A multi-source POI data matching and fusion method
By preprocessing and spatial position correction of multi-source POI data, combined with data mining and name similarity calculation, the mismatch and data incomplete data in POI matching and fusion in the prior art are solved, and more efficient and accurate POI data matching and fusion.
Patent Information
- Application Number
- CN202210205311.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-03-02
- Publication Date
- 2024-11-29
- Estimated Expiration
- 2042-03-02
AI Technical Summary
The existing POI matching methods have problems such as missed matching and incomplete data, especially in high-density urban environments, resulting in excessive matching amount and low data quality.
By acquiring multi-source POI data for preprocessing, the spatial position is corrected based on the topological relationship between building two-dimensional surface data and POI data, the matching scheme is determined and matched, and combined with data mining and name similarity calculation, the efficient matching and fusion of POI data is achieved. .
It effectively corrects the errors in the space of POI, improves data integrity and matching accuracy, solves the missed allocation problem, and improves the matching efficiency and quality of high-density urban POI data.
Smart Images

Figure CN114676754B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of spatial big data technology, and in particular to a multi-source POI data matching and fusion method. Background Art
[0002] POI matching and fusion are important means to enrich digital map content and improve map quality through multi-source data, and have been widely used in various fields of people's lives. However, existing POI matching methods may have missed matches, and existing POI fusion methods may have incomplete data after fusion.
[0003] Therefore, the existing technology still needs to be improved and developed. Summary of the invention
[0004] The technical problem to be solved by the present invention is to provide a multi-source POI data matching and fusion method in view of the above-mentioned defects of the prior art, aiming to solve the problems of POI data missing and incomplete data quantity in the prior art.
[0005] The technical solution adopted by the present invention to solve the problem is as follows:
[0006] In a first aspect, an embodiment of the present invention provides a multi-source POI data matching and fusion method, wherein the method comprises:
[0007] Acquire multi-source POI data, and perform data preprocessing on the multi-source POI data to obtain multi-source preprocessed POI data;
[0008] Based on the topological relationship between the two-dimensional building surface data and the POI data, correcting the spatial position of the multi-source pre-processed POI data;
[0009] Determine a plurality of matching schemes according to the multi-source pre-processed POI data after the spatial position is corrected, and match the multi-source pre-processed POI data after the spatial position is corrected according to the plurality of matching schemes to obtain a matching result;
[0010] Based on the matching result, the multi-source pre-processed POI data are fused to obtain POI fused data.
[0011] In one implementation, the preprocessing the multi-source POI data to obtain the multi-source preprocessed POI data includes:
[0012] Unifying the POI names in the multi-source POI data based on a preset standard;
[0013] Parsing the addresses of the multi-source POI data to obtain a plurality of address segments, and adding the plurality of address segments as new attribute values to the multi-source POI data;
[0014] Data mining is performed on the plurality of address segment data sets to obtain mining results, and based on preset association rules and the mining results, the addresses of the multi-source POI data are supplemented.
[0015] In one implementation, the correcting the spatial position of the multi-source pre-processed POI data based on the topological relationship between the two-dimensional building surface data and the POI data includes:
[0016] Based on spatial geometry algorithms, the spatial position of the centroid of all building two-dimensional surface data is extracted;
[0017] Adding the spatial position of the centroid of the two-dimensional building surface data as an attribute value to the multi-source pre-processed POI data;
[0018] Based on the topological relationship between the two-dimensional building surface data and the POI data, the multi-source pre-processed POI data with added attribute values is input into a preset spatial analysis tool to obtain a POI data set with positioning errors;
[0019] The centroid in the incorrectly positioned POI data set is relocated to correct the spatial position of the multi-source pre-processed POI data.
[0020] In one implementation, after relocating the centroid in the POI data set with the wrong positioning to correct the spatial position of the multi-source pre-processed POI data, the method further includes:
[0021] When the relocated POI dataset has a positioning error, the centroid in the POI dataset with the positioning error is relocated again.
[0022] In one implementation, the multi-source pre-processed POI data after the spatial position is corrected includes a POI to be matched and a plurality of heterogeneous POIs; determining a plurality of matching schemes based on the multi-source pre-processed POI data after the spatial position is corrected, and matching the multi-source pre-processed POI data after the spatial position is corrected according to the plurality of matching schemes, and obtaining a matching result includes:
[0023] Analyzing the multi-source pre-processed POI data after correcting the spatial position to obtain attribute features corresponding to the multi-source pre-processed POI data;
[0024] Determine several matching schemes according to the attribute characteristics;
[0025] According to some of the matching schemes, the multi-source pre-processed POI data after spatial position correction is matched to obtain a matching result.
[0026] In one implementation, the attribute feature includes a first attribute feature, a second attribute feature, or a third attribute feature; and determining a plurality of matching schemes according to the attribute feature includes:
[0027] When the attribute feature is the first attribute feature, the matching scheme is to combine and match the POI to be matched with a plurality of heterogeneous POIs according to the first address fine granularity;
[0028] When the attribute feature is the second attribute feature, the matching scheme is to combine and match the POI to be matched with the heterogeneous POI according to the second address fine granularity;
[0029] When the attribute feature is the third attribute feature or the POI to be matched does not match the heterogeneous POI, the matching solution is to combine and match the POI to be matched with the heterogeneous POI at a preset distance value.
