
- sensed data
- data
- land type
- attribute
- 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.)
- Pending
Links
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to the technical field of remote sensing data identification, and provides a building method for a decision support unit. The method comprises the following steps: S2, obtaining remote sensing data of the identified land type, wherein the remote sensing data comprises a plurality of attribute data; S4, selecting at least two attribute data as the characteristic attribute data for clustering, and dividing the remote sensing data to several class clusters; S6, rebuilding the remote sensing data in each class cluster on the basis of the land type and the characteristic attribute data; S8, performing the association rule discovery to the rebuilt remote sensing data, and obtaining the association result; S 10, storing the association result of which the association probability is greater than the probability threshold T and forming the decision support unit. The decision support unit obtained by the method has many association rules with higher relevancy, and can be used as the reference for the unknown land type with higher relevancy, and the decision is supported.
Description
Technical field
The present invention relates to remotely-sensed data recognition technology field, be specifically related to decision support unit, land type identification and verification system.
Background technology
Various two-dimensional imaging technique is all the center of many technical fields now.Such as, in environmental analysis and in city planning, satellite image is used for mapping as original image or with treated form.Usually, before use comprises the image of such information, need to carry out decipher (i.e. identification, such as image type identification etc.) to it.The object of this decipher normally identifies, the structure in image becomes, object in such as background and so on; Delimit the different sections to another or a large amount of different colourity, structure or mutual all different boundaries in other respects; Some examples have: the road in identification satellite image, house, forest land, meadow, arable land etc.At present, this decipher work mainly carries out simple human interpretation based on the experience of professional and limited image data information, and (namely professional is by the analysis of the every attribute of image and understanding, identify the type of its correspondence), the workload of decipher is large, inefficiency, with high costs, and very high to decipher personnel requirement.This decipher is operated in can not satisfy the demand today that quantity of information increases completely, so, a kind of fast, the decipher of low cost or land type recognition methods become a kind of active demand in this field.
Summary of the invention
In view of this, the present invention solves that land type identification workload in prior art is large, the technical matters of inefficiency, provides following method, unit, module and system.
The construction method of a kind of decision support unit that the embodiment of the present invention provides, wherein, comprises the steps:
S2, obtain and identified the remotely-sensed data of land type, described remotely-sensed data comprises some attribute datas;
S4, choose at least two described attribute datas and carry out cluster as characteristic attribute data, by described remotely-sensed data divide into several classes bunch;
S6, the remotely-sensed data in each class bunch to be reconstructed based on described land type and characteristic attribute data;
S8, to reconstruct after remotely-sensed data carry out associated rule discovery, obtain association results;
The association results that S10, storage association probability are greater than probability threshold value T forms decision support unit.
Further, described step S6 comprises the steps:
S62, select one or more in described characteristic attribute data as reconfiguration attribute data;
S64, every part of remotely-sensed data carried out to entry and indicate and set up, the entry land type of every part of remotely-sensed data and reconfiguration attribute data thereof being constructed as simultaneously this part of remotely-sensed data indicates;
S66, the described entry in each class bunch indicated and carries out permutation and combination, and add up the quantity of each permutation and combination.
Further, described step S62 also comprises:
To select characteristic attribute data in one or more carry out classification process, and using classification process after result substitute the characteristic attribute data of described selection as reconfiguration attribute data.
Further, probability threshold value T is more than or equal to 65%.
Further, described characteristic attribute data to comprise in land area, centre coordinate and soil girth one or more.
Further, described cluster adopts the one in DBSCAN algorithm, K-means algorithm and KNN algorithm.
Further, described associated rule discovery adopts Apriori algorithm.
A kind of decision support unit that the embodiment of the present invention provides, wherein, adopts above-mentioned construction method to build and obtains.
A kind of land type recognition methods that the embodiment of the present invention provides, wherein, comprises the steps:
S100, above-mentioned construction method is adopted to obtain decision support unit;
S120, obtain the attribute data of remotely-sensed data to be identified and at least one land type that identify remotely-sensed data adjacent with described remotely-sensed data to be identified and attribute data thereof;
S130, the attribute data of described remotely-sensed data to be identified and described adjacent identification remotely-sensed data to be handled as follows:
S132, obtain described remotely-sensed data to be identified and the described adjacent reconfiguration attribute data identifying remotely-sensed data;
S134, described adjacent identification remotely-sensed data carried out to entry and indicate and set up, the described adjacent land type identifying remotely-sensed data and reconfiguration attribute data thereof are constructed as the described adjacent entry identifying remotely-sensed data simultaneously and indicate;
S140, in conjunction with describedly adjacent identifying that the entry of remotely-sensed data indicates, the reconfiguration attribute data of described remotely-sensed data to be identified and described decision support unit, provide the land type of described remotely-sensed data to be identified.
A kind of land type identification module that the embodiment of the present invention provides, wherein, comprising:
Adopt the decision support unit that above-mentioned construction method obtains, for providing decision support;
First acquiring unit, for obtaining the attribute data of remotely-sensed data to be identified, and at least one land type that identify remotely-sensed data adjacent with described remotely-sensed data to be identified and attribute data thereof;
First attribute processing unit, for obtaining described remotely-sensed data to be identified and the described adjacent reconfiguration attribute data identifying remotely-sensed data; And the establishment of entry sign is carried out to described adjacent identification remotely-sensed data, the described adjacent land type identifying remotely-sensed data and reconfiguration attribute data thereof are constructed as the described adjacent entry identifying remotely-sensed data simultaneously and indicate;
First processing unit, in conjunction with the described adjacent reconfiguration attribute data identifying the entry sign of remotely-sensed data, described remotely-sensed data to be identified and described decision support unit, provides the land type of described remotely-sensed data to be identified.
