Urban parking zoning method and zoning system based on multi-source big data
Through multi-source big data analysis and clustering algorithm, parking partitions are constructed in combination with road network and administrative boundaries, the problems of rough partitioning and poor practicality in traditional methods are solved, efficient and accurate urban parking partitions are achieved, and differentiated management is supported. Improve the scientific nature of urban traffic management.
Patent Information
- Application Number
- CN202311325162.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-10-13
- Publication Date
- 2024-11-22
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
The traditional urban parking partitioning method has rough partitioning, few considerations, and weak practicality. It is difficult to process massive data and quickly build highly practical partition boundaries, and it is difficult to achieve differentiated partition management.
Using a multi-source big data method, through data import, collection, dimension screening, traffic community clustering and parking partition network structure, the weighted improved K-mean clustering algorithm and aggregation clustering algorithm are used to combine road network and administration The boundary constructs the parking partition to realize the mapping of clustering attributes.
It has achieved efficient and accurate urban parking zoning, improved scientificity and operability, can quickly respond to urban changes, supports differentiated management, and improved the scientific nature of urban traffic management.
Smart Images

Figure CN117392868B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of urban traffic and parking management, and in particular to an urban parking zoning method and zoning system based on multi-source big data. Background Art
[0002] 1. Overview of Urban Parking Zoning Research
[0003] Urban transportation is an important foundation for the efficient development of social activities. With the continuous advancement of my country's urbanization process, more and more megacities and super-large cities have emerged. These cities have a high share of car travel, prominent contradictions between parking supply and demand, and increasingly serious parking problems.
[0004] There are great differences in social economy, population vitality, land use, transportation facilities and operations between different areas of large cities. Exploring and classifying the characteristics of urban transportation and parking operations in different areas, and adopting appropriate management policies and measures to achieve refined and differentiated management of urban parking are effective ways to improve the problem of urban parking difficulties.
[0005] The main research idea of urban parking zoning is to mine the characteristic differences of zoning objects in urban traffic and parking operations from multi-source big data related to urban parking, so as to identify the clusters to which the zoning objects belong and form a method for urban parking zoning. This method can be applied to parking construction index zoning, parking fee zoning, parking management zoning, etc., and has broad and urgent application prospects.
[0006] 2. Current status of urban parking zoning research
[0007] Existing urban parking zoning research mainly uses the following three methods: qualitative analysis method, quantitative analysis method, and multi-level zoning method.
[0008] (1) Qualitative analysis
[0009] In the initial research stage, due to the difficulty of data collection, in order to facilitate practical operation and management, managers considered the overall supply and demand characteristics of urban parking at the administrative level and tended to delineate zones from the city center outward. Taking Beijing's parking construction index zoning as an example, managers divided Beijing into four parking zones based on the Second Ring Road, the Fourth Ring Road, and the Fifth Ring Road, and adopted differentiated parking construction policies.
[0010] (2) Quantitative analysis method
[0011] With the development of technology, managers have gradually mastered the data related to urban parking. They usually take urban traffic areas or urban matrix grids as basic units, and use methods such as fuzzy clustering, multi-index weighted evaluation, and kernel density analysis to quantitatively analyze the traffic attributes of urban basic units, and then delineate parking zones according to the scores of basic units.
[0012] (3) Multi-level zoning method
[0013] Due to the different planning, construction, operation and management characteristics of urban parking in the three spatial dimensions of urban macro, regional meso and parking lot micro, some researchers have proposed that urban parking zoning should have a corresponding hierarchical system to guide parking construction and management at different levels. The macro level is mainly responsible for the strategic zoning of urban parking, the meso level is mainly responsible for the regional parking planning zoning, and the micro level is mainly responsible for the specific parking lot implementation plan zoning of local sections and plots.
[0014] 3. Problems with existing urban parking zones
[0015] Traditional urban parking zoning has problems such as rough zoning, few considerations, and poor practicality:
[0016] First, the traditional urban parking zoning method usually combines administrative boundaries to ensure operability, which is less scientific and accurate.
[0017] Second, due to the limitations of early data collection technology and computer performance, the urban parking data used for analysis was single-dimensional and small in volume, lacking a multi-dimensional and in-depth analysis of urban parking characteristics;
[0018] Third, the traditional urban parking zoning method is difficult to handle the massive parking big data generated by cities with highly popular transportation information equipment in the Internet era;
[0019] Fourth, the traditional urban parking zoning method makes it difficult to quickly construct urban parking zoning boundaries that are practical and easy to manage in a short period of time.