[0030] In one implementation, matching the multi-source pre-processed POI data after spatial position correction according to the plurality of matching schemes to obtain a matching result includes:
[0031] According to the matching schemes, the multi-source pre-processed POI data after spatial position correction are combined in pairs;
[0032] Calculating the similarity of the combined point pair names; wherein the point pair names are names of the point pairs formed by combining the multi-source pre-processed POI data in pairs;
[0033] When the similarity of the point pair names is greater than a preset similarity threshold, the matching result is a successful match, and the successfully matched POI to be matched and the heterogeneous POI are stored in the target data set.
[0034] In one implementation, the fusing the multi-source pre-processed POI data based on the matching result to obtain POI fused data includes:
[0035] When there are missing attribute values in the to-be-matched POI and the heterogeneous POI in the target data set, the attribute values are supplemented;
[0036] When the names of the POI to be matched in the target dataset and the POI from the heterogeneous source are inconsistent, no operation is performed;
[0037] When there are several similar attribute values in the POI to be matched and the heterogeneous POI in the target data set, the preset first attribute value replaces the original attribute values in the POI to be matched and the heterogeneous POI;
[0038] When the same attribute values in the to-be-matched POI and the heterogeneous POI in the target data set have unique values, the preset second attribute value replaces the original attribute values in the to-be-matched POI and the heterogeneous POI;
[0039] The multi-source pre-processed POI data after attribute value supplementation and attribute value replacement is used as POI fusion data.
[0040] In a second aspect, an embodiment of the present invention further provides a multi-source POI data matching and fusion device, wherein the device comprises:
[0041] A multi-source pre-processed POI data acquisition module acquires multi-source POI data and performs data pre-processing on the multi-source POI data to obtain multi-source pre-processed POI data;
[0042] A spatial position correction module, used to correct the spatial position of the multi-source pre-processed POI data based on the topological relationship between the building two-dimensional surface data and the POI data;
[0043] A matching result determination module, used to determine a plurality of matching schemes according to the multi-source pre-processed POI data after the spatial position is corrected, and to match the multi-source pre-processed POI data after the spatial position is corrected according to the plurality of matching schemes to obtain a matching result;
[0044] The POI fusion data determination module is used to fuse the multi-source pre-processed POI data based on the matching result to obtain POI fusion data.
[0045] In a third aspect, an embodiment of the present invention further provides an intelligent terminal, comprising a memory and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by one or more processors, wherein the one or more programs include a method for executing multi-source POI data matching and fusion as described in any one of the above.
[0046] In a fourth aspect, an embodiment of the present invention further provides a non-temporary computer-readable storage medium. When the instructions in the storage medium are executed by a processor of an electronic device, the electronic device can execute the multi-source POI data matching and fusion method as described in any one of the above.
[0047] Beneficial effects of the present invention: the embodiment of the present invention first obtains multi-source POI data, and performs data preprocessing on the multi-source POI data to obtain multi-source preprocessed POI data; then, based on the topological relationship between the two-dimensional building surface data and the POI data, the spatial position of the multi-source preprocessed POI data is corrected; then, according to the multi-source preprocessed POI data after the spatial position is corrected, several matching schemes are determined, and according to the several matching schemes, the multi-source preprocessed POI data after the spatial position is corrected is matched to obtain matching results; finally, based on the matching results, the multi-source preprocessed POI data is fused to obtain POI fused data; it can be seen that the embodiment of the present invention fully mines the existing information of POI, completes incomplete attributes, and corrects conflicting attributes; the method of efficiently matching point pairs by name similarity solves the problem of POI missing caused by inaccurate spatial position, is particularly effective for high-density urban POI data, and can solve the problem of excessive matching caused by too dense POI points; the wrong spatial position of POI is corrected, the POI fusion result is better, and the data is more complete. BRIEF DESCRIPTION OF THE DRAWINGS
[0048] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.
[0049] Figure 1 A schematic flow chart of a multi-source POI data matching and fusion method provided in an embodiment of the present invention.
[0050] Figure 2 A schematic diagram of a building address data set association rule provided in an embodiment of the present invention.
[0051] Figure 3 A flow chart of a POI matching solution provided in an embodiment of the present invention.
[0052] Figure 4 A schematic diagram of the spatial distribution of a POI before and after spatial position correction according to an embodiment of the present invention.
[0053] Figure 5 A schematic diagram of POI fusion results provided by an embodiment of the present invention.
[0054] Figure 6 This is a principle block diagram of a multi-source POI data matching and fusion device provided in an embodiment of the present invention.
[0055] Figure 7This is a block diagram of the internal structure principle of the intelligent terminal provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0056] The present invention discloses a multi-source POI data matching and fusion method. In order to make the purpose, technical solution and effect of the present invention clearer and more specific, the present invention is further described in detail with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.