A kind of land type identification and evaluation method that the embodiment of the present invention provides, wherein, comprises the steps:
S200, above-mentioned construction method is adopted to obtain decision support unit;
S220, obtain the remotely-sensed data of remotely-sensed data to be evaluated and described remotely-sensed data periphery to be evaluated;
S240, the remotely-sensed data of remotely-sensed data to be evaluated and remotely-sensed data periphery to be evaluated carried out to entry and indicate and set up;
S260, obtain P(c, A according to decision support unit and remotely-sensed data to be evaluated), this A is remotely-sensed data to be evaluated, c is the numbering of A periphery remotely-sensed data, P(c, A) represent the land type of remotely-sensed data based on being numbered c, the probable value of the land type of A remotely-sensed data;
S280, according to formula:
obtain the land type identification and evaluation value of remotely-sensed data to be evaluated, wherein, Q
arepresent the evaluation of estimate of the land type identification of A remotely-sensed data, n is A periphery remotely-sensed data number summation.
A kind of land type identification and evaluation module that the embodiment of the present invention provides, wherein, comprising:
Adopt the decision support unit that above-mentioned construction method obtains, for providing decision support;
Second acquisition unit, for obtaining the remotely-sensed data of remotely-sensed data to be evaluated and described remotely-sensed data periphery to be evaluated;
Second attribute processing unit, indicates establishment for carrying out entry to the remotely-sensed data of remotely-sensed data to be evaluated and remotely-sensed data periphery to be evaluated;
Probability calculation unit, obtains P(c, A according to decision support unit and remotely-sensed data to be evaluated), this A is remotely-sensed data to be evaluated, and c is the numbering of A periphery remotely-sensed data, P(c, A) land type based on the remotely-sensed data being numbered c is represented, the probable value of the land type of A remotely-sensed data;
Evaluate computing unit: according to formula:
obtain the land type identification and evaluation value of remotely-sensed data to be evaluated, wherein, Q
arepresent the evaluation of estimate of the land type identification of A remotely-sensed data, n is A periphery remotely-sensed data number summation.
A kind of land type verification system that the embodiment of the present invention provides, wherein, comprises above-mentioned land type identification module and/or above-mentioned land type identification and evaluation module.
The land type identification module of the embodiment of the present invention provides multiple land type alternatively or land type, the identification of supervision land type can have been realized, user only need simple judge just can comparatively fast, accuracy rate completes land type identification work higher, reduce the workload identifying land type, this reduces the skill set requirements of land type identification worker.
Accompanying drawing explanation
By the detailed description of carrying out below in conjunction with accompanying drawing, object of the present invention and feature will become apparent, wherein:
Fig. 1 is your construction method process flow diagram of decision support unit of the embodiment of the present invention;
Fig. 2 is the cluster process flow diagram of the embodiment of the present invention;
Fig. 3 is the reconstructing method process flow diagram of the embodiment of the present invention;
Fig. 4 is the reconstruct example schematic of the embodiment of the present invention;
Fig. 5 is the land type recognition methods process flow diagram of the embodiment of the present invention;
Fig. 6 is the structural representation of the land type identification module of the embodiment of the present invention;
Fig. 7 is the land type identification and evaluation method flow diagram of the embodiment of the present invention;
Fig. 8 is the land type identification and evaluation module diagram of the embodiment of the present invention.
Embodiment
Be described in detail embodiments of the invention now, its sample table shows in the accompanying drawings, and wherein, identical label represents identical parts all the time.Be described to explain the present invention to embodiment below with reference to the accompanying drawings.In the following description, obscuring of the present invention's design that the unnecessary detailed description in order to avoid known features and/or function causes, may omit the detailed description do not wanted of known features and/or function.
Described in inside background technology, soil satellite remote sensing picture to be understood or decipher is common occurrence feelings, and in soil satellite remote sensing picture, often comprise each soil remote sensing images of be connected to each other (adjoining).A soil remote sensing images periphery has multiple soils remote sensing images and adjoins with it.Therefore, being separated the soil remote sensing images in soil satellite remote sensing picture, identifying is the groundwork of decipher.Each soil remote sensing images can have land type data and a lot of attribute data, are referred to as remotely-sensed data or soil remotely-sensed data.Carrying out identification to each soil remote sensing images needs experienced expert to be competent at, and generally comprises attribute data identification and land type identification (i.e. land type data identification).The attribute data of soil remote sensing images generally has: soil label, land area, soil centre coordinate and soil girth etc., and these attribute data ratios are easier to obtain.And the land type of soil remote sensing images, in the prior art, being only needs experienced expert to carry out identifying, land type identification workload is large, inefficiency.Core of the present invention is by building decision support unit, and the land type of remotely-sensed data to be identified is provided by attribute data (i.e. the attribute data of the remotely-sensed data to be identified) acting in conjunction of decision support unit and soil to be identified remote sensing images, alleviate the workload of land type identification, improve the efficiency of land type identification.