[0020] 4. Technical issues to be solved
[0021] The technical difficulties in solving the above problems are as follows:
[0022] 1) With the increasing coupling of urban parking systems and the dramatic increase in data volume, most cities in China are still in a stage of rapid development, and the iteration of urban parking zoning is extremely fast. Therefore, there is a lack of a method and system for efficiently and quickly analyzing urban parking zoning;
[0023] 2) As urban parking becomes more digital and intelligent, and sensors and devices that collect data become more and more popular, the number of data dimensions used for analysis increases, and the validity of the data needs to be clarified;
[0024] 3) It is difficult to achieve differentiated management of parking zones based on the macro-dimension of administrative divisions; and it is difficult to quickly transform parking zones based on meso- and micro-level traffic communities into highly practical parking zones through standardized measures. Summary of the invention
[0025] In order to solve the defects and deficiencies in the above-mentioned prior art, this paper proposes a city parking zoning method and zoning system based on multi-source big data. The present invention provides a city parking zoning method and zoning system based on multi-source big data. First, multi-source big data related to urban parking and the boundaries of urban traffic zones are imported, and various types of data are counted on the traffic zones; secondly, effective data feature dimensions are identified through correlation analysis to complete data dimension screening; then, a weighted improved K-means clustering algorithm is used to identify the clusters to which the traffic zones belong, and an agglomerative clustering algorithm is used to merge similar clusters, so that the number of cluster types meets industry standards or related research; finally, parking zones are constructed using road networks, administrative boundaries and other elements, and the clustering attributes of traffic zones are mapped thereon, thereby obtaining parking zones with zoning attributes and strong operability. This zoning method and zoning system can solve urban parking zoning for a variety of application scenarios, and it is scientific, reasonable, efficient and feasible.
[0026] The technical solution of the present invention is as follows: A method for urban parking zoning based on multi-source big data, the main steps of which include:
[0027] 1) Data import;
[0028] 2) Data set accounting;
[0029] 2.1) Complete the cleaning of outliers, missing values, and duplicate values in the data set;
[0030] 2.2) The big data corresponding to the main influencing factors of the partition object are calculated on each traffic zone;
[0031] 3) Dimension screening
[0032] 3.1) Calculate the correlation coefficient matrix between the influencing factors, clarify the degree of correlation between the influencing factors of the partition object, and identify the degree of influence of each influencing factor on the partition result;
[0033] 3.2) After completing the calculation of the correlation coefficient of the influencing factors, retain the influencing factors with an absolute value of the correlation coefficient less than 0.5 as valid influencing factors, and update the total number of influencing factors M after screening;
[0034] 4) Traffic community clustering
[0035] 4.1) Traffic community clustering: Introducing the weight of influencing factors, the optimal clustering objective function of traffic communities is constructed as follows:
[0036]
[0037] In the formula, k is the cluster center of the cluster, u mn is the m-dimensional influencing factor value of traffic area n, c mk represents the m-dimensional influencing factor value of cluster center k, w j is the weight of the m-dimensional influencing factor;
[0038] 4.2) Using the improved weighted k-means algorithm, calculate the optimal clustering objective function value for each case within the range of k∈[3,15], use the elbow method to determine the optimal k value, and obtain the preliminary clustering of traffic communities under the optimal clustering condition;
[0039] 4.3) Agglomerative clustering: Reduce the number of clusters to the number of clusters specified by the industry practice or specification to which the partition object belongs; calculate the average value of each influencing factor attribute of each preliminary cluster, use the agglomerative clustering function of SPSS to obtain the dendrogram of agglomerative clustering, determine the clustering results after the number of clusters is reduced, and update the final cluster to which each traffic zone belongs;
[0040] 5) Parking zone network structure, including:
[0041] 5.1) Select linear features with high management relevance to the zoning object, including administrative district boundaries, municipal road networks, and river systems, import them into ArcGIS and merge them into the same layer;
[0042] 5.2) Construct a topological network in ArcGIS to check and process the problem nodes; load the topological toolbar, and based on the initial parking partition network, convert the initial parking partition network into a topological network according to the addition rules that there should be no hanging points and no pseudo nodes;
[0043] 5.3) Enable topological rule diagnosis for the topological network, query the hanging points and pseudo nodes, and superimpose the actual road network map to perform trimming, extension and splicing correction operations to obtain a parking partition network composed of closed polygon networks;
[0044] 6) Cluster mapping, including:
[0045] 6.1) Load the traffic zone boundaries with the final clustering attributes in ArcGIS, and perform the intersection operation on the two layers with the revised parking zone network, and decompose the traffic zone and fill it into each parking zone;
[0046] 6.2) Calculate the clustering attribute of each parking partition, take the decomposed traffic zone it contains as the calculation unit, and use the area ratio of the parking partition it occupies as the weight. Calculate the weighted average clustering attribute of the traffic zone and round it up. The cluster mapping formula is:
[0047]
[0048]
[0049] In the formula, the parking partition is represented by the symbol l, and there are L parking partitions in total. nl is the decision variable for whether the parking partition intersects with the traffic zone, a nl is the area where traffic zone n intersects with parking zone l; clu n is the final clustering attribute of parking partition l, A l is the area of parking partition l, and the urban parking partition is completed.