[0057] It will be understood by those skilled in the art that, unless expressly stated, the singular forms "one", "said", and "the" used herein may also include plural forms. It should be further understood that the term "comprising" used in the specification of the present invention refers to the presence of the features, integers, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components and / or groups thereof. It should be understood that when we refer to an element as being "connected" or "coupled" to another element, it may be directly connected or coupled to the other element, or there may be intermediate elements. In addition, the "connection" or "coupling" used herein may include wireless connection or wireless coupling. The term "and / or" used herein includes all or any unit and all combinations of one or more associated listed items.
[0058] It will be understood by those skilled in the art that, unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as generally understood by those skilled in the art in the art to which the present invention belongs. It should also be understood that terms such as those defined in general dictionaries should be understood to have meanings consistent with the meanings in the context of the prior art, and will not be interpreted with idealized or overly formal meanings unless specifically defined as herein.
[0059] In the prior art, the most commonly used POI matching method is a POI similarity calculation method based on the spatial position of the POI. This method first calculates the similarity of POIs from different sources, selects the POI point pairs with the highest similarity as the POI matching point pairs, and then corrects and completes the POI attribute values through data fusion rules.
[0060] Most of the current POI fusion methods focus on using machine learning to determine the POI category, and then combine the spatial location clustering and non-spatial attribute similarity of POI to find the correct fusion object from POI data sets from different sources. This method is effective in determining the POI category, but due to insufficient data preprocessing, the following deficiencies have occurred: 1) There is no systematic study on the POI address algorithm, and there is a lack of effective solutions for mining, completing address information and standardizing address formats; 2) For algorithms that introduce machine learning or artificial intelligence models, the robustness is low and cannot be well applied to various cities to truly and effectively improve the fusion accuracy of POI. 3) The point-to-point distance calculated from the spatial location of the POI is greatly affected by the quality of the POI spatial location information. Setting the distance too small is prone to missing matching problems, and setting the spatial range too large is prone to excessive matching. 4) Due to incomplete data mining, many POIs are missed.
[0061] In order to solve the problems of the prior art, the present embodiment provides a multi-source POI data matching and fusion method, which fully mines the existing POI information, completes incomplete attributes, and corrects conflicting attributes; the method of efficiently matching point pairs by name similarity solves the problem of POI missing caused by inaccurate spatial position, which is especially effective for high-density urban POI data and can solve the problem of excessive matching volume caused by too dense POI points; the incorrect spatial position of POI is corrected, and the POI fusion result is better and the data is more complete.
[0062] Exemplary Methods
[0063] This embodiment provides a multi-source POI data matching and fusion method, which can be applied to smart terminals of spatial big data. Figure 1 As shown, the method includes:
[0064] Step S100, acquiring multi-source POI data, and performing data preprocessing on the multi-source POI data to obtain multi-source preprocessed POI data;
[0065] Specifically, a POI (Point of Interest) in a geographic information system can be a house, a shop, a mailbox, a bus stop, etc. Multi-source POI data can be obtained from a website or a database. Multi-source POI data may have inconsistent formats or erroneous information, so it is necessary to preprocess the multi-source POI data. The preprocessing method can be to unify the format, correct the erroneous information, or remove the non-compliant data, without specific restrictions.
[0066] Step S100 includes the following steps:
[0067] S101, unifying POI names in the multi-source POI data based on a preset standard;
[0068] S102, parsing the addresses of the multi-source POI data to obtain a plurality of address segments, and adding the plurality of address segments as new attribute values to the multi-source POI data;
[0069] S103: performing data mining on the plurality of address segment data sets to obtain mining results, and supplementing the addresses of the multi-source POI data based on preset association rules and the mining results.
[0070] Specifically, the preset standard is the main language of the POI data, such as Chinese, that is, the POI names of the multi-source POI data are unified based on Chinese. In this embodiment, the present invention uses web crawler technology to automatically batch search the English names of POIs in Google map, and then uses regular expressions to identify and store their corresponding Chinese names, and unifies the English names in each source POI data set into simplified Chinese. In addition, by comparing the data sets before and after the name standardization, the accuracy of the POI name is evaluated, and the inaccurate judgment standard is: POIs that cannot be searched on Google map and names with typos compared to the Chinese names on Google map are inaccurate POI names. After unifying the POI names in the multi-source POI data, the addresses of the multi-source POI data are parsed to obtain several address segments, and the several address segments are added to the multi-source POI data as new attribute values. In this embodiment, each POI address is parsed into 6 address segments using regular expressions and added to each source POI data set as new attribute values: street, street number, building name, building seat number, floor, room number; multiple data sets with completely split addresses are obtained. In addition, if any part of the six parts of a POI address after being parsed is missing, the address is considered to be incomplete. By counting the proportion of incomplete addresses in the POI data, the completeness of the POI addresses from each source is evaluated. A POI address has been parsed into six parts, and any part missing is considered to be incomplete. Data mining is performed on several address segment data sets, and the mining results are obtained by filtering out all addresses whose building names are not empty from the address parsed data set, and establishing a data set of building addresses through re-operation, which includes streets, street numbers, and building names (building names). The purpose is to complete the missing parts in the POI address, correct erroneous addresses, and use address attributes to search for POI point pairs in a small range more quickly during the matching process. Finally, based on the preset association rules and the mining results, the addresses of the multi-source POI data are supplemented. In this embodiment, the preset association rules are the association rules between streets, street numbers and buildings, as shown in the following example. Figure 2As shown. The number of association rules from different sources is compared to evaluate the relative completeness of association rules from different sources. At the same time, in order to correct the writing errors of POI street names, all street-level address information is screened out from the address parsed data set to establish a street name data set. Using the building address data set, for POIs that are originally located in buildings and lack building names in the address, the missing building information of the POI is completed through the POI street and street number; for POIs that only have room numbers in the address but lack building seat numbers and floor numbers, the seat number, floor number, and room number of the building where the POI is located can be mined based on the room number information in the address parsed data set; using the street address data set, through the fuzzy matching algorithm, that is, using the algorithm written in python language to operate on the data set in CSV, the street names of the POI data sets from each source are standardized. In this way, the quality of the name and address data before matching is improved, and other multi-source POI data sets Si (i>=1) are obtained. In one implementation method, the timeliness of POI data from different sources is judged by the data update time.
