Geocoding calculation method and device based on big data

Abstract

The invention relates to a geocoding calculation method and device based on big data, and the method comprises the steps: carrying out the structural conversation of an address through a simplified classification system (11 types of core elements), optimizing an address analysis process, carrying out the forward prediction through a geocoding technology, completing the reverse verification through a reverse geocoding technology, and improving the reliability of an analysis result; carrying out clustering analysis on address coordinates by utilizing an aggregation algorithm, and enhancing the capturing capability of spatial distribution characteristics; abnormal data which does not conform to geographical boundary conditions are eliminated through address fence filtering, and the data quality is improved; a modular device is constructed and comprises an address structuring module, a forward and reverse address matching module, an aggregation grouping module, an address fence filtering module and a screening module, and full-process automatic processing is achieved. The method not only improves the precision and efficiency, but also reduces the system complexity, provides a feasible technical path for the acquisition of high-reliability and high-precision spatial data, and has a wide application prospect.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to a geocoding calculation method and apparatus based on big data. Background Technology

[0002] Geocoding, also known as address-matching, is the process of establishing a connection between geographic coordinates and given address information. It involves finding and marking the location coordinates corresponding to each address on a map, typically expressed as a point, line, or polygon coordinate vector. Reverse geocoding is the inverse process. Geocoding and reverse geocoding are crucial functions in Geographic Information Systems (GIS). With the accelerating pace of urban digitization, the application depth and breadth of GIS in key scenarios such as logistics, emergency rescue, intelligent transportation, urban planning, and commercial investment continue to expand. These scenarios place higher demands on the accuracy and reliability of address coordinate calculations. However, existing geocoding technologies face the following structural challenges:

[0003] Low accuracy: Traditional methods rely on static matching between structured addresses and limited data sources. This is limited by the standardization of address descriptions and makes it difficult to integrate multi-dimensional environmental information (such as main building features and surrounding topological relationships), resulting in large parsing errors.

[0004] Large errors: Mainstream map platforms exhibit three key differences at their underlying layers: different data sources, divergent core algorithm architectures, and independent coordinate system encryption (BD09 / GCJ02 / WGS84). Experiments show that while the overall accuracy of reverse geocoding across platforms can reach 98%, the overall accuracy of geocoding is only 90%. Even within the same coordinate system on the same platform, the accuracy rate for resolving the same standardized address with an offset of less than 20 meters across different interfaces / systems is less than 95%.

[0005] High complexity: The complex classification system built by commercial platforms to cover all scenarios (such as Baidu Maps providing at least 32 address types) significantly increases the cost of business adaptation, while 90% of actual applications only require 11 categories (5 categories of statutory administrative division system + 6 categories of basic system), namely, province / prefecture / county / township / village / development zone / road / road number / building / building number / POI (Point of Interest, a spatial entity with address coordinates and attribute information, which is the smallest unit of map data).

[0006] Lack of verification mechanism: The lack of an effective verification mechanism makes it difficult to guarantee the accuracy of the calculation results.

[0007] Existing geocoding technologies have significant shortcomings in terms of accuracy, error, complexity, and verification mechanisms, necessitating optimization through the introduction of high-precision parsing algorithms, simplification of classification systems, and the establishment of verification mechanisms. With the accelerating pace of urban digitalization, resolving these issues has become a top priority, directly impacting the operational efficiency and security of several key industries, including logistics, transportation, and emergency response. Summary of the Invention

[0008] To address the issues of low accuracy, large errors, high complexity, and lack of verification mechanisms in geocoding technology, a geocoding calculation method and device based on big data is proposed. The address resolution process is optimized by simplifying the classification system (11 core elements), geocoding technology is used for forward prediction, and reverse geocoding technology is used to complete reverse verification, thereby achieving high accuracy and high reliability of the coding.

