Crowdsourcing acquisition and freshness maintenance system for self-updating spatial database

Abstract

The invention relates to the technical field of data updating, in particular to a crowdsourcing collecting and freshness maintaining system of a self-updating spatial database. According to the system, a to-be-selected updating node corresponding to a maximum position matching characteristic value is used as an updating node; calculating a track change characteristic value between the historical road tracks and the road tracks, and judging whether the historical road condition information is updated or not based on the track change characteristic value; performing interactive comparison processing on the updated image features and the historical image features to obtain a building change index, and judging whether to update the historical building image information based on the building change index; and the historical road condition information and the historical building image information in the spatial database are updated, so that the updating efficiency of the data in the spatial database can be improved and the updating hysteresis can be reduced.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data updating, in particular to a crowdsourcing collection and freshness maintenance system for self-updating spatial database. BACKGROUND

[0002] As the core infrastructure of digital city, intelligent transportation, environmental monitoring and location-based services, the accuracy and freshness of spatial database directly determine the service quality and reliability of upper-layer applications. Crowdsourcing data collection mode is a data collection mode in which a large, undefined group is recruited to complete the data collection task that was traditionally performed by internal staff.

[0003] Limitations of existing spatial data updating mode: on the one hand, the cost of professional collection is high, relying on professional surveying and mapping team to collect data, which is time-consuming and laborious, with high economic cost and long update cycle, and cannot meet the real-time requirements of modern applications; on the other hand, the data updating is lagging, and the road condition information and building image information are not updated in time.

[0004] Therefore, there is an urgent need for a crowdsourcing collection and freshness maintenance system for self-updating spatial database to solve the above problems. SUMMARY

[0005] The purpose of the present application is to provide a crowdsourcing collection and freshness maintenance system for self-updating spatial database, which solves the technical problems of high data updating cost and high data updating lag in the existing spatial data updating mode.

[0006] The crowdsourcing collection and freshness maintenance system for self-updating spatial database comprises: The crowdsourcing collection unit is configured to collect the to-be-updated data uploaded by the selected participants who have passed legal authentication based on the crowdsourcing platform, wherein the to-be-updated data comprises location information, road condition information and building image information of the selected participants; The data matching unit is configured to calculate the location matching feature value of the location information of the selected participants and each to-be-updated node in the spatial database, and take the to-be-updated node corresponding to the maximum location matching feature value as the update node; The road analysis unit is configured to determine the historical road condition information of the update node based on the road condition information and the update node corresponding to the road condition information, track the road trajectory of the historical road condition information and the road condition information respectively, obtain the historical road trajectory and the road trajectory, calculate the trajectory change feature value between the historical road trajectory and the road trajectory, and determine whether to update the historical road condition information based on the trajectory change feature value; The building analysis unit is configured to respectively perform feature extraction on the building image information and the historical building image information in the spatial database to obtain updated image features and historical image features, and to perform interactive comparison processing on the updated image features and the historical image features to obtain a building change index, and to determine whether to update the historical building image information based on the building change index. The data updating unit is configured to update the historical road condition information and the historical building image information in the spatial database.

[0007] Further, the calculation of the position matching feature value of the location information of the selected participant and each candidate updating node in the spatial database includes the following process: Based on the location information of the selected participant, the coordinates of the mobile terminal of the selected participant are obtained ; Based on the coordinates , the azimuth of the selected participant pointing to the candidate updating node is calculated using an inverse tangent function : The calculated azimuth is converted into degrees , wherein the coordinates of the candidate updating node ; Based on the coordinates , the distance between the selected participant and the candidate updating node is calculated using the Euclidean distance formula ; The position proximity of the location information of the selected participant and the candidate updating node is calculated The degrees corresponding to the azimuth , the distance , and the position proximity are de-dimensioned to obtain their numerical values, which are substituted into the position matching feature value calculation formula to calculate the position matching feature value WZT, and the position matching feature value calculation formula is: ; wherein the value of c is , the number of candidate updating nodes in the spatial database.