[0050] Preferably, 1) data import includes:
[0051] 1.1) Determine the research subject and partition object of parking partition, and import the big data set corresponding to the main influencing factors of the partition object into the partition system;
[0052] 1.2) Import the boundaries of the traffic zones within the study area into the zoning system.
[0053] Preferably, 2.2) the big data corresponding to the main influencing factors of the partition object are calculated on each traffic zone, specifically including:
[0054] Assume that the traffic zone of a city is represented by symbol n, there are N in total, and the area of the traffic zone is denoted by s n The point elements of the big data set related to urban parking are represented by the symbol i, and there are I in total; the surface elements are represented by j, and there are J in total. The area of the surface element is recorded as a j ; The influencing factors are represented by the symbol m, and there are M factors in total. The decision variable r of whether the point-surface element intersects with the traffic zone is introduced in ,r jn , the aggregation formula of point and surface elements is:
[0055]
[0056]
[0057] U m =(u m1 ,...,u mn ) T
[0058] In the formula, u mn is the m-dimensional influencing factor value of traffic area n, U m is the column vector of the m-th dimension influencing factors of all traffic areas, v mi is the m-dimensional influencing factor value of the i-th point element, v mj is the m-dimensional influencing factor value of the j-th surface element, ajn It is the area where the j-th polygon element intersects with traffic zone n.
[0059] Preferably, in 3.1), M is a set of influencing factors, assuming a, b∈M, a≠b, then the correlation coefficient calculation formula between any two influencing factors is:
[0060]
[0061] In the formula, u an is the a-dimensional influencing factor value of traffic community n, is the mean of the influencing factors in dimension a, u bn is the b-dimensional influencing factor value of traffic community n, is the mean of the b-dimensional influencing factors.
[0062] A parking zoning system for implementing the urban parking zoning method based on multi-source big data mainly includes:
[0063] Data import module: implement the above step 1);
[0064] Data collection module: implements the above step 2);
[0065] Dimension screening module: implements the above step 3);
[0066] Traffic community clustering module: implement the above step 4);
[0067] Parking partition network construction module: implement the above step 5);
[0068] Cluster mapping module: implements the above step 6).
[0069] A city parking zoning system terminal based on multi-source big data comprises a memory, a processor and at least one instruction or at least one computer program stored in the memory and loadable and run on the processor, wherein the processor loads and runs at least one instruction or at least one computer program to implement the city parking zoning method based on multi-source big data.
[0070] A non-transitory computer-readable storage medium stores a computer program, which, when executed by a processor, implements a method for urban parking zoning based on multi-source big data.
[0071] Compared with the prior art, the present invention has the following advantages:
[0072] 1) The present invention provides a city parking zoning method and zoning system based on big data. On the basis of obtaining big data related to urban traffic communities and urban parking, the city parking zoning based on the urban parking big data at this scale can be realized.
[0073] 2) The present invention identifies effective data analysis dimensions through correlation analysis to avoid interference from invalid or highly correlated data.
[0074] 3) The present invention utilizes existing linear elements such as road networks and administrative boundaries to construct a partition network, and maps the clustering attributes of traffic zones that cannot be directly applied to actual management onto the partition network, thereby ensuring the high operability of urban parking partitions.
[0075] This method efficiently and accurately constructs urban parking zoning with strong practicability. The zoning results can be quickly updated with the underlying data to cope with the rapid changes in the city. It provides technical support for the differentiated management of urban traffic and urban parking zoning, assists planning management and decision-making, and improves the scientific nature of urban traffic management. BRIEF DESCRIPTION OF THE DRAWINGS
[0076] Figure 1 A schematic diagram of a city parking zoning system based on multi-source big data in the present invention;
[0077] Figure 2 A schematic diagram of a process flow of an urban parking zoning method based on multi-source big data in the present invention;
[0078] Figure 3 This is a flow chart for clustering traffic zones;
[0079] Figure 4 To correct the schematic diagram of hanging points and pseudo nodes;
[0080] Figure 5 It is a schematic diagram of the overlay of the traffic zone and the parking zone;
[0081] Figure 6 Schematic diagram of the results of large data sets;
[0082] Figure 7 Schematic diagram of the elbow method for determining the optimal clustering K value;
[0083] Figure 8 is the average value of the attributes of each influencing factor in the initial partition;
[0084] Fig. 9 This is an agglomerative clustering dendrogram. DETAILED DESCRIPTION
[0085] See attached Figure 1 The present invention provides an urban parking zoning system based on multi-source big data, which mainly includes:
[0086] Data import module: After determining the research topic of parking zoning, import the big data set corresponding to the factors that have significant influencing effects on the left and right zoning effects and the boundaries of traffic zones into the system. Big data related to urban parking usually includes data on urban social economy, population occupation and residence distribution, land use, other transportation service levels, parking operations and supply and demand, etc. The above data are used as the basic data for urban parking zoning.