Embodiment one
The present embodiment provides a kind of decision support cell formation method.Fig. 1 is the decision support cell formation method flow diagram of the embodiment of the present invention.Please refer to Fig. 1, the construction method of this decision support unit, comprises the steps:
S2, obtain and identified the remotely-sensed data of land type, described remotely-sensed data comprises some attribute datas.
Soil remote sensing images have a remotely-sensed data, and remotely-sensed data and soil remote sensing images are one to one.A soil satellite remote sensing picture comprises some soils remote sensing images.Generally, this has identified that the remotely-sensed data of land type needs for many parts, and to meet below needed for cluster, association analysis, concrete volume cost those skilled in the art can prepare as required voluntarily.Portion has identified that the remotely-sensed data (being also called for short " remotely-sensed data identified " or " identifying remotely-sensed data " herein) of land type comprises a land type and several attribute datas.The a remotely-sensed data identified has its oneself clause name, is generally its land type.This remotely-sensed data identified is the basis of this decision support unit, is also the basis of carrying out the accuracy of decision-making and evaluation based on this decision support unit.Therefore, choosing of this remotely-sensed data identified is very important, preferably through the remotely-sensed data identifying land type that authoritative expert proves repeatedly.Land type generally has: meadow, forest land, field, arable land, construction land, water are watered, Gobi desert, reservoir swag, alkaline land etc. type, are main targets of the present invention to the identification of these land types.This attribute data has: the girth in the area in soil, the central point (such as, dimension and longitude) in soil and soil etc.
S4, choose at least two described attribute datas and carry out cluster as characteristic attribute data, by described remotely-sensed data divide into several classes bunch.
The object of this step is that the remotely-sensed data feature based attribute data identified many parts is divided into some inhomogeneities bunch.Remotely-sensed data in each class bunch has certain contact, and this contact of data between class bunch is just fewer, processes like this and is convenient to associated rule discovery.The method of this cluster has a variety of, and the preferred clustering method of the embodiment of the present invention has DBSCAN algorithm, K-means algorithm and KNN algorithm etc.
Be described for DBSCAN algorithm below.Fig. 2 is the cluster process flow diagram of the embodiment of the present invention.The clustering method of the present embodiment is described in detail below in conjunction with Fig. 2.
This clustering method comprises the steps:
Step S42, chooses at least two attribute datas in described some attribute datas as characteristic attribute data.These characteristic attribute data should get rid of land type, also should get rid of irrelevant attribute, such as soil label etc.Therefore, this step has the function getting rid of uncorrelated attribute data, reduces the interference of uncorrelated attribute data to subsequent step, improves recognition accuracy or evaluates accuracy rate.These characteristic attribute data are the attribute data relevant to land type, are generally land area, soil centre coordinate and soil girth etc.These characteristic attribute data to comprise in land area, centre coordinate and soil girth one or more.
Step S44, setting cluster device parameter, so that the work of cluster device.This parameter has radius (Eps) to be at least 6 for 0-1 and threshold value (MinPts).This cluster device can revise this cluster device parameter according to cluster result, to make cluster result optimum.
Step S46, cluster device carry out the computing of DBSCAN clustering algorithm to every part of this remotely-sensed data identified and export cluster result.
Step S48, described DBSCAN clustering algorithm operation result to be evaluated.Such as, judge whether the class number of clusters amount after cluster meets the demands.If meet the demands, then pass through, perform step S50; Otherwise, return step S44 and reset parameter, namely change Eps, MinPts value.
Step S50, export some classes bunch, such as output class bunch 1, class bunches 2, class bunches 3 etc.May 300 parts of remotely-sensed datas identified be contained in class bunch 1, in class bunches 2, may 100 parts of remotely-sensed datas identified be contained, may 100 parts of remotely-sensed datas identified be contained in class bunches 3, the remotely-sensed data that M part has identified in class bunch N, may be comprised.
Cluster is the one in data mining technology, and those skilled in the art are familiar, is not described in detail it realizes cluster or cluster device at this.
Certainly, adopt K-means cluster device, the scope of this clustering parameter K is 5 to 20.
S6, the remotely-sensed data in each class bunch to be reconstructed based on described land type and characteristic attribute data.
This step is core procedure of the present invention, can improve the correlativity of identification, is also the basis of associated rule discovery.The clause name of the remotely-sensed data identified in each class bunch is its land type, does not comprise property parameters, be unfavorable for identification and the process of processing unit in this clause name.This step carries out secondary processing and process to remotely-sensed data just, makes it to be formed the effective data having and identify meaning.Certainly, the reconstruct in this step is never simply carry out rename, but undergoes technological transformation for limited data, makes it the object reaching valid data.
The following detailed description of the reconstructing method of the present embodiment.Fig. 3 is the reconstructing method process flow diagram of the embodiment of the present invention.Please refer to Fig. 3, comprising the steps: of this reconstructing method
S62, select one or more in described characteristic attribute data as reconfiguration attribute data.In certain embodiments, this step needs to carry out classification process to one or more in characteristic attribute data further, and the result after classification processes is substituted these characteristic attribute data as reconfiguration attribute data.Such as, area value (i.e. area attribute data) in remotely-sensed data is divided into five classes, is respectively more than I(10000 sq-km), more than II(7000 ~ 10000 sq-km), more than III(4000 ~ 7000 sq-km), more than IV(1000 ~ 4000 sq-km), below V(1000 sq-km).If the area attribute data of A remotely-sensed data is 3000 sq-kms, the area attribute data of D remotely-sensed data is 4005 sq-kms, the area attribute data of F remotely-sensed data is 800 sq-kms, so, the reconfiguration attribute data (area) of A remotely-sensed data are IV, the reconfiguration attribute data (area) that the reconfiguration attribute data (area) of D remotely-sensed data are III, F remotely-sensed data are V.