[0008] Further, the calculation of the position proximity of the location information of the selected participant and the candidate updating node includes the following process: Based on the location information of the selected participant, the IP address information, node code information and geographic location information of the selected participant are selected. The Conditional Random Field (CRF) model is used to extract entity references from the IP address information, node code information and geographic location information, and the entity references are converted into vector matrices based on the bag-of-words model. Calculate vector matrix and nodes to be updated The modularity vector between the corresponding node vector matrices : ; Where K is the number of vectors in the vector matrix, represents the adjacency matrix of the vector matrix and the node vector matrix, Represents the node number of the adjacency matrix, 2P represents the adjacency matrix The sum of all edges in , represents the degree of i, represents the degree of j, represents the community in which node i exists, represents the community in which node j exists, Is the conditional restriction coefficient, if the vector matrix and the node to be updated are The node encoding lengths of the corresponding node vector matrices are different, then Is 0, if the vector matrix and the node to be updated The node encoding lengths of the corresponding node vector matrices are the same, then is 1; Modularity vector The modulus value is recorded as the location information of the selected participant and the update node to be selected Position proximity .

[0009] Furthermore, determining the historical road condition information of the update node based on the road condition information and the update node corresponding to the road condition information specifically includes the following process: The road condition information and the update nodes corresponding to the road condition information are input into the deep convolutional neural network model, and the historical road condition information of the update nodes is output.

[0010] Furthermore, the historical road condition information and the road condition information are tracked to obtain the historical road trajectory and the road trajectory, which specifically includes the following processes: Step 1: Based on the historical road condition information or road condition information, obtain the corresponding historical road route or road route, set Set the same sampling points for the combined parameters, historical road routes or road routes, and The sampling points are combined according to the sampling order of the sampling interval values to obtain a plurality of combination groups; Step two: the first sampling point and the second sampling point of each combination group form a reference vector m, and the remaining zth sampling point and the z-1th sampling point form a vector , wherein, , the angle between is calculated respectively , wherein the angle is calculated as follows: ; Step three: compare with the threshold angle , if the angle of the zth point is greater than the threshold angle , the sampling points arranged in front of the zth sampling point are subdivided into a group, and the sampling points arranged in front of the zth point are repeated step two and step three to complete the subdivision of the combination group, and the sampling points in each subdivided group are connected to obtain the historical road trajectory and the road trajectory.

[0011] Further, the calculation of the trajectory change characteristic value between the historical road trajectory and the road trajectory includes the following process: The length of the trajectory segment curve of the historical road trajectory and the road trajectory is counted, and the lengths of the trajectory curves corresponding to the historical road trajectory and the road trajectory are counted respectively, the maximum trajectory curve length is determined, and the ratio between the length of the trajectory segment curve and the maximum trajectory curve length is calculated. The ratio is recorded as the trajectory change characteristic value.

[0012] Further, the judgment of whether to update the historical road condition information based on the trajectory change characteristic value includes the following process: Load the trajectory change characteristic value threshold, judge whether the trajectory change characteristic value exceeds the trajectory change characteristic value threshold, if yes, determine that the historical road condition information needs to be updated, if no, determine that the historical road condition information does not need to be updated, wherein the trajectory change characteristic value threshold is set by the system and stored in the system.

[0013] Further, the feature extraction of the building image information and the historical building image information in the spatial database obtains the updated image feature and the historical image feature, which includes the following process: The building image information and the historical building image information are respectively input into the Transformer encoding layer for feature extraction to obtain the updated image feature and the historical image feature , wherein, is the length of the feature respectively, is the dimension of the feature, and R is a set of real numbers.

[0014] Furthermore, interactive comparison processing is performed on the updated image features and the historical image features to obtain the building change index, which specifically includes the following process: based on Generate graph structure , where V is the node set representing the target building feature information in the updated image feature, and the node set , is the number of nodes, E is the edge set representing the characteristic information of the target building in the historical image features, , the number of edges is A, where two adjacent nodes Constituent edge ; Edge centrality is defined as the average of the centrality scores of two adjacent nodes as follows: ; in, For the edge The centrality is used to express the influence of the importance between two connected nodes. represents the node centrality measure; Calculate the probability of each edge based on the centrality of each edge: Set To reduce the impact of dense nodes, after normalization, the interaction probability between the two is calculated : ; in, is a hyperparameter whose purpose is to control the overall probability of interaction; yes The maximum value of yes The average value of , used to limit the maximum value of probability; Calculate the building change index: ; Among them, the greater the interaction probability, the more similar the updated image features are to the historical image features, and the smaller the change in the building.

[0015] Furthermore, judging whether to update the historical building image information based on the building change index specifically includes the following process: Load the building change index threshold and determine whether the building change index exceeds the building change index threshold. If so, determine that the historical building image information needs to be updated. If not, determine that the historical building image information does not need to be updated. The building change index threshold is set by the system and stored in the system.