[0087] Data aggregation module: cleans the abnormal values, missing values and duplicate values of the data set, completes data preprocessing, and aggregates and integrates various large data sets into each traffic community. Urban traffic big data is generally expressed as points and surfaces in space. Through the intersection operation of geometric elements in GIS software, data aggregation and integration are realized.
[0088] Dimension screening module: After completing the aggregation of the large data set corresponding to the influencing factors into the traffic zone, in order to judge the effectiveness of the influencing factors and avoid the interference of invalid factors, the correlation coefficients between various types of data are calculated, the correlation analysis is completed, and the effective influencing factors with low correlation are retained as the data basis for subsequent partitioning.
[0089] Traffic zone clustering module: This module consists of two steps. Step 1 is traffic zone clustering, which takes the traffic zones after the aggregate fusion data as the clustering object, takes the attributes of the influencing factors carried by the traffic zones as the characteristic dimensions of clustering, and adopts the weighted improved K-means method to identify the preliminary clusters to which each traffic zone belongs; Step 2 is agglomerative clustering, the number of preliminary clusters of traffic zones is generally large, which cannot be directly applied to actual management and needs to be reduced. The agglomerative clustering algorithm is used to merge similar clusters to obtain the final clusters to which each traffic zone belongs.
[0090] Parking partition network construction module: Traffic zone boundaries are difficult to be directly applied to actual management, and a parking partition network suitable for practical application needs to be constructed separately. Therefore, linear elements such as administrative district boundaries and municipal road networks are selected to construct parking partition networks. By correcting pseudo nodes and hanging points in the network, it is ensured that the parking partition network consists of closed areas.
[0091] Cluster mapping module: The traffic zones with final zoning attributes are spatially superimposed with the parking zoning network, and the clustering attributes of the former are mapped to the latter, so that the zoning results are associated with the administrative boundaries and municipal road networks of the city, thereby constructing the city parking zoning that can eventually be used for actual management.
[0092] See attached Figure 2 The present invention provides an urban parking zoning method based on multi-source big data, the main steps of which include:
[0093] 1) Use the data import module to import urban parking-related big data and traffic communities;
[0094] 2) Using the data statistics module to realize the integration of multi-source big data into traffic communities;
[0095] 3) Use the dimension screening module to implement correlation analysis of urban parking-related big data and screen effective analysis dimensions;
[0096] 4) Use the traffic zone clustering module to realize traffic zone clustering;
[0097] 5) Using the parking partition network construction module to realize the parking partition construction based on the existing linear elements;
[0098] 6) The cluster mapping module is used to map the clustering attributes of traffic zones to parking zones to ensure the operability of the zones.
[0099] The goal of this method is to optimally partition the research subject of urban parking.
[0100] Each step is described in detail below with reference to the accompanying drawings.
[0101] 1. Data import
[0102] This step aims to import big data related to urban parking and traffic zone boundaries into the system. Big data usually refers to the massive data related to urban parking collected by various equipment sensors. This method and system assume that the data has been acquired, which is the data basis for carrying out parking zoning work. After determining the research subject and zoning object of parking zoning, combine industry specifications and related research to judge the main influencing factors of the zoning object and import the corresponding big data. The data generally include urban social economy, population occupation and residence distribution, land use, other transportation service levels, parking operation and supply and demand data. Among them, urban socio-economic data refers to regional output value, regional income, etc., population occupation and residence distribution refers to the population's residence and employment distribution, land use data refers to land price, land use type, land volume ratio, etc., other transportation service levels refer to the distribution and operation of facilities of various modes of transportation, and parking operation and supply and demand refer to the dynamic operation level and supply and demand level of parking lots in the region. The traffic zone is the basic analysis unit for studying traffic in traffic planning. Generally, traffic boundaries such as roads and railways, and natural barriers such as rivers are used as zoning boundaries. The socio-economic, traffic and travel characteristics of the same traffic zone are similar and homogeneous. The traffic planning departments of various cities have complete basic data on traffic zones, which can be used as a spatial carrier to finely characterize the differences in characteristics of urban traffic at the meso-micro level. It is the basic analysis unit for urban parking zoning.