S64, every part of remotely-sensed data carried out to entry and indicate and set up, the entry land type of every part of remotely-sensed data and reconfiguration attribute data thereof being constructed as simultaneously this part of remotely-sensed data indicates; The form that this entry indicates is land type+attribute.Certainly have multiple reconfiguration attribute data, the form that this entry indicates is land type+reconfiguration attribute 1+ reconfiguration attribute 2+......+ reconfiguration attribute N.
S66, the described entry in each class bunch indicated and carries out permutation and combination, and add up the quantity of each permutation and combination.
Enumerate an example below, so that reconstruct to be described.
Fig. 4 is the reconstruct example schematic of the embodiment of the present invention.Please refer to Fig. 4, form T1 is through cluster and remotely-sensed data after reconfiguration attribute data decimation, and this remotely-sensed data is divided into 4 classes bunch (cluster), is respectively c1 ~ c4.Every part of corresponding a piece of land of remotely-sensed data, comprise land type, attribute 1 and attribute 2, and shown as clause name using land type, namely soil is rendered as land type in a database, and such as Gobi desert, water water etc.Form T2 is through entry and indicates the remotely-sensed data after setting up.The clause name of all remotely-sensed datas is carried out entry and indicate establishment, form entry and indicate.Indicated the land type both can knowing this remotely-sensed data by entry, also can know the association attributes of this remotely-sensed data.Such as, land type is saline and alkaline, attribute 1(and reconfiguration attribute) be A, attribute 2(and reconfiguration attribute) remotely-sensed data of 1, after setting up, its entry is denoted as saline and alkaline _ A_1.This entry indicates to set up and generally sets up successively each class bunch.Form T3 is the result entry in each class bunch being indicated to permutation and combination.In form T3, data have not had having divided of class bunch, and the combination that each entry indicates is shown in the left side one list, and the number of times (this number of times statistics obtains based on all classes bunch) that this combination occurs is shown in that list of the right.Such as, water waters _ and being combined in all classes bunch of B_1 and saline and alkaline _ A_1 only occurred once, the number of times of therefore this combination is the 1(frequency is 1).The combination of Gobi desert _ B_1, exposed soil ground _ c_2 and reservoir swag _ c_2 respectively occurred once respectively in class bunch c2, class bunch c4, and the number of times of therefore this combination is the 2(frequency is 2).The result (combination+number of times) of this permutation and combination, the result namely reconstructed, using the input as associated rule discovery.
S8, to reconstruct after remotely-sensed data carry out associated rule discovery, obtain association results.
Correlation degree between this step entry mainly obtained in permutation and combination indicates, represents with association probability.Associated rule discovery algorithm has a variety of and all very ripe, and the present embodiment preferably this associated rule discovery algorithm is Apriori algorithm (i.e. the algorithm of Boolean Association Rules Mining Frequent Itemsets Based).The result that this associated rule discovery obtains is the degree of correlation of each combination inside.Such as, water waters _ B_1 → saline and alkaline _ A_1, and support is 1/16, and degree of confidence is 50%.Saline and alkaline _ A_1 → water waters _ B_1, and support is 1/16, and degree of confidence is 25%.
The association results that S10, storage association probability are greater than probability threshold value T forms decision support unit.
Association probability is greater than probability threshold value T, and represent that correlation degree is between the two higher, have decision references meaning, the set of such result just defines decision support unit.The span of general this probability threshold value T, for being greater than 65%, is specifically preferably 80%.
Through the decision support unit that said method obtains, there is the correlation rule that a lot of degree of correlation is higher, when knowing given data, other unknown data higher with this given data degree of correlation can be provided as a reference, support decision-making.The decision support unit of the application of the invention embodiment carries out land type identification, can reduce the workload of land type identification, improves land type recognition efficiency, also can improve the accuracy of identification simultaneously, reduces the requirement to the personnel of identification.
Embodiment two
The present embodiment provides a kind of land type recognition methods and module.
Fig. 5 is the land type recognition methods process flow diagram of the embodiment of the present invention.Please refer to Fig. 5, this land type recognition methods, comprises the steps:
S100, acquisition decision support unit; This step adopts above-mentioned decision support cell formation method to obtain this decision support unit.
S120, obtain the attribute data of remotely-sensed data to be identified and at least one land type that identify remotely-sensed data adjacent with described remotely-sensed data to be identified and attribute data thereof.
A soil satellite remote sensing images (remotely-sensed data namely to be identified) to be identified, the attribute data can known in this remotely-sensed data to be identified has a lot, such as size, center point coordinate, girth etc., comprise reconfiguration attribute data in this attribute data; The land type of the identification remotely-sensed data (the soil satellite remote sensing images that soil satellite remote sensing images namely to be identified with this is adjacent) of this remotely-sensed data to be identified and attribute data, comprise reconfiguration attribute data in these attribute datas.
Step S100 and step S120 provides the basis of land type identification.