[0071] After obtaining multi-source pre-processed POI data, you can execute the following Figure 1 The following steps are shown: S200, based on the topological relationship between the building two-dimensional surface data and the POI data, correcting the spatial position of the multi-source pre-processed POI data; accordingly, step S200 includes the following steps:
[0072] S201, extracting the spatial position of the centroid of all building two-dimensional surface data based on a spatial geometry algorithm;
[0073] S202, adding the spatial position of the centroid of the two-dimensional building surface data as an attribute value to the multi-source pre-processed POI data;
[0074] S203, based on the topological relationship between the two-dimensional building surface data and the POI data, inputting the multi-source pre-processed POI data with added attribute values into a preset spatial analysis tool to obtain a POI data set with positioning errors;
[0075] S204: relocate the centroid of the POI data set with the wrong positioning, so as to correct the spatial position of the multi-source pre-processed POI data.
[0076] Specifically, based on the spatial geometry algorithm, the spatial position of the centroid of all building two-dimensional surface data is extracted to calculate the spatial position (centro_lon, centro_lat) of the centroid of all building surface data by geometry, and then the spatial position of the centroid of the building two-dimensional surface data is added to the multi-source pre-processed POI data as an attribute value, such as retaining it as a new attribute for navigation. Next, based on the topological relationship between the building two-dimensional surface data and the POI data, the multi-source pre-processed POI data with added attribute values is input into a preset spatial analysis tool to obtain a POI data set with positioning errors, such as using the spatial analysis tool of ArcGIS to obtain the topological relationship between the POI point data and the Building two-dimensional surface data in the matched source data set A, and outputting the POI data set A_wrong_loc with absolute positioning errors, such as restaurants and hotels located outside the building or bus stations located inside the building; finally, the centroid of the POI data set with positioning errors is relocated to correct the spatial position of the multi-source pre-processed POI data, for example, according to the centroid of the building corresponding to the "building name" of the POI in A_wrong_loc in the 2D data set, the POI is relocated by the following formula:
[0077] [lon1,lat1]=[centro_lon+random(-0.000001,0.000001),centro_lat+random(-0.000001,0.000001)].
[0078] Among them, [centro_lon, centro_lat] represents the spatial position of the centroid of the building surface data on the two-dimensional plane. The longitude and latitude 0.000001 is 10 meters in real life. The POI is relocated within 10m of the building centroid to ensure that the corrected POI is located inside the building when the electronic map is zoomed and does not overlap.
[0079] In one implementation, the accuracy of spatial positions from different sources (such as the proportion of POIs with absolutely wrong positions) is counted to obtain the credibility of spatial positions of POIs from different sources.
[0080] In one implementation, the corrected spatial position [lon1, lat1] replaces A and S i The spatial position [lon, lat] in .
[0081] In one implementation, after relocating the centroid of the erroneously positioned POI data set to correct the spatial position of the multi-source preprocessed POI data, the following step is also included: when the relocated POI data set has a positioning error, relocating the centroid of the erroneously positioned POI data set again.
[0082] Specifically, when the relocated POI dataset still has errors, further processing is required. In this embodiment, it is determined whether the spatially corrected POI is located within the 2D building surface. If it is outside the building surface, a smaller spatial range is used to make it fall within the building surface.
[0083] After correcting the spatial position of the multi-source pre-processed POI data, the following can be performed: Figure 1 The following steps are shown: S300, according to the multi-source pre-processed POI data after the spatial position is corrected, several matching schemes are determined, and according to the several matching schemes, the multi-source pre-processed POI data after the spatial position is corrected is matched to obtain a matching result; accordingly, step S300 includes the following steps:
[0084] S301, parsing the multi-source pre-processed POI data after correcting the spatial position, and obtaining attribute features corresponding to the multi-source pre-processed POI data;
[0085] S302, determining several matching schemes according to the attribute characteristics;
[0086] S303: According to the plurality of matching schemes, the multi-source pre-processed POI data after spatial position correction is matched to obtain a matching result.