[0103] 2 Dataset
[0104] This step aims to preprocess the imported large data set, clean the data set of abnormal values, missing values, and duplicate values, and aggregate and fuse them into each traffic zone. The spatial form of big data is generally expressed as points and surfaces. Point elements are matched to each traffic zone through the point-surface intersection operation of GIS software, and the data carried by the point elements are counted into each traffic zone through the fusion operation; surface elements are decomposed into each traffic zone through the surface-surface intersection operation of GIS software, and the data carried by the surface elements are counted into each traffic zone through the fusion operation.
[0105] Assume that the traffic zone of a city is represented by symbol n, there are N in total, and the area of the traffic zone is denoted by s n ; The point element of the large data set is represented by the symbol i, and there are I in total; the area element is represented by j, and there are J in total. The area of the area element is recorded as a j ; The influencing factors are represented by the symbol m, and there are M factors in total. The decision variable r of whether the point-surface element intersects with the traffic zone is introduced in ,r jn , the calculation formula of the data set is as follows:
[0106]
[0107] U m =(u m1 ,...,u mn ) T Formula 3
[0108] In the formula, u mn is the m-dimensional influencing factor value of traffic area n, U m is the column vector of the m-th dimension influencing factors of all traffic areas, v mi is the m-dimensional influencing factor value of the i-th point element, v mj is the m-dimensional influencing factor value of the j-th surface element, a jn It is the area where the j-th polygon element intersects with traffic zone n.
[0109] 3-dimensional screening
[0110] This step aims to screen out the effective influencing factors of the urban traffic big data set aggregated into traffic zones. The partition object is usually affected by multiple factors. If the influencing factors with high correlation are used as basic data, it will interfere with the partition result. The column vectors of each influencing factor at the traffic zone level have been calculated in the previous step. By calculating the correlation coefficient matrix between the column vectors corresponding to the influencing factors, the correlation between the influencing factors of the partition object is clarified, and the importance of each influencing factor to the partition result is further judged. Those with low correlation are retained, thereby optimizing the partition algorithm and reducing the amount of calculation.
[0111] M is a set of influencing factors. Assuming a, b∈M, a≠b, the correlation coefficient between any two influencing factors is calculated as follows:
[0112]
[0113] In the formula, u an is the a-dimensional influencing factor value of traffic community n, is the mean of the influencing factors in dimension a, u bn is the b-dimensional influencing factor value of traffic community n, is the mean of the b-dimensional influencing factors. After completing the calculation of the pairwise correlation coefficients of the influencing factors, the influencing factors with an absolute value of the correlation coefficient less than 0.5 are retained as valid influencing factors, and the total number of influencing factors M after screening is updated.
[0114] 4 Traffic Community Clustering
[0115] Urban parking zoning is a systematic and complex project. The essence of zoning is to divide traffic areas with high similarity in urban parking attribute characteristics into the same zone, that is, the essence of zoning is to cluster traffic areas. This step includes two sub-steps: preliminary clustering of traffic areas and agglomerative clustering.
[0116] Step 1: The data set corresponding to the screened effective influencing factors is used as the data chassis. The data set fully reflects the impact of urban parking-related data on the zoning. According to the industry standards and related research of the zoning object, the weight of the influencing factor is introduced to characterize the importance of each influencing factor. Each traffic zone is used as a clustering sample, and the minimum weighted square norm (SSE) between each sample point and the center of the cluster to which it belongs is taken as the optimization goal. The optimal clustering objective function of the traffic zone is as follows. The clustering process is attached. Figure 3 .
[0117]
[0118] In the formula, k is the cluster center of the cluster, u mn is the m-dimensional influencing factor value of traffic area n, c mk represents the m-dimensional influencing factor value of cluster center k, w j is the weight of the m-dimensional influencing factor.
[0119] Since the number of clusters k of urban parking partitions cannot be determined in advance, according to relevant research and actual management, k∈[3,15] is given, and the objective function value SSE under various cluster numbers is calculated. The optimal k value is determined by the elbow method, so as to obtain the preliminary clustering of traffic zones under the optimal clustering condition. The preliminary cluster to which traffic zone n belongs is clu n , clu n ∈[1,k], the optimal clustering algorithm process of traffic cells is as follows Figure 3 shown.
[0120] Step 2: The number of preliminary clusters of traffic zones is generally large and cannot be directly applied to actual management. It is necessary to use an agglomerative clustering algorithm to merge similar clusters to further reduce the number of clusters to the number of clusters specified by the industry practice or specification of the zone object, thereby obtaining the final clustering attributes of each traffic zone. The specific process is as follows:
[0121] (1) Taking the traffic communities included in each preliminary cluster as a whole, calculate the average value of the attributes of each influencing factor;
[0122] (2) Import each preliminary cluster and the corresponding influencing factor means into SPSS software;
[0123] (3) Select the menu [Analysis → Classification → System Clustering], select the influencing factors involved in the agglomerative clustering analysis in the [Variable] box, select the clustering type as agglomerative clustering in the [Clustering] box, click OK, and judge the clustering results after the number of clusters is reduced based on the dendrogram of agglomerative clustering, and update the final clustering clu of each traffic zone. n .