S140, the attribute data of described remotely-sensed data to be identified and described adjacent identification remotely-sensed data to be handled as follows:
S142, obtain described remotely-sensed data to be identified and the described adjacent reconfiguration attribute data identifying remotely-sensed data; Such as, reconfiguration attribute can be area and center point coordinate.
S144, described adjacent identification remotely-sensed data carried out to entry and indicate and set up, the described adjacent land type identifying remotely-sensed data and reconfiguration attribute data thereof are constructed as the described adjacent entry identifying remotely-sensed data simultaneously and indicate; Such as, what this was adjacent identifies that the entry of remotely-sensed data indicates and can water _ B_1 for water.
S160, in conjunction with describedly adjacent identifying that the entry of remotely-sensed data indicates, the reconfiguration attribute data of described remotely-sensed data to be identified and described decision support unit, provide the land type of described remotely-sensed data to be identified.
The land type recognition methods of the present embodiment can provide multiple land type alternatively, for user's reference, also can provide the land type of 1 land type as this remotely-sensed data to be identified.
The present embodiment is preferred, provides multiple land type alternatively, can realize the identification of supervision land type, user only need simple judge just can comparatively fast, accuracy rate completes land type identification work higher, this reduces the skill set requirements to user.This also overcomes simultaneously and carrys out error problem directly to going out soil type belt, reduces the artificial erroneous judgement of user as far as possible.
Fig. 6 is the structural representation of the land type identification module of the embodiment of the present invention.Please refer to Fig. 6, the land type identification module of the present embodiment, comprising:
Adopt the decision support unit 100 that above-mentioned construction method obtains, for providing decision support.
First acquiring unit 120, for obtaining the attribute data of remotely-sensed data to be identified, and at least one land type that identify remotely-sensed data adjacent with described remotely-sensed data to be identified and attribute data thereof.
A soil satellite remote sensing images (remotely-sensed data namely to be identified) to be identified, the attribute data can known in this remotely-sensed data to be identified has a lot, such as size, center point coordinate, girth etc., comprise reconfiguration attribute data in this attribute data; The land type of the identification remotely-sensed data (the soil satellite remote sensing images that soil satellite remote sensing images namely to be identified with this is adjacent) of this remotely-sensed data to be identified and attribute data, comprise reconfiguration attribute data in these attribute datas.
First attribute processing unit 140, is connected with the first acquiring unit 120, for obtaining described remotely-sensed data to be identified and the described adjacent reconfiguration attribute data identifying remotely-sensed data; Such as, reconfiguration attribute can be area and center point coordinate.This first attribute processing unit 140 also carries out entry to described adjacent identification remotely-sensed data and indicates establishment, the described adjacent land type identifying remotely-sensed data and reconfiguration attribute data thereof is constructed as the described adjacent entry identifying remotely-sensed data simultaneously and indicates; Such as, what this was adjacent identifies that the entry of remotely-sensed data indicates and can water _ B_1 for water.
First processing unit 160, be connected with the first attribute processing unit 140 and decision support unit 100, for in conjunction with the described adjacent reconfiguration attribute data identifying the entry sign of remotely-sensed data, described remotely-sensed data to be identified and described decision support unit, provide the land type of described remotely-sensed data to be identified.
The land type identification module of the present embodiment can provide multiple land type alternatively, for user's reference, also can provide the land type of 1 land type as this remotely-sensed data to be identified.
The present embodiment is preferred, land type identification module provides multiple land type alternatively, realization has supervision land type identification, user only need simple judge just can comparatively fast, accuracy rate completes land type identification work higher, this reduces the requirement to user.Which overcome and carry out error problem directly to going out soil type belt.
This first processing unit 160 receiving this adjacent entry identifying remotely-sensed data that the first attribute processing unit 140 provides and indicate < water and water _ B_1>(below be called for short known entry and indicate) time, known objective with this can be found from this decision support unit 100 and be shown with the correlation rule of pass, such as water waters _ B_1 → saline and alkaline _ A_1, water waters _ B_1 → ground coverage grass _ A_1; The reconfiguration attribute (A, 1) of the remotely-sensed data to be identified that this first processing unit 160 provides according to the correlation rule of this acquisition and this first attribute processing unit 140 again, obtain the highest one or more of correlation rules of the degree of association (water water _ B_1 → saline and alkaline _ A_1, water water _ B_1 → ground coverage grass _ A_1), and provide the land type (saline and alkaline or ground coverage grass) of this remotely-sensed data to be identified according to the correlation rule of this highest degree of association.
The land type recognition methods of the present embodiment and module, provide land type or candidate's land type, reduce the requirement of user, improve recognition efficiency, also improve recognition accuracy simultaneously.
Embodiment three
The present embodiment provides a kind of land type identification and evaluation method and module.
Fig. 7 is the land type identification and evaluation method flow diagram of the embodiment of the present invention.Please refer to Fig. 7, the land type identification and evaluation method of the embodiment of the present invention, comprises the steps:
S200, above-mentioned construction method is adopted to obtain decision support unit.
S220, obtain the remotely-sensed data of remotely-sensed data to be evaluated and described remotely-sensed data periphery to be evaluated.Remotely-sensed data in this step comprises reconfiguration attribute data.In this step, the remotely-sensed data of periphery refers to the remotely-sensed data that remotely-sensed data all and to be evaluated is adjacent.