[0009] The technical solution of the present invention is: a geocoding calculation method based on big data, which performs real-time verification calculation on a real-time updated geocoding database and a reverse geocoding database, specifically including the following steps: 1) Input address and other dimension information, and perform address structuring on the input information;

[0010] 2) Based on the roads or buildings in the structured data from step 1), under the constraints of legal administrative divisions or development zones including provinces, prefectures, and counties, match them with the geocoding database to obtain the matched structured address information; 3) Use a density-based spatial clustering algorithm to group the matched structured address coordinates to obtain address coordinate groups, that is, classify all addresses of the same roads or buildings into one coordinate group, and then filter them through the following steps;

[0011] 4) Using the legal administrative divisions or development zones in the structured data as corresponding coordinate fences, filter the address coordinate groups in step 3) to remove abnormal data that do not meet the geographical boundary conditions;

[0012] 5) Using a reverse geocoding database, perform range queries on the filtered address coordinate groups to retrieve the reverse address coordinate groups and related information; after verification and comparison, retain the address coordinate groups with consistent information.

[0013] 6) Finally, based on road + road number, building + building number, POI, and other dimensions, filter the address coordinates with the highest matching degree, and select the final coordinates as the system output parameters.

[0014] Furthermore, the address structuring specifically involves obtaining 11 types of address information, including road and building information. The 11 types of address information are: province, prefecture, county, township, village, development zone, road, road number, building, building number, and POI.

[0015] Furthermore, the density-based spatial clustering algorithm performs cluster analysis on address coordinates to obtain address coordinate groups, thereby enhancing the ability to capture spatial distribution characteristics.

[0016] Furthermore, step 4) filtering verifies the input address and filters out abnormal data that contain roads and buildings with the same name.

[0017] Furthermore, the clustering region boundaries obtained after filtering in step 4) are irregular two-dimensional discrete polygons. To simplify the calculation, they are replaced with 2×2 two-dimensional coordinate sets. In step 5), the minimum longitude X of each address coordinate set in step 4) is obtained. min Maximum longitude X max Minimum dimension Y min Maximum dimension Y max The query scope of the reverse geocoding database must include the road or building information from step 3); otherwise, it should be considered invalid.

[0018] A geocoding computing device based on big data includes: an address structuring module, a geocoding matching module, an aggregation grouping module, a coordinate fence filtering module, a reverse geocoding verification module, and a filtering module;

[0019] The address structuring module is responsible for structuring the input address, outputting 11 types of address information, and sending them to the geocoding matching module.

[0020] The geocoding matching module, through the roads or buildings in the structured address information, under the limitation of administrative division or development zone information, calls the geocoding database to obtain the corresponding address coordinates and related information in the geocoding database to match the structured address information, and the structured address information after matching is sent to the aggregation and grouping module.

[0021] The aggregation and grouping module uses an aggregation algorithm to aggregate and group the matched address coordinates.

[0022] The coordinate fence filtering module uses administrative division information or development zone information from the address structuring module results as coordinate fences to filter the aggregated grouping results.

[0023] The reverse geocoding verification module performs a range query on the reverse geocoding database of the filtered aggregated grouping results and performs further filtering by comparing feature points.

[0024] The filtering module selects the coordinates with the highest matching degree from the aggregated grouping results after secondary filtering and outputs them as the final result.

[0025] The beneficial effects of this invention are as follows: The geocoding calculation method and apparatus based on big data utilize geocoding technology for forward prediction, followed by reverse geocoding technology for reverse verification, improving the reliability of the parsing results; it employs aggregation algorithms to perform cluster analysis on address coordinates, enhancing the ability to capture spatial distribution characteristics; it uses address fence filtering to remove abnormal data that does not meet geographical boundary conditions, improving data quality; and it constructs a modular device, including an address structuring module, a forward and reverse address matching module, an aggregation and grouping module, an address fence filtering module, and a screening module, achieving fully automated processing. This not only improves accuracy and efficiency but also reduces system complexity, providing a feasible technical path for acquiring highly reliable and high-precision spatial data, and has broad application prospects. Attached Figure Description

[0026] Figure 1 This is a flowchart of the geocoding calculation method based on big data according to the present invention;

[0027] Figure 2 This is a schematic diagram of the geocoding computing device based on big data according to the present invention. Detailed Implementation

[0028] The present invention will now be described in detail with reference to the accompanying drawings and specific embodiments. These embodiments are based on the technical solution of the present invention and provide detailed implementation methods and specific operating procedures. However, the scope of protection of the present invention is not limited to the following embodiments.