[0124] 5. Parking zone network structure
[0125] This step aims to construct a parking partition network. Since the boundaries of traffic zones cannot be directly applied to actual management, after completing the clustering of traffic zones, it is necessary to construct a boundary that is connected to the actual management elements, namely the parking partition network. According to the research subject and partition object of urban parking, linear elements such as administrative district boundaries, municipal road networks, and river systems that are highly relevant to their research or management are selected. The closed polygonal network formed by the interweaving of the above elements is the parking partition network. This network can serve as a transitional carrier for the clustered traffic zones to extend to administrative districts, municipal roads, etc. The specific steps for constructing the parking partition network are as follows:
[0126] (1) Merge linear features: Combine the administrative boundaries of the research area, import different levels of municipal single-line networks, river systems and other vector files suitable for zoning objects into ArcGIS, and use the merge operation to merge the linear features of each layer into the same layer as the initial parking zoning network;
[0127] (2) Construct a topological network: Since the linear layers have breakpoints and other situations, the initial network obtained is not completely composed of closed polygons. Therefore, a topological network should be constructed to check and deal with problematic nodes. Load the topological toolbar and convert the parking partition initial network into a topological network based on the rules that there should be no hanging points or pseudo nodes added.
[0128] (3) Correcting the topological network: Enable topological rule diagnosis on the topological network, query the hanging points and pseudo nodes, and overlay the actual road network map to perform trimming, extension and splicing correction operations to ensure that the final parking partition network composed of closed polygons is obtained, as shown in the attached figure. Figure 4 shown.
[0129] 6 Cluster Mapping
[0130] This step aims to map the clustering attributes of the traffic zone to the parking zone network to obtain the final urban parking zone. Open ArcGIS and load the traffic zone boundaries with the final clustering attributes and the revised parking zone network. The parking zone is represented by the symbol l, and there are L in total. The decision variable r is introduced to determine whether the parking zone intersects with the traffic zone. nl . Perform the intersection operation of the two layers, decompose the traffic zone and fill it into each parking zone, thereby mapping the final clustering attributes of the traffic zone to the parking zone. Calculate the clustering attributes of each parking zone, take the decomposed traffic zone it contains as the calculation unit, use its area as the weight, calculate the weighted average clustering attribute of the traffic zone and round it up, the calculation formula is as follows:
[0131]
[0132] In the formula, a nl is the area where the traffic zone n and the parking zone l intersect, as shown in the attached figure. Figure 5 As shown. l is the area of parking partition l, Clu l is the final clustering attribute of parking partition l. So far, the urban parking partition is completed.
[0133] The following is a study of six districts in the city center (A, B, C, D, E, and F). The above-mentioned urban parking zoning method and zoning system are used to demonstrate the expected effect by taking the parking construction index zoning as an example. The six steps and results of the model are as follows:
[0134] (1) Data import
[0135] The research subject and zoning object of this example is parking construction zoning, the purpose is to formulate differentiated parking facility supply strategies according to different functional spaces in the city. According to the "Guidelines for Urban Parking Facility Planning" issued by the Ministry of Housing and Urban-Rural Development, parking zoning involves many factors such as population, land, and transportation. Therefore, five influencing factors, including population density, land use volume ratio, road congestion, rail station coverage, and parking space demand-supply ratio, are initially selected to collect and organize the corresponding big data sets. The 2,374 traffic community boundaries included in the research scope and the big data sets corresponding to the five influencing factors are imported into the zoning system.
[0136] (2) Dataset
[0137] In the initial big data set, population density and parking space demand-supply ratio are point elements, while land use volume ratio, road congestion and rail station coverage are area elements. The data collection module can quickly collect big data into each traffic zone to form a data collection matrix with 2374 rows and 5 columns. Some matrices are shown in the attached figure. Figure 6 shown.
[0138] (3) Dimension screening
[0139] According to the dimension screening steps, the correlation coefficients between the five influencing factors are calculated to form a correlation coefficient matrix. It can be seen that the absolute value of the correlation coefficient of each influencing factor is less than 0.5, and it is believed that the influencing factors initially selected are all effective.
[0140] (4) Traffic community clustering
[0141] Step 1: The research subject of the example is parking zoning, giving a greater weight to the parking space demand-supply ratio, using the weighted k-means algorithm, combined with the elbow method to determine the optimal clustering K value is 6 (see attached) Figure 7 , the traffic areas with parking facilities as the main body are initially divided into 6 categories, and the clustering results corresponding to the clustering K value are retrieved and assigned to each traffic area.