S240, the remotely-sensed data of remotely-sensed data to be evaluated and remotely-sensed data periphery to be evaluated carried out to entry and indicate and set up.The entry land type of the remotely-sensed data of remotely-sensed data to be evaluated and periphery thereof and reconfiguration attribute data thereof being constructed as simultaneously the remotely-sensed data of described remotely-sensed data to be evaluated and remotely-sensed data periphery to be evaluated indicates.This step is similar to step S144, S64, is not described in detail.
S260, according to decision support unit and set up after entry indicate obtain P(c, A); Wherein, this A is remotely-sensed data to be evaluated; C is the numbering of A periphery remotely-sensed data; P(c, A) represent the land type of remotely-sensed data based on being numbered c, the probable value of the land type of A remotely-sensed data.This P(c, A) preparation method know for those skilled in the art, no longer carefully state at this.
S280, according to formula:
obtain the land type identification and evaluation value of remotely-sensed data to be evaluated, wherein, Q
arepresent the evaluation of estimate of the land type identification of A remotely-sensed data, n is A periphery remotely-sensed data number summation.
By the evaluation method of the present embodiment, land type recognition result can be quantized, an objective identification and evaluation standard is provided.This Q
avalue larger, indicate this land type identification more accurate.
Fig. 8 is the land type identification and evaluation module diagram of the embodiment of the present invention.Please refer to Fig. 8, the land type identification and evaluation module of the embodiment of the present invention, comprising:
Adopt the decision support unit 200 that above-mentioned construction method obtains, for providing decision support;
Second acquisition unit 220, for obtaining the remotely-sensed data of remotely-sensed data to be evaluated and described remotely-sensed data periphery to be evaluated; The remotely-sensed data that this second acquisition unit 200 obtains comprises reconfiguration attribute data.The remotely-sensed data of this periphery refers to the remotely-sensed data that remotely-sensed data all and to be evaluated is adjacent.
Second attribute processing unit 240, is connected with second acquisition unit 220, indicates establishment for carrying out entry to the remotely-sensed data of remotely-sensed data to be evaluated and remotely-sensed data periphery to be evaluated.The entry land type of the remotely-sensed data of remotely-sensed data to be evaluated and periphery thereof and reconfiguration attribute data thereof being constructed as simultaneously the remotely-sensed data of described remotely-sensed data to be evaluated and remotely-sensed data periphery to be evaluated indicates.This second attribute processing unit 240 is similar to the first attribute processing unit 140, can refer to the first attribute processing unit 140 and understands, therefore be not described in detail.
Probability calculation unit 260, be connected with this second attribute processing unit 240 and decision support unit 200, indicate according to decision support unit 200 and the entry after setting up and obtain P(c, A), this A is remotely-sensed data to be evaluated, and c is the numbering of A periphery remotely-sensed data, P(c, A) land type based on the remotely-sensed data being numbered c is represented, the probable value of the land type of A remotely-sensed data;
Evaluate computing unit 280, be connected with this probability calculation unit 260, for according to formula:
obtain the land type identification and evaluation value of remotely-sensed data to be evaluated, wherein, Q
arepresent the evaluation of estimate of the land type identification of A remotely-sensed data, n is A periphery remotely-sensed data number summation.
This second attribute processing unit 240 receive that second acquisition unit 220 obtains for obtain remotely-sensed data to be evaluated and described remotely-sensed data periphery to be evaluated remotely-sensed data after, entry is carried out to the remotely-sensed data of remotely-sensed data to be evaluated and remotely-sensed data periphery to be evaluated and indicates to set up and form entry and indicate.Probability calculation unit 260 indicates according to the entry after establishment and the correlation rule of decision support unit 200 calculates P(c, A).P(c, A that this evaluation computing unit 280 acquisition probability computing unit 260 provides), and to P(c, A) square carry out cumulative summation and obtain Q
a.
By the evaluation module of the present embodiment, land type recognition result can be quantized, an objective identification and evaluation standard is provided.This Q
avalue larger, indicate this land type identification more accurate.
Embodiment four
The present embodiment provides a kind of land type verification system, and this land type verification system comprises above-mentioned land type identification module and/or above-mentioned land type identification and evaluation module.
The land type verification system of the present embodiment, has above-mentioned land type identification module and above-mentioned land type identification and evaluation module, can realize the land type accuracy identified identification and the evaluation of land type simultaneously.This is the function that existing land type verification system does not have.
Although illustrate and describe the present invention with reference to specific embodiment; but it will be apparent to one skilled in the art that the various changes can made in form and details all should fall into the protection domain of claims of the present invention when not departing from the spirit and scope of the present invention of scope by claim and equivalents thereof.
Claims (13)
1. a construction method for decision support unit, is characterized in that, comprises the steps:
S2, obtain and identified the remotely-sensed data of land type, described remotely-sensed data comprises some attribute datas;
S4, choose at least two described attribute datas and carry out cluster as characteristic attribute data, by described remotely-sensed data divide into several classes bunch;
S6, the remotely-sensed data in each class bunch to be reconstructed based on described land type and characteristic attribute data;
S8, to reconstruct after remotely-sensed data carry out associated rule discovery, obtain association results;
The association results that S10, storage association probability are greater than probability threshold value T forms decision support unit.