[0087] Specifically, Figure 3 As shown, the multi-source pre-processed POI data after the spatial position correction is parsed, and then the POI addresses parsed in the pre-processing are identified. Based on the idea of parallel, the POI is divided into three data sets: Class a has "building" information; Class b has a null value for the "building" attribute, but has "street" and "street number" information; Class c has null "street", "street number" and "building" information, and the small-scale spatial range in which it is located cannot be identified through the address information. The address granularity is: a>b>c. The "building" information is taken as the first attribute feature, the "building" attribute is null, and the "street" and "street number" information are taken as the second attribute feature; the "street", "street number" and "building" information are all null as the third attribute feature.
[0088] The attribute feature includes a first attribute feature, a second attribute feature or a third attribute feature; step S302 includes the following steps: when the attribute feature is the first attribute feature, the matching scheme is to combine and match the POI to be matched with a plurality of heterogeneous POIs according to the first address fine granularity; when the attribute feature is the second attribute feature, the matching scheme is to combine and match the POI to be matched with the heterogeneous POI according to the second address fine granularity; when the attribute feature is the third attribute feature or the POI to be matched does not match the heterogeneous POI, the matching scheme is to combine and match the POI to be matched with the heterogeneous POI according to a preset distance value.
[0089] Specifically, when the first address granularity is class a and the second address granularity is class b, when the attribute feature is the first attribute feature, the matching scheme is to combine and match the POI to be matched with several heterogeneous POIs according to the first address granularity, to combine and match heterogeneous POIs in the same spatial scale in pairs; the class a POI in the source A to be matched with the heterogeneous S i POIs in the same "building" in category a are combined and matched in pairs, as the source A to be matched and fused, and a POI in A will be combined with S i When the attribute feature is the second attribute feature, the matching scheme is to combine and match the POI to be matched with several heterogeneous POIs according to the second address fine-grainedness, that is, to combine and match heterogeneous POIs in the same spatial scale in pairs; the b-type POIs of the source to be matched in A and the POIs in the b-type in Si in the same "street" and "street number" are combined and matched in pairs, as the source A to be matched and fused, and a POI in A will be matched with S i When the attribute feature is the third attribute feature or the POI to be matched does not match the heterogeneous POI, the matching scheme is to combine and match the POI to be matched with the heterogeneous POI at a preset distance value, such as i The matched POIs in Match1 are deleted from the dataset, and the remaining POIs and class C are put into the same dataset class D. Then, the A source POI is used as a unit and the Si source point pairs within 600m are combined and matched.
[0090] Step S303 includes the following steps: combining the multi-source pre-processed POI data after spatial position correction in pairs according to several matching schemes; calculating the similarity of the combined point pair names; wherein the point pair names are the names of the point pairs formed after the multi-source pre-processed POI data are combined in pairs; when the similarity of the point pair names is greater than a preset similarity threshold, the matching result is a successful match, and the successfully matched POI to be matched and the heterogeneous POI are stored in the target data set.
[0091] Specifically, when the matching scheme is to combine and match the POI to be matched with several heterogeneous POIs according to the first address fine-grainedness, heterogeneous POIs at the same address fine-grainedness and the same spatial scale are combined in pairs to generate a POI pre-matching point pair set. i The POIs in the same "building" in category a are combined in pairs, and then the similarity of the combined point pair names is calculated; wherein, the point pair names are the names of the point pairs formed by the pairwise combination of the multi-source preprocessed POI data; when the similarity of the point pair names is greater than the preset similarity threshold, the matching result is a successful match, and the successfully matched POIs to be matched and the heterogeneous POIs are stored in the target data set. The name similarity of the point pairs is calculated, and the algorithm used is a similarity algorithm for non-semantic ordered strings based on the DTW dynamic time series theory. Taking the A source POI as the unit, the POI point pairs with the largest name similarity are screened out. The name similarities of the point pairs screened out in the previous step are sorted, and when the name similarity is found to be greater than 0.7, the POI point pair is considered to be the same POI that has been successfully matched. All the attributes of these point pairs are put into the matching result data set Match1. When the matching scheme is to combine and match the POI to be matched with the heterogeneous POI according to the second address fine granularity, the multi-source pre-processed POI data after spatial position correction are combined in pairs, such as: the b-type POI of the source to be matched in A and the POI of the b-type in Si on the same "street" and "street number" are combined in pairs. As the source A to be matched and fused, a POI in A will be combined with S i . Then calculate the similarity of the combined point pair names; wherein, the point pair names are the names of the point pairs formed by the combination of the multi-source pre-processed POI data in pairs; when the similarity of the point pair names is greater than the preset similarity threshold, the matching result is a successful match, and the successfully matched POI to be matched and the heterogeneous POI are stored in the target data set, for example: the name similarity of the point pair is calculated, and the algorithm used is a similarity algorithm of non-semantic ordered strings based on DTW dynamic time series theory. Taking A source POI as the unit, filter out the POI point pairs with the largest name similarity. Sort the name similarities of the point pairs filtered out in the previous step, and when it is found that the name similarity is greater than 0.7, the POI point pair is considered to be the same POI that has been successfully matched. All attributes of these point pairs are put into the matching result data set Match1. When the matching scheme is to combine and match the POI to be matched with the heterogeneous POI at a preset distance value, the multi-source pre-processed POI data after spatial position correction are combined in pairs, such as in A and S iDelete the matched POIs in Match1 from the data set, put the remaining POIs and class c into the same data set class d, and then combine the A-source POI with the Si-source point pairs within 600m to generate a point pair data set to be matched. Calculate the similarity of the combined point pair names; wherein the point pair names are the names of the point pairs formed by combining the multi-source preprocessed POI data in pairs; when the similarity of the point pair names is greater than the preset similarity threshold, the matching result is a successful match, and the successfully matched POIs to be matched and the heterogeneous POIs are stored in the target data set, such as: matching by calculating the name similarity of the POI point pairs. Filter the point pairs with a name similarity greater than 0.7 again and they are considered to pass through a POI, and they are stored in the matching result Match1. Points with a maximum name similarity less than 0.5 are considered to have no matching points and are stored in the result No_Match.