[0142] Step 2: According to the "Guidelines for Urban Parking Facility Planning" issued by the Ministry of Housing and Urban-Rural Development, parking zones are generally divided into three categories: strictly restricted areas, generally restricted areas, and moderately developed areas. The agglomerative clustering algorithm is used to reduce the number of preliminary clusters to three categories. The average values of the attributes of each influencing factor of the six preliminary clustering zones are calculated. The results are shown in the attached Figure 8 After the agglomerative clustering operation of SPSS, the number of clusters to which the traffic community belongs is reduced from 6 to 3, see the attached Fig. 9 .
[0143] (5) Parking zone network structure
[0144] The parking partition network was constructed using the road network above the secondary trunk road level and the administrative district boundaries within the research scope. The pseudo nodes and hanging points were corrected in ArcGIS. The parking partition network finally contained 439 closed polygons.
[0145] (6) Cluster Mapping
[0146] The traffic subdistrict with the final clustering attributes is superimposed on the parking partition network constructed in the fifth step, and the intersection operation is performed using ArcGIS to obtain 5099 decomposed traffic subdistricts. The clustering attributes of the decomposed traffic subdistrict are transferred to the parking partition using the cluster mapping formula to obtain the final three types of parking configuration partitions.
[0147] Summary: The final generated parking zoning space is as follows: Category I areas are mainly distributed along both sides of Zhongshan Avenue and Huangpu Avenue, from Kangwang Road in the west to Machang Road in the east, basically covering the areas with the best public transportation services, the highest openness, and the most tense parking supply and demand in Guangzhou within the research scope; Category II areas expand the management scope outward, from Baiyun New Town in the north to the west side of the Inner Ring Road, Fangcun, and Xilang in the west, to Industrial Avenue in the south, and to Pazhou West in Area C and Chebei in Area D in the east, basically covering the old city areas with high built-up density and the new areas with high development intensity under construction, with greater parking supply pressure; the remaining areas are Category III areas, basically covering the areas with low public transportation service levels and good parking supply and demand levels within the research scope.
[0148] The parking zoning obtained by using the urban parking zoning method and zoning system comprehensively considers multiple factors such as the city's built-up level, population density, public transportation service level, car operation level, parking supply and demand level, etc. The zoning method and system can quickly, scientifically and accurately identify the clusters to which traffic communities belong from big data, and the results are in line with reality. By constructing a parking zoning network, the clustering attributes of traffic communities are mapped to parking zoning composed of road networks and administrative district boundaries. Managers can directly use the zoning results to assist in their decision-making in the entire process of planning, construction, operation and management. The method is highly versatile and can be applied to urban parking zoning in a variety of scenarios, with strong operability and practical significance.
[0149] Compared with the prior art, the present invention has the following advantages:
[0150] 1) The present invention is a method and system for urban parking zoning based on big data, which can realize urban parking zoning based on multi-source big data related to the zoning object, using an improved weighted k-means algorithm and agglomerative clustering algorithm, and based on the optimal clustering identification of urban traffic zones.
[0151] 2) The present invention is a big data-based urban parking zoning method and zoning system, which can build a parking zoning network based on any linear elements related to the study or management of zoning objects, such as administrative district boundaries and municipal road networks. The optimal clustering of traffic zones is mapped to the parking zoning network using a cluster mapping module.
[0152] 3) The present invention is a city parking zoning method and zoning system based on big data, which can efficiently and accurately construct urban parking zoning that is directly used for management. The zoning results can be quickly updated with the iteration of basic data to cope with the rapid development of urban changes, assist managers in decision-making, and diagnose the boundaries of urban parking problems. In the era of big data, scientific methods are used to truly realize differentiated zoning management and improve the scientific nature of urban traffic management.
[0153] The above is only a preferred specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.