[0029] like Figure 1 The flowchart shown is a geocoding calculation method based on big data according to the present invention. It is used to perform real-time verification calculations on a real-time updated geocoding database and a reverse geocoding database. The method specifically includes the following steps:

[0030] Step 1: Input information, which can be considered as system parameters, includes address and other dimension information. The address is structured based on the input information to obtain 11 types of address information, including road and building information. The 11 types of address information are: province, prefecture, county, township, village, development zone, road, road number, building, building number, and POI. For example, "Room 705, Block A, Wande Building, No. 199 Taishan Road, Xincheng Science and Technology Park, Jianye District, Nanjing City, Jiangsu Province" can be structured as: Province = Jiangsu Province; Prefecture = Nanjing City; County = Jianye District; Township = ; Village = ; Development Zone = Xincheng Science and Technology Park; Road = Taishan Road; Road Number = No. 199; Building = Wande Building; Building Number = Block A; POI = Room 705.

[0031] The fields in the geocoding database include, but are not limited to, address, geocode, unit where the address is located, province, prefecture, county, development zone, road, and building.

[0032] Fields in the reverse geocoding database include, but are not limited to, geocoding, address, province, prefecture, county, development zone, road, and building.

[0033] Step 2: Based on the roads or buildings identified in Step 1, and within the constraints of legal administrative divisions or development zones including provinces, prefectures, and counties, match them with the geocoding database to obtain the matched structured address information. That is, road = Taishan Road or building = Wande Building. Since the information on provinces, prefectures, counties, and development zones in Step 3 is not empty, additional constraints are added (e.g., province = Jiangsu Province; prefecture = Nanjing City; county = Jianye District). The complete query statement is: (road = Taishan Road or building = Wande Building) and ((province = Jiangsu Province and prefecture = Nanjing City and county = Jianye District) or (development zone = Xincheng Science and Technology Park)). This query retrieves the address coordinates and related information from the geocoding database for matching.

[0034] The geocoding database is existing, with fields including address, geocode, address location, province, prefecture, county, development zone, road, and building, and is updated in real time. At least one of the road or building must match; if neither matches, no processing is performed. There are three reasons for a mismatch: 1. The database lacks the data; 2. The address information provided by the user is not precise enough to specify a road or building (e.g., "Nanjing City, Jiangsu Province" cannot be used to construct a road or building); 3. The address is complex or incorrect, preventing the construction of a road or address. In actual production, the first and third reasons for mismatches trigger a feedback mechanism, leading to iterative algorithm upgrades. This feedback mechanism is not an improvement in this invention and will not be elaborated upon here.

[0035] Once a match is found, proceed to the next step for filtering and selection.

[0036] Step 3: The DBSCAN (Density-Based Spatial Clustering of Applications with Noise) density-based spatial clustering algorithm is used to group the matched structured address coordinates to obtain address coordinate groups. This method utilizes the aggregation effect of similar roads / buildings in the same area to achieve coordinate grouping, grouping closely spaced coordinates together and extracting road and building information from the same coordinate groups. Using aggregation algorithms to cluster address coordinates enhances the ability to capture spatial distribution characteristics.

[0037] The process involves grouping all addresses of the same road or building into a single coordinate group, and then filtering them using the following steps.