[0092] After getting the matching results, you can execute Figure 1 The following steps are shown: S400, based on the matching result, the multi-source pre-processed POI data is fused to obtain POI fused data. Accordingly, step S400 includes the following steps:
[0093] S401, when there are missing attribute values in the to-be-matched POI and the heterogeneous POI in the target data set, supplement the attribute values;
[0094] S402: When the names of the POI to be matched in the target data set and the POI from a different source are inconsistent, no operation is performed;
[0095] S403, when there are several similar attribute values in the POI to be matched and the heterogeneous POI in the target data set, the original attribute values in the POI to be matched and the heterogeneous POI are replaced by the preset first attribute value;
[0096] S404, when the same attribute values in the to-be-matched POI and the heterogeneous POI in the target data set have unique values, the original attribute values in the to-be-matched POI and the heterogeneous POI are replaced by the preset second attribute value;
[0097] S405: Use the multi-source pre-processed POI data after attribute value supplementation and attribute value replacement as POI fusion data.
[0098] Specifically, first determine whether the matched POI successfully matches the POI from another source, put the successfully matched POI point pairs into the data set Result1, put the unmatched POI into the data set Result2, and put the POI that does not exist in the matched source but exists in other sources, that is, the missing POI, into the data set Result3. The missing POI is the POI that has not been matched multiple times in other sources.
[0099] When there are missing attribute values in the POI to be matched and the heterogeneous POI in the target data set, the attribute values are supplemented as follows: for the missing attributes in the matched POI in Result1, the attribute missing information completion scheme is used, that is, the empty value is supplemented with information from other sources, and the attributes of the matching points are directly used to supplement the fused data set.
[0100] When the names of the POI to be matched in the target dataset and the POI from the other source are inconsistent, no operation is performed: For the name conflict of the POI matching point pair in Result1, since the name has been corrected in the preprocessing process, the conflict ignoring strategy (i.e., no operation) is adopted, and the attribute values involved in the conflict are recorded and displayed together. For example, the names of the same POI in source A and source B are Chinese name name1 and English name name2, and both name1 and name2 are recorded and stored as name attribute values.
[0101] When there are several similar attribute values in the POI to be matched and the heterogeneous POI in the target data set, the preset first attribute value replaces the original attribute value in the POI to be matched and the heterogeneous POI. Specifically: for the attributes of the POI that can have multiple values, such as the merchant contact information of the hotel accommodation POI in the implementation of the present invention can have multiple numbers, a conflict resolution policy can be adopted to use the preset first attribute value (a more authoritative data source) to enrich the POI attributes.
[0102] When the same attribute values in the to-be-matched POI and the heterogeneous POI in the target data set have unique values, the preset second attribute value replaces the original attribute values in the to-be-matched POI and the heterogeneous POI. Specifically, a conflict avoidance strategy is adopted for the attributes with unique values of the POI, including the POI spatial position in the implementation of the present invention, and the preset second attribute value (the corrected spatial position or the spatial position in a more authoritative data source such as government data) is used as the fused POI position.
[0103] In one implementation, for attributes that are not included in the fused POI, a scheme for completing missing attribute information is used. For example, in the implementation of the present invention, the license plate information and number of rooms of the hotel accommodation service POI are missing, and the relevant information from the government matching point is directly added to the fused data set. For another example, if the attributes of the fused source POI do not include the contact information of the merchant, the attributes of the fused source POI data will have more attribute information of the merchant contact information after enrichment.
[0104] In another implementation, the data sets before and after fusion are compared to evaluate the number completeness of the single-source POIs before fusion and the data quality of the POIs in the fused data set after fusion. By comparing the number of POIs of the fused source before fusion and the number of POIs after fusion, the number completeness of the POIs before fusion is obtained. Then, the data quality of the POIs after data fusion is evaluated, including name accuracy, address completeness, and attributes.
[0105] In order to illustrate the effect of the POI matching and fusion method provided in the embodiment of the present invention, taking the POI of hotel accommodation services on Amap as an example, POI data from Hong Kong is used to match and fuse the POI of Amap. After the fusion, the number of POIs increases significantly, and the effect of enriching the attribute values is remarkable.
[0106] like Figure 4 The positions of the POI points before and after the spatial position correction are given. It can be seen that most of the hotel accommodation service POIs located outside the building have been corrected to the inside of the building.
[0107] Table 1 further shows the data quality evaluation results after preprocessing in the embodiment of the present invention.