Claims
1. A method for urban parking zoning based on multi-source big data, characterized in that: The main steps include: 1) Data import; 2) Data set accounting; 2.1) Complete the cleaning of outliers, missing values, and duplicate values in the data set; 2.2) The big data corresponding to the main influencing factors of the partition objects are calculated on each traffic zone; 3) Dimension screening 3.1) Calculate the correlation coefficient matrix between the influencing factors, clarify the degree of correlation between the influencing factors of the partition object, and identify the degree of influence of each influencing factor on the partition result; 3.2) After completing the calculation of the correlation coefficient of the influencing factors, retain the influencing factors with an absolute value of the correlation coefficient less than 0.5 as valid influencing factors, and update the total number of influencing factors M after screening; 4) Traffic community clustering 4.1) Traffic community clustering: Introducing the weight of influencing factors, the optimal clustering objective function of traffic communities is constructed as follows: In the formula, k is the cluster center of the cluster, u mn is the m-dimensional influencing factor value of traffic area n, c mk represents the m-dimensional influencing factor value of cluster center k, w j is the weight of the m-dimensional influencing factor; 4.2) Using the improved weighted k-means algorithm, calculate the optimal clustering objective function value for each case within the range of k∈[3,15], use the elbow method to determine the optimal k value, and obtain the preliminary clustering of traffic communities under the optimal clustering condition; 4.3) Agglomerative clustering: Reduce the number of clusters to the number of clusters specified by the industry practice or specification to which the partition object belongs; calculate the average value of each influencing factor attribute of each preliminary cluster, use the agglomerative clustering function of SPSS to obtain the dendrogram of agglomerative clustering, determine the clustering results after the number of clusters is reduced, and update the final cluster to which each traffic zone belongs; 5) Parking zone network structure, including: 5.1) Select linear features with high management relevance to the zoning object, including administrative district boundaries, municipal road networks, and river systems, import them into ArcGIS and merge them into the same layer; 5.2) Construct a topological network in ArcGIS to check and process the problem nodes; load the topological toolbar, and based on the initial parking partition network, convert the initial parking partition network into a topological network according to the addition rules that there should be no hanging points and no pseudo nodes; 5.3) Enable topological rule diagnosis for the topological network, query the hanging points and pseudo nodes, and superimpose the actual road network map to perform trimming, extension and splicing correction operations to obtain a parking partition network composed of closed polygon networks; 6) Cluster mapping, including: 6.1) Load the traffic zone boundaries with the final clustering attributes in ArcGIS, and perform the intersection operation on the two layers with the revised parking zone network, and decompose the traffic zone and fill it into each parking zone; 6.2) Calculate the clustering attribute of each parking partition, take the decomposed traffic zone it contains as the calculation unit, and use the area ratio of the parking partition it occupies as the weight. Calculate the weighted average clustering attribute of the traffic zone and round it up. The cluster mapping formula is: In the formula, the parking partition is represented by the symbol l, and there are L parking partitions in total. nl is the decision variable for whether the parking partition intersects with the traffic zone, a nl is the area where traffic zone n intersects with parking zone l; clu n is the final clustering attribute of parking partition l, A l is the area of parking partition l, and the urban parking partition is completed.
2. The urban parking zoning method based on multi-source big data according to claim 1 is characterized in that: 1) Data import includes: 1.1) Determine the research subject and partition object of parking partition, and import the big data set corresponding to the main influencing factors of the partition object into the partition system; 1.2) Import the boundaries of the traffic zones within the study area into the zoning system.
3. The urban parking zoning method based on multi-source big data according to claim 1 is characterized in that: 2.2) The big data corresponding to the main influencing factors of the zoning object are calculated on each traffic zone, including: Assume that the traffic zone of a city is represented by symbol n, there are N in total, and the area of the traffic zone is denoted by s n The point elements of the big data set related to urban parking are represented by the symbol i, and there are I in total; the surface elements are represented by j, and there are J in total. The area of the surface element is recorded as a j ; The influencing factors are represented by the symbol m, and there are M factors in total. The decision variable r of whether the point-surface element intersects with the traffic zone is introduced in ,r jn , the aggregation formula of point and surface elements is: IN m =(in m1 ,...,in mn ) T In the formula, u mn is the m-dimensional influencing factor value of traffic area n, U m is the column vector of the m-th dimension influencing factors of all traffic areas, v mi is the m-dimensional influencing factor value of the i-th point element, v mj is the m-dimensional influencing factor value of the j-th surface element, a jn It is the area where the j-th polygon element intersects with traffic zone n.
4. The urban parking zoning method based on multi-source big data according to claim 1 is characterized in that: 3.1), M is the set of influencing factors. Assume a, b∈M, a≠b, then the correlation coefficient between any two influencing factors is calculated as: In the formula, u an is the a-dimensional influencing factor value of traffic community n, is the mean of the influencing factors in dimension a, u bn is the b-dimensional influencing factor value of traffic community n, is the mean of the b-dimensional influencing factors.
5. A parking partitioning system for implementing the urban parking partitioning method based on multi-source big data as described in any one of claims 1 to 4, characterized in that: Mainly include: Data import module: implement the above step 1); Data collection module: implements the above step 2); Dimension screening module: implements the above step 3); Traffic community clustering module: implement the above step 4); Parking partition network construction module: implement the above step 5); Cluster mapping module: implements the above step 6).
6. An urban parking partition system terminal based on multi-source big data, comprising a memory, a processor, and at least one instruction or at least one computer program stored in the memory and loadable and run on the processor, characterized in that: The processor loads and runs at least one instruction or at least one computer program to implement an urban parking zoning method based on multi-source big data as described in any one of claims 1-4.
7. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, an urban parking zoning method based on multi-source big data as described in any one of claims 1 to 4 is implemented.