2. construction method as claimed in claim 1, it is characterized in that, described step S6 comprises the steps:
S62, select one or more in described characteristic attribute data as reconfiguration attribute data;
S64, every part of remotely-sensed data carried out to entry and indicate and set up, the entry land type of every part of remotely-sensed data and reconfiguration attribute data thereof being constructed as simultaneously this part of remotely-sensed data indicates;
S66, the described entry in each class bunch indicated and carries out permutation and combination, and add up the quantity of each permutation and combination.
3. construction method as claimed in claim 2, it is characterized in that, described step S62 further comprises:
To select characteristic attribute data in one or more carry out classification process, and using classification process after result substitute the characteristic attribute data of described selection as reconfiguration attribute data.
4. construction method as claimed in claim 1, it is characterized in that, probability threshold value T is more than or equal to 65%.
5. the construction method as described in any one of Claims 1-4, is characterized in that, described characteristic attribute data to comprise in land area, centre coordinate and soil girth one or more.
6. the construction method as described in any one of Claims 1-4, is characterized in that, described cluster adopts the one in DBSCAN algorithm, K-means algorithm and KNN algorithm.
7. the construction method as described in any one of Claims 1-4, is characterized in that, described associated rule discovery adopts Apriori algorithm.
8. a decision support unit, is characterized in that, adopts the construction method described in any one of claim 1 to 7 to build and obtains.
9. a land type recognition methods, is characterized in that, comprises the steps:
S100, the construction method described in any one of claim 1 to 7 is adopted to obtain decision support unit;
S120, obtain the attribute data of remotely-sensed data to be identified and at least one land type that identify remotely-sensed data adjacent with described remotely-sensed data to be identified and attribute data thereof;
S130, the attribute data of described remotely-sensed data to be identified and described adjacent identification remotely-sensed data to be handled as follows:
S132, obtain described remotely-sensed data to be identified and the described adjacent reconfiguration attribute data identifying remotely-sensed data;
S134, described adjacent identification remotely-sensed data carried out to entry and indicate and set up, the described adjacent land type identifying remotely-sensed data and reconfiguration attribute data thereof are constructed as the described adjacent entry identifying remotely-sensed data simultaneously and indicate;
S140, in conjunction with describedly adjacent identifying that the entry of remotely-sensed data indicates, the reconfiguration attribute data of described remotely-sensed data to be identified and described decision support unit, provide the land type of described remotely-sensed data to be identified.
10. a land type identification module, is characterized in that, comprising:
Adopt the decision support unit that the construction method described in any one of claim 1 to 7 obtains, for providing decision support;
First acquiring unit, for obtaining the attribute data of remotely-sensed data to be identified, and at least one land type that identify remotely-sensed data adjacent with described remotely-sensed data to be identified and attribute data thereof;
First attribute processing unit, for obtaining described remotely-sensed data to be identified and the described adjacent reconfiguration attribute data identifying remotely-sensed data; And
Entry is carried out to described adjacent identification remotely-sensed data and indicates establishment, the described adjacent land type identifying remotely-sensed data and reconfiguration attribute data thereof are constructed as the described adjacent entry identifying remotely-sensed data simultaneously and indicate;
First processing unit, in conjunction with the described adjacent reconfiguration attribute data identifying the entry sign of remotely-sensed data, described remotely-sensed data to be identified and described decision support unit, provides the land type of described remotely-sensed data to be identified.
11. 1 kinds of land type identification and evaluation methods, is characterized in that, comprise the steps:
S200, the construction method described in any one of claim 1 to 7 is adopted to obtain decision support unit;
S220, obtain the remotely-sensed data of remotely-sensed data to be evaluated and described remotely-sensed data periphery to be evaluated;
S240, the remotely-sensed data of remotely-sensed data to be evaluated and remotely-sensed data periphery to be evaluated carried out to entry and indicate and set up;
S260, obtain P(c, A according to decision support unit and remotely-sensed data to be evaluated), this A is remotely-sensed data to be evaluated, c is the numbering of A periphery remotely-sensed data, P(c, A) represent the land type of remotely-sensed data based on being numbered c, the probable value of the land type of A remotely-sensed data;
S280, according to formula:
obtain the land type identification and evaluation value of remotely-sensed data to be evaluated, wherein, Q
arepresent the evaluation of estimate of the land type identification of A remotely-sensed data, n is A periphery remotely-sensed data number summation.
12. 1 kinds of land type identification and evaluation modules, is characterized in that, comprising:
Adopt the decision support unit that the construction method described in any one of claim 1 to 7 obtains, for providing decision support;
Second acquisition unit, for obtaining the remotely-sensed data of remotely-sensed data to be evaluated and described remotely-sensed data periphery to be evaluated;
Second attribute processing unit, indicates establishment for carrying out entry to the remotely-sensed data of remotely-sensed data to be evaluated and remotely-sensed data periphery to be evaluated;
Probability calculation unit, obtains P(c, A according to decision support unit and remotely-sensed data to be evaluated), this A is remotely-sensed data to be evaluated, and c is the numbering of A periphery remotely-sensed data, P(c, A) land type based on the remotely-sensed data being numbered c is represented, the probable value of the land type of A remotely-sensed data;
Evaluate computing unit: according to formula:
obtain the land type identification and evaluation value of remotely-sensed data to be evaluated, wherein, Q
arepresent the evaluation of estimate of the land type identification of A remotely-sensed data, n is A periphery remotely-sensed data number summation.