[0108] index Accuracy Name accuracy 91.4% Address Integrity 42.7% Association rule completeness 72.6% Spatial position accuracy 83%
[0109] Table 1
[0110] Table 2 further provides a quantitative analysis of the matching results of the embodiments of the present invention.
[0111] category Proportion Matched POI 49% Missing POI 26% Confirm that no POI is matched 32.5%
[0112] Table 2
[0113] Table 3 further gives the data quality evaluation results after fusion in the embodiment of the present invention.
[0114] index Accuracy POI quantity completeness 64.4% Name accuracy 80% Address Integrity 42.7% Timeliness of data sources 13.7% Spatial position accuracy 86.7%
[0115] Table 3
[0116] like Figure 5 This is the final fusion result of POI. It can be found that the present invention effectively corrects and fuses the POI attribute values.
[0117] Exemplary Devices
[0118] like Figure 6 As shown in , an embodiment of the present invention provides a multi-source POI data matching and fusion device, which includes a multi-source pre-processed POI data acquisition module 501, a spatial position correction module 502, a matching result determination module 503 and a POI fusion data determination module 504, wherein:
[0119] A multi-source pre-processed POI data acquisition module 501 acquires multi-source POI data and performs data pre-processing on the multi-source POI data to obtain multi-source pre-processed POI data;
[0120] A spatial position correction module 502, for correcting the spatial position of the multi-source pre-processed POI data based on the topological relationship between the building two-dimensional surface data and the POI data;
[0121] A matching result determination module 503 is used to determine a plurality of matching schemes according to the multi-source pre-processed POI data after the spatial position is corrected, and to match the multi-source pre-processed POI data after the spatial position is corrected according to the plurality of matching schemes to obtain a matching result;
[0122] The POI fusion data determination module 504 is used to fuse the multi-source pre-processed POI data based on the matching result to obtain POI fusion data.
[0123] Based on the above embodiments, the present invention further provides an intelligent terminal, whose principle block diagram can be shown as follows: Figure 7 As shown. The intelligent terminal includes a processor, a memory, a network interface, a display screen, and a temperature sensor connected through a system bus. Among them, the processor of the intelligent terminal is used to provide computing and control capabilities. The memory of the intelligent terminal includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the intelligent terminal is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a multi-source POI data matching and fusion method is implemented. The display screen of the intelligent terminal can be a liquid crystal display screen or an electronic ink display screen, and the temperature sensor of the intelligent terminal is pre-set inside the intelligent terminal to detect the operating temperature of the internal device.
[0124] Those skilled in the art will understand that Figure 7 The schematic diagram in the figure is only a block diagram of a part of the structure related to the scheme of the present invention, and does not constitute a limitation on the smart terminal to which the scheme of the present invention is applied. The specific smart terminal may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.
[0125] In one embodiment, a smart terminal is provided, comprising a memory and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by one or more processors, and the one or more programs include instructions for performing the following operations:
[0126] Acquire multi-source POI data, and perform data preprocessing on the multi-source POI data to obtain multi-source preprocessed POI data;
[0127] Based on the topological relationship between the two-dimensional building surface data and the POI data, correcting the spatial position of the multi-source pre-processed POI data;
[0128] Determine a plurality of matching schemes according to the multi-source pre-processed POI data after the spatial position is corrected, and match the multi-source pre-processed POI data after the spatial position is corrected according to the plurality of matching schemes to obtain a matching result;
[0129] Based on the matching result, the multi-source pre-processed POI data are fused to obtain POI fused data.
[0130] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided by the present invention can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).
[0131] In summary, the present invention discloses a multi-source POI data matching and fusion method, the method comprising: acquiring multi-source POI data, and performing data preprocessing on the multi-source POI data to obtain multi-source preprocessed POI data; correcting the spatial position of the multi-source preprocessed POI data based on the topological relationship between the building two-dimensional surface data and the POI data; determining a number of matching schemes according to the multi-source preprocessed POI data after the spatial position is corrected, and matching the multi-source preprocessed POI data after the spatial position is corrected according to the several matching schemes to obtain a matching result; based on the matching result, fusing the multi-source preprocessed POI data to obtain POI fusion data; in the embodiment of the present invention, the existing information of POI is fully mined, incomplete attributes are supplemented, and conflicting attributes are corrected; the method of efficiently matching point pairs by name similarity solves the problem of POI missing caused by inaccurate spatial position, is particularly effective for POI data in high-density cities, and can solve the problem of excessive matching caused by too dense POI points; the wrong spatial position of POI is corrected, the POI fusion result is better, and the data is more complete.
[0132] Based on the above embodiments, the present invention discloses a multi-source POI data matching and fusion method. It should be understood that the application of the present invention is not limited to the above examples. For ordinary technicians in this field, it can be improved or transformed according to the above description. All these improvements and transformations should fall within the scope of protection of the claims attached to the present invention.