[0038] Step 4: Utilize the legally defined administrative divisions or development zones (development zones often geographically span multiple administrative regions, but administratively belong to only one) in the structured dataset as corresponding coordinate fences to filter the address coordinate groups from Step 3. Specifically, retrieve the coordinate fences for Province = Jiangsu Province; Prefecture = Nanjing City; County = Jianye District; Development Zone = Xincheng Science and Technology Park. This initially filters out erroneous address coordinate groups from the large dataset. The remaining coordinate groups should meet the following condition: at least one coordinate within the set should fall within the coordinate fence of the legally defined administrative division (Province = Jiangsu Province; Prefecture = Nanjing City; County = Jianye District) or development zone (Development Zone = Xincheng Science and Technology Park). Address fence filtering can eliminate abnormal data that does not conform to geographical boundary conditions, improving data quality.

[0039] The main function of legally designated administrative divisions or development zones is to filter multiple coordinate sets. Any coordinate set that intersects with or is close to the coordinate boundary of a legally designated administrative division or development zone is considered a valid coordinate set. Considering the boundary, in practical use, the determination is made in two steps: first, it is determined whether any coordinate in the coordinate set is within the coordinate boundary; if so, it is considered valid. If not, the shortest distance between the coordinate set and the coordinate boundary is calculated to see if it falls within the error range; if it does, it is considered valid.

[0040] Step 4 serves two purposes: firstly, it validates the input address. For example, if the address is 199 Taishan Road, Hangzhou, and the coordinates of Taishan Road and Wande Building are not within the Hangzhou city limits, then the input address will be validated. Secondly, it filters out cases where there are roads and buildings with the same name.

[0041] Step 5: Using a reverse geocoding database, perform range queries on the filtered address coordinate groups to retrieve the inverse address coordinate groups and related information. After verification and comparison, retain the address coordinate groups with consistent information. The clustering region boundaries obtained in Step 4 are irregular two-dimensional discrete polygons. To simplify calculations, they are replaced with 2×2 two-dimensional coordinate groups; that is, obtain the minimum longitude X of each address coordinate group in Step 4. min Maximum longitude X max Minimum dimension Y min Maximum dimension Y max Range query reverse geocoding database (X min -δ <= Longitude <= X max +δand Y min -δ <= Longitude <= Y max +δ, where δ is the error), the result must include the road or building information from step 3; otherwise, it should be considered invalid.

[0042] Step 6: Finally, filter the address coordinates with the highest matching degree based on road + road number, building + building number, POI, and other dimensions. That is, prioritize filtering the data with the highest matching degree of address information. If multiple data are matched, then filter the final coordinates through other dimensions (such as company name) and return them. The final coordinates are used as the system output parameters.

[0043] This invention primarily targets addresses containing road or building information, utilizing a large amount of existing geocoding and reverse geocoding data to provide highly reliable coordinates. Many map providers offer geocoding and reverse geocoding services; however, in practice, geocoding services have generally had lower accuracy compared to reverse geocoding services. For example, in real-world business scenarios, when querying which companies are within a 1km radius of a given business, after defining the coordinate range, it's often found that while some company addresses are geographically close, the road / building information within those addresses indicates they are actually quite far apart. This obvious error leads to frequent complaints. To improve this situation, the invention leverages the aggregation effect of identical roads / buildings, along with forward and reverse address matching and address fencing, to obtain highly reliable coordinates, ensuring the real-time reliability and effectiveness of the updated geocoding and reverse geocoding databases.

[0044] The present invention also proposes a geocoding calculation device based on big data, the device comprising: an address structuring module, a geocoding matching module, an aggregation grouping module, a coordinate fence filtering module, a reverse geocoding verification module, and a filtering module.

[0045] The address structuring module is responsible for structuring the input address, outputting 11 types of address information, and sending them to the geocoding matching module.

[0046] The geocoding matching module, through the road or building information in the structured address information, as well as the administrative division (province, prefecture, county) or development zone information, calls and queries the geocoding database to obtain the corresponding address coordinates and related information in the geocoding database to match the structured address information. The structured address information after matching is sent to the aggregation and grouping module.

[0047] The aggregation and grouping module uses an aggregation algorithm to aggregate and group the matched address coordinates.

[0048] The coordinate fence filtering module uses administrative division information or development zone information from the address structuring module results as coordinate fences to filter the aggregated grouping results.