[0016] Compared with the prior art, the application has the beneficial effects of: In one aspect, the application can intelligently and automatically associate the crowd-sourced data with the most relevant candidate update nodes in the database by calculating accurate position matching feature values through the data matching unit, thereby avoiding data mismatch from the source, ensuring the correctness of the subsequent analysis processing object, greatly reducing the cost of manual review and error correction, and improving the update efficiency and accuracy. In another aspect, the road analysis unit can sensitively capture substantial changes in road topology and direction (such as new road opening, one-way adjustment, road closure, etc.) by comparing real-time road trajectories with historical trajectories and calculating trajectory change feature values; the building analysis unit can effectively identify complex changes such as new building, demolition, and facade reconstruction of buildings by deep feature extraction and interactive comparison, and generate a building change index, based on a freshness judgment mechanism of multi-dimensional feature analysis, so that the system can accurately identify which building information needs to be updated, greatly improving the purpose and value density of the update. Finally, data updating based on the crowd-sourcing platform can reduce data updating cost and data updating lag. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description are only some embodiments described in the present application, and other drawings can also be obtained by those skilled in the art based on these drawings.

[0018] Figure 1 is a system block diagram of a crowd-sourcing collection and freshness maintenance system of a self-updating spatial database according to an embodiment of the present application; Figure 2 is a workflow diagram of a crowd-sourcing collection and freshness maintenance system of a self-updating spatial database according to an embodiment of the present application; Figure 3 is a structural block diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0019] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of the present application.

[0020] Moreover, the described features, structures, or characteristics can be combined in any suitable manner in one or more example embodiments. In the following description, numerous specific details are provided to give a thorough understanding of example embodiments of the disclosure. One skilled in the relevant art will recognize, however, that the techniques of the disclosure can be practiced without one or more of the specific details, or

[0021] The embodiment provides a crowdsourcing collection and freshness maintenance system of a self-updating spatial database, Figure 1 is a system block diagram of a crowdsourcing collection and freshness maintenance system of a self-updating spatial database according to an embodiment of the present application, as shown in the figure, the system comprises: Figure 1 The crowdsourcing collection unit is configured to collect to-be-updated data uploaded by selected participants who pass legal authentication based on a crowdsourcing platform, wherein the to-be-updated data comprises location information of the selected participants, road condition information and building image information. It is worth noting that the selected participants who pass legal authentication can upload the to-be-updated data to the crowdsourcing platform through a mobile terminal or a computer terminal, wherein the crowdsourcing platform is a digital intermediary or an online market, which decomposes, distributes and outsources a task or a project of an organization (a contractor) to a large, undefined network group (contributors or "the public") to complete through an open call form, and coordinates the work, management and payment of rewards.

[0022] The data matching unit is configured to calculate a location matching feature value of the location information of the selected participants and each to-be-updated node in the spatial database, and take a to-be-updated node corresponding to a maximum location matching feature value as an updated node. The road analysis unit is configured to determine historical road condition information of the updated node based on the road condition information and the updated node corresponding to the road condition information, perform road track tracking on the historical road condition information and the road condition information respectively, obtain historical road track and road track, calculate a track change feature value between the historical road track and the road track, and judge whether to update the historical road condition information based on the track change feature value. The building analysis unit is configured to perform feature extraction on the building image information and historical building image information in the spatial database respectively, obtain updated image features and historical image features, perform interactive comparison processing on the updated image features and the historical image features, obtain a building change index, and judge whether to update the historical building image information based on the building change index. The data updating unit is configured to update the historical road condition information and the historical building image information in the spatial database.​

[0023] In summary, the present invention calculates precise location matching feature values ​​through the data matching unit, and can intelligently and automatically associate crowdsourced data with the most relevant candidate update nodes in the database, avoiding data mismatching from the source, ensuring the correctness of subsequent analysis and processing objects, greatly reducing the cost of manual review and error correction, and improving update efficiency and accuracy; the road analysis unit calculates trajectory change feature values ​​by comparing real-time road trajectories with historical trajectories, and can keenly capture substantial changes in road topology and direction (such as new road openings, one-way street adjustments, road closures, etc.); the building analysis unit generates a building change index through deep feature extraction and interactive comparison, and can effectively distinguish complex changes such as new construction, demolition, and facade renovation of buildings. The freshness judgment mechanism based on multi-dimensional feature analysis enables the system to accurately identify which building information needs to be updated, greatly improving the purposefulness and value density of the update; updating data in the spatial database based on the crowdsourcing platform can reduce data update costs and data update lags.