Claims
1. A multi-source POI data matching and fusion method, characterized in that: The method comprises: Acquire multi-source POI data, and perform data preprocessing on the multi-source POI data to obtain multi-source preprocessed POI data; Based on the topological relationship between the two-dimensional building surface data and the POI data, correcting the spatial position of the multi-source pre-processed POI data; Determine a plurality of matching schemes according to the multi-source pre-processed POI data after the spatial position is corrected, and match the multi-source pre-processed POI data after the spatial position is corrected according to the plurality of matching schemes to obtain a matching result; Based on the matching result, the multi-source pre-processed POI data are fused to obtain POI fused data; The correcting the spatial position of the multi-source pre-processed POI data based on the topological relationship between the two-dimensional building surface data and the POI data comprises: Based on spatial geometry algorithms, the spatial position of the centroid of all building two-dimensional surface data is extracted; Adding the spatial position of the centroid of the two-dimensional building surface data as an attribute value to the multi-source pre-processed POI data; Based on the topological relationship between the two-dimensional building surface data and the POI data, the multi-source pre-processed POI data with added attribute values is input into a preset spatial analysis tool to obtain a POI data set with positioning errors; Relocating the centroid in the POI data set with the wrong positioning to correct the spatial position of the multi-source pre-processed POI data; The multi-source pre-processed POI data after correcting the spatial position includes a POI to be matched and a plurality of hetero-source POIs; Determining a plurality of matching schemes according to the multi-source pre-processed POI data after the spatial position is corrected, and matching the multi-source pre-processed POI data after the spatial position is corrected according to the plurality of matching schemes to obtain a matching result includes: Analyzing the multi-source pre-processed POI data after correcting the spatial position to obtain attribute features corresponding to the multi-source pre-processed POI data; Determine several matching schemes according to the attribute characteristics; According to some of the matching schemes, the multi-source pre-processed POI data after spatial position correction is matched to obtain a matching result.
2. The multi-source POI data matching and fusion method according to claim 1, characterized in that: The performing data preprocessing on the multi-source POI data to obtain the multi-source preprocessed POI data comprises: Unifying the POI names in the multi-source POI data based on a preset standard; Parsing the addresses of the multi-source POI data to obtain a plurality of address segments, and adding the plurality of address segments as new attribute values to the multi-source POI data; Data mining is performed on the plurality of address segment data sets to obtain mining results, and based on preset association rules and the mining results, the addresses of the multi-source POI data are supplemented.
3. The multi-source POI data matching and fusion method according to claim 1, characterized in that: After relocating the centroid in the erroneously positioned POI data set to correct the spatial position of the multi-source pre-processed POI data, the method further includes: When the relocated POI dataset has a positioning error, the centroid in the POI dataset with the positioning error is relocated again.
4. The multi-source POI data matching and fusion method according to claim 1, characterized in that: The attribute feature includes a first attribute feature, a second attribute feature, or a third attribute feature; and determining a plurality of matching schemes according to the attribute feature includes: When the attribute feature is the first attribute feature, the matching scheme is to combine and match the POI to be matched with a plurality of heterogeneous POIs according to the first address fine granularity; When the attribute feature is the second attribute feature, the matching scheme is to combine and match the POI to be matched with the heterogeneous POI according to the second address fine granularity; When the attribute feature is the third attribute feature or the POI to be matched does not match the heterogeneous POI, the matching solution is to combine and match the POI to be matched with the heterogeneous POI at a preset distance value.
5. The multi-source POI data matching and fusion method according to claim 4, characterized in that: According to the plurality of matching schemes, matching the multi-source pre-processed POI data after spatial position correction to obtain matching results includes: According to the matching schemes, the multi-source pre-processed POI data after spatial position correction are combined in pairs; Calculating the similarity of the combined point pair names; wherein the point pair names are names of the point pairs formed by combining the multi-source pre-processed POI data in pairs; When the similarity of the point pair names is greater than a preset similarity threshold, the matching result is a successful match, and the successfully matched POI to be matched and the heterogeneous POI are stored in the target data set.
6. The multi-source POI data matching and fusion method according to claim 5, characterized in that: Based on the matching result, fusing the multi-source pre-processed POI data to obtain POI fused data includes: When there are missing attribute values in the to-be-matched POI and the heterogeneous POI in the target data set, the attribute values are supplemented; When the names of the POI to be matched in the target dataset and the POI from the heterogeneous source are inconsistent, no operation is performed; When there are several similar attribute values in the POI to be matched and the heterogeneous POI in the target data set, the preset first attribute value replaces the original attribute values in the POI to be matched and the heterogeneous POI; When the same attribute values in the to-be-matched POI and the heterogeneous POI in the target data set have unique values, the preset second attribute value replaces the original attribute values in the to-be-matched POI and the heterogeneous POI; The multi-source pre-processed POI data after attribute value supplementation and attribute value replacement is used as POI fusion data.
7. An intelligent terminal, characterized in that: The device comprises a memory and one or more programs, wherein the one or more programs are stored in the memory and are configured to be executed by one or more processors, and the one or more programs include being used to execute the method according to any one of claims 1 to 6.
8. A non-transitory computer-readable storage medium, characterized in that: When the instructions in the storage medium are executed by a processor of an electronic device, the electronic device is enabled to execute the method as claimed in any one of claims 1 to 6.
Citation Information
Patent Citations
Fault diagnosis method and system for gear bearing based on multi-source information fusion
AU2020103923A4
Multi-source road network data automatic fusion method based on road element topological feature classification
CN113704853A