[0049] The reverse geocoding verification module performs a range query on the reverse geocoding database of the filtered aggregated grouping results and performs further filtering by comparing feature points.

[0050] The filtering module selects the coordinates with the highest matching degree from the aggregated grouping results after secondary filtering and outputs them as the final result.

[0051] This invention uses data from 100,000 enterprises to test the geocoding calculation method based on big data provided by this invention, with a sample size of 1,000. The test results show that the original interface has an accuracy rate of approximately 90%, while the new algorithm based on this invention can achieve an accuracy rate of 98%. For addresses with road + road number / building + building number, the overall accuracy rate can be improved to 99%.

[0052] The embodiments described above merely illustrate specific implementations of the present invention, and while the descriptions are detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. A big data based geocoding computation method, characterized in that, Real-time updating of the geographic coding database and the inverse geographic coding database is checked in real time, and specifically includes the following steps: 1) inputting an address and other dimensional information, and performing address structuring on the input information; 2) matching the address with the geographic coding database according to the road or building in the structuring in step 1) within the limits of the legal administrative division or the development zone including the province, the region and the county, to obtain matched structured address information; 3) grouping the matched structured address coordinates by using a density-based spatial clustering algorithm, to obtain address coordinate groups, that is, all addresses of the same road or building are classified into one coordinate group, and then the following steps are used for screening; 4) filtering the address coordinate groups in step 3) by using the legal administrative division or the development zone in the structuring as a corresponding coordinate fence, to remove abnormal data that does not meet the geographic boundary condition; 5) performing range query on the filtered address coordinate groups by using the inverse geographic coding database, to query out inverse address coordinate groups and related information; After checking and comparing, the address coordinate groups with consistent information are retained; 6) finally, the address coordinates with the highest matching degree are screened according to the road + road number, the building + building number, the POI and other dimensional information, and the final coordinates are screened out as the system output parameters.

2. The big data based geocoding computation method according to claim 1 or 2, characterized in that, The address structuring specifically refers to obtaining 11 types of address information including road and building information, and the 11 types of address information are: province, region, county, township, village, development zone, road, road number, building, building number and POI.

3. The big data based geocoding computation method of claim 2, wherein, The density-based spatial clustering algorithm is used to analyze and cluster the address coordinates, to obtain address coordinate groups and enhance the ability to capture spatial distribution characteristics.

4. The big data based geocoding computation method of claim 2, wherein, The filtering in step 4) realizes the checking of the input address, and simultaneously screens and removes abnormal data with the same-named road and building.

5. The big data based geocoding computation method of claim 2, wherein, The clustering region boundaries obtained after filtering in step 4) are irregular two-dimensional discrete polygons. To simplify the calculation, they are replaced with 2×2 two-dimensional coordinate sets. In step 5), the minimum longitude X of each address coordinate set in step 4) is obtained. min Maximum longitude X max Minimum dimension Y min Maximum dimension Y max The query scope of the reverse geocoding database must include the road or building information from step 3); otherwise, it should be considered invalid. 6.A big data based geocoding computing device, characterized in that, It includes: an address structuring module, a geographic coding matching module, an aggregation grouping module, a coordinate fence filtering module, an inverse geographic coding checking module and a screening module; The address structuring module is responsible for structuring the input address, and outputs 11 types of address information to the geographic coding matching module; The geographic coding matching module calls the geographic coding database to obtain corresponding address coordinates and related information in the geographic coding database by using the road or building in the structured address information within the limits of the administrative division or the development zone information, matches the structured address information, and sends the matched structured address information to the aggregation grouping module; The aggregation grouping module aggregates and groups the matched address coordinates by using an aggregation algorithm; The coordinate fence filtering module filters the results of the aggregation grouping according to the administrative division information or the development zone information in the results of the address structuring module; The inverse geographic coding checking module performs range query on the filtered aggregation grouping results in the inverse geographic coding database, and further filters by comparing feature points; The screening module screens out the coordinates with the highest matching degree from the twice-filtered aggregation grouping results as the final output.

Images