[0024] In some embodiments, calculating the location information of the selected participant and the location matching feature value of each candidate update node in the spatial database specifically includes the following process: Get the coordinates of the selected participant's mobile terminal based on the location information of the selected participant ; The location information of the selected participant is collected through a positioning system, and the coordinates of the selected participant's mobile terminal are obtained; Based on coordinates Use the inverse tangent function to calculate the selected participant's point to the update node to be selected Azimuth : , the calculated azimuth Convert to degrees , where the node to be updated is Coordinates ; It is worth noting that within a specific interval, the selected participant points to the update node to be selected Azimuth The larger the value, the more the selected participants and the candidate update nodes The closer the distance, the higher the matching degree.

[0025] Based on coordinates Use the Euclidean distance formula to calculate the selected participants and the candidate update nodes distance : ; It is worth noting that the selected participants and the candidate update nodes distance The larger the value, the more the selected participants and the candidate update nodes The farther the distance, the lower the matching degree; Calculate the location information of the selected participants and the update nodes to be selected Position proximity ; Among them, based on the location information of the selected participant, the IP address information, node code information and geographic location information of the selected participant are selected. The CRF model is used to extract entity references from the IP address information, node code information and geographic location information, and the entity references are converted into vector matrices based on the bag-of-words model: The conditional random field (CRF) model is a statistical modeling method that is very suitable for sequence labeling tasks. It can well consider contextual features and determine which words in the text belong to specific entity references; The bag-of-words model converts entity references into a vector matrix and constructs a vocabulary: all extracted entity values ​​(such as all different IPs, node IDs, and place names) are collected to form a vocabulary, and each word corresponds to a unique index. Generate a vector: For a data sample (such as a log record), create a vector based on the entities it contains. The length of the vector is equal to the size of the vocabulary. If a word appears in the sample, the corresponding index position in the vector is set to 1 (or the number of occurrences), otherwise it is 0. Form a matrix: Stack the vectors of all samples to form a huge, sparse vector matrix. Each row represents a sample, and each column represents an entity word.

[0026] Calculate vector matrix and nodes to be updated The modularity vector between the corresponding node vector matrices : ; Where K is the number of vectors in the vector matrix, represents the adjacency matrix of the vector matrix and the node vector matrix, Represents the node number of the adjacency matrix, 2P represents the adjacency matrix The sum of all edges in , represents the degree of i, represents the degree of j, represents the community in which node i exists, represents the community in which node j exists, Is the conditional restriction coefficient, if the vector matrix and the node to be updated are The node encoding lengths of the corresponding node vector matrices are different, then Is 0, if the vector matrix and the node to be updated The node encoding lengths of the corresponding node vector matrices are the same, then is 1; Modularity vector The modulus value is recorded as the location information of the selected participant and the update node to be selected Position proximity .

[0027] Furthermore, the azimuth Corresponding degree ,distance Proximity to location Remove the dimension and substitute its value into the position matching eigenvalue calculation formula to obtain the position matching eigenvalue WZT. The position matching eigenvalue calculation formula is: ; Among them, the value of c is , is the number of nodes to be updated in the spatial database.

[0028] In some embodiments, determining the historical road condition information of the update node based on the road condition information and the update node corresponding to the road condition information specifically includes the following process: The road condition information and the update nodes corresponding to the road condition information are input into the deep convolutional neural network model, and the historical road condition information of the update nodes is output.

[0029] The training process of the deep convolutional neural network model is as follows: The deep convolutional neural network model consists of several cells, each cell contains N nodes. Represents the node parameters of each Cell, Represents a node, The value is ; Obtain the candidate historical road condition information of all roads of the update node, where different roads of the update node correspond to a historical road condition information. All the candidate historical road condition information are combined into a dataset X. The process of training the deep convolutional neural network model is equivalent to determining each neural network operation. The process: make Between the i-th and j-th nodes, The probability distribution of a parameterized neural network operation, Represents a joint distribution, each Cell consists of N operations, for a given data set X and a candidate operation set , calculate an operation distribution Conditional entropy : ; in, is the probability distribution of the data set X, a set of neural network operation operations; conditional entropy of all operations in the entire neural network Cell is: ; calculating log-likelihood function based on conditional entropy : ; wherein, is a discrete feature indicator function, if the value is less than a preset threshold, then is 1, otherwise, is 0, is the Euclidean distance between node i and node j; wherein, ; update by solving the maximum increase the maximum entropy constraint condition.