13. 1 kinds of land type verification systems, is characterized in that, comprise land type identification module according to claim 10 and/or land type identification and evaluation module according to claim 12.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310740391.6A CN104750707A (en) | 2013-12-27 | 2013-12-27 | Decision support unit, land type identifying and verifying system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201310740391.6A CN104750707A (en) | 2013-12-27 | 2013-12-27 | Decision support unit, land type identifying and verifying system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN104750707A true CN104750707A (en) | 2015-07-01 |
Family
ID=53590410
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201310740391.6A Pending CN104750707A (en) | 2013-12-27 | 2013-12-27 | Decision support unit, land type identifying and verifying system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN104750707A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107315647A (en) * | 2017-06-26 | 2017-11-03 | 广州视源电子科技股份有限公司 | Outlier detection method and system |
| CN111898014A (en) * | 2020-08-12 | 2020-11-06 | 哈尔滨工业大学 | Cognitive service-oriented user intention identification method and system |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101650748A (en) * | 2009-09-14 | 2010-02-17 | 杨敬锋 | Soil quality evaluation method and system |
| CN102073879A (en) * | 2010-12-02 | 2011-05-25 | 南京大学 | Method for identifying characteristic land categories of ocean remote sensing images of coast on basis of semi-supervised learning |
| CN102521598A (en) * | 2011-12-16 | 2012-06-27 | 北京市城市规划设计研究院 | Identification method based on remote-sensing image |
| CN102622345A (en) * | 2011-01-26 | 2012-08-01 | 杨存建 | High-precision land-utilization remote sensing updating technology with synergistic multisource spatio-temporal data |
-
2013
- 2013-12-27 CN CN201310740391.6A patent/CN104750707A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101650748A (en) * | 2009-09-14 | 2010-02-17 | 杨敬锋 | Soil quality evaluation method and system |
| CN102073879A (en) * | 2010-12-02 | 2011-05-25 | 南京大学 | Method for identifying characteristic land categories of ocean remote sensing images of coast on basis of semi-supervised learning |
| CN102622345A (en) * | 2011-01-26 | 2012-08-01 | 杨存建 | High-precision land-utilization remote sensing updating technology with synergistic multisource spatio-temporal data |
| CN102521598A (en) * | 2011-12-16 | 2012-06-27 | 北京市城市规划设计研究院 | Identification method based on remote-sensing image |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107315647A (en) * | 2017-06-26 | 2017-11-03 | 广州视源电子科技股份有限公司 | Outlier detection method and system |
| CN111898014A (en) * | 2020-08-12 | 2020-11-06 | 哈尔滨工业大学 | Cognitive service-oriented user intention identification method and system |
| CN111898014B (en) * | 2020-08-12 | 2021-07-13 | 哈尔滨工业大学 | Cognitive service-oriented user intention identification method and system |
| WO2022033073A1 (en) * | 2020-08-12 | 2022-02-17 | 哈尔滨工业大学 | Cognitive service-oriented user intention recognition method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Chacón et al. | Data-driven density derivative estimation, with applications to nonparametric clustering and bump hunting | |
| CN106506705A (en) | Listener clustering method and device based on location-based service | |
| CN104751477A (en) | Space domain and frequency domain characteristic based parallel SAR (synthetic aperture radar) image classification method | |
| CN105243139A (en) | Deep learning based three-dimensional model retrieval method and retrieval device thereof | |
| CN103888541A (en) | Method and system for discovering cells fused with topology potential and spectral clustering | |
| CN109726749A (en) | A kind of Optimal Clustering selection method and device based on multiple attribute decision making (MADM) | |
| CN110993037A (en) | Protein activity prediction device based on multi-view classification model | |
| CN108804473A (en) | The method, apparatus and Database Systems of data query | |
| Fowler et al. | All pixels are useful, but some are more useful: Efficient in situ data collection for crop-type mapping using sequential exploration methods | |
| Al-Naymat et al. | Dimensionality reduction for long duration and complex spatio-temporal queries | |
| Khurshid et al. | A residual-dyad encoder discriminator network for remote sensing image matching | |
| CN105740917A (en) | High-resolution remote sensing image semi-supervised multi-view feature selection method with tag learning | |
| CN107451617A (en) | One kind figure transduction semisupervised classification method | |
| CN104331711B (en) | SAR image recognition methods based on multiple dimensioned fuzzy mearue and semi-supervised learning | |
| CN103905993A (en) | Region misjudgment solution method based on indoor positioning | |
| CN104750707A (en) | Decision support unit, land type identifying and verifying system | |
| CN109447943A (en) | A kind of object detection method, system and terminal device | |
| Brunello et al. | Towards interpretability in fingerprint based indoor positioning: May attention be with us | |
| Srinivasa Raju et al. | Classification of Indian meteorological stations using cluster and fuzzy cluster analysis, and Kohonen artificial neural networks | |
| CN104657473B (en) | A kind of large-scale data method for digging guaranteeing quality monotonicity | |
| Galba et al. | Public transportation bigdata clustering | |
| Gupta et al. | Spatial data classification using decision tree models | |
| Wang et al. | Mining common spatial-temporal periodic patterns of animal movement | |
| Petrovska et al. | Machine Learning with Remote Sensing Image Data Sets | |
| Love et al. | Topological convolutional neural networks |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20150701 |
|
| RJ01 | Rejection of invention patent application after publication |