[0030] In some embodiments, the historical road condition information and the road condition information are respectively subjected to road trajectory tracking to obtain the historical road trajectory and the road trajectory, which specifically includes the following processes: Step one: based on the historical road condition information or the road condition information, the corresponding historical road route or the road route is obtained, and is a combination parameter, the historical road route or the road route is set with the same sampling points, and each sampling point is combined according to the sampling order of the sampling interval value to obtain a plurality of combination groups; Step two: the first sampling point and the second sampling point of each combination group are combined to form a reference vector m, and the remaining z-th sampling point and the z-1-th sampling point are combined to form a vector wherein, , the angle between is calculated, wherein the calculation formula of the angle is as follows: ; Step three: the is compared with the threshold angle , if the of the z-th point is greater than the threshold angle , the sampling points arranged in front of the z-th sampling point are subdivided into a group, and the sampling points arranged in front of the z-th point are repeated step two and step three to complete the subdivision of the combination group, and the sampling points in each subdivided group are connected to obtain the historical road trajectory and the road trajectory.

[0031] In some embodiments, Figure 2is a workflow diagram of a crowdsourcing collection and freshness maintenance system of a self-updating spatial database according to an embodiment of the present application, as shown in Figure 2 The calculation of the trajectory change feature value between the historical road trajectory and the road trajectory specifically includes the following steps: Step one, count the trajectory segment curve length of the historical road trajectory and the road trajectory, and count the corresponding trajectory curve length of the historical road trajectory and the road trajectory respectively; Step two, determine the maximum trajectory curve length of the two; Step three, calculate the ratio between the trajectory segment curve length and the maximum trajectory curve length, and record the ratio as the trajectory change feature value.

[0032] In some embodiments, the judgment of whether to update the historical road condition information based on the trajectory change feature value specifically includes the following process: Load the trajectory change feature value threshold, judge whether the trajectory change feature value exceeds the trajectory change feature value threshold, if yes, determine that the historical road condition information needs to be updated, if not, determine that the historical road condition information does not need to be updated, wherein the trajectory change feature value threshold is set by the system and stored in the system. As an optional embodiment, the setting process of the trajectory change feature value threshold is as follows: in the early stage of system operation, collect a large number of trajectory change feature values under the scene of "no need to update" (i.e. the feature value change caused by normal factors such as GPS error, different driving lanes when the road does not change substantially), calculate the statistics: the average value of the trajectory change feature value calculated from this batch of historical data: the central tendency of normal fluctuations. Standard deviation: the degree of dispersion of normal fluctuations. Set threshold: a common and effective method is to use the mean value plus several times the standard deviation.

[0033] In some embodiments, the feature extraction of the building image information and the historical building image information in the spatial database is specifically as follows: The building image information and the historical building image information are respectively input into the Transformer encoding layer for feature extraction, and the update image feature and the historical image feature are obtained respectively, wherein, is the length of the feature respectively, is the dimension of the feature, and R is the real number set.

[0034] Further, the update image feature and the historical image feature are interactively compared to obtain the building change index, which specifically includes the following process: Based on a graph structure , where V is the node set representing the target building feature information in the updated image feature, and the node set , is the number of nodes, E is the edge set representing the characteristic information of the target building in the historical image features, , the number of edges is A, where two adjacent nodes Constituent edge ; Edge centrality is defined as the average of the centrality scores of two adjacent nodes as follows: ;in, For the edge The centrality is used to express the influence of the importance between two connected nodes. represents the node centrality measure; Calculate the probability of each edge based on the centrality of each edge: Set To reduce the impact of dense nodes, after normalization, the interaction probability between the two is calculated : ; in, is a hyperparameter whose purpose is to control the overall probability of interaction; yes The maximum value of yes The average value of , used to limit the maximum value of probability; Calculate the building change index: ; Among them, the greater the interaction probability, the more similar the updated image features are to the historical image features, and the smaller the change in the building.

[0035] In some embodiments, determining whether to update historical building image information based on the building change index specifically includes the following process: Load the building change index threshold and determine whether the building change index exceeds the building change index threshold. If so, determine that the historical building image information needs to be updated. If not, determine that the historical building image information does not need to be updated. The building change index threshold is set and stored by the system. The building change index threshold is set according to the process for setting the trajectory change characteristic value threshold described above.

[0036] In some embodiments, updating the historical road condition information and historical building image information in the spatial database includes replacing the spatial data corresponding to the node in the spatial database with the data to be updated uploaded by the selected participant.

[0037] The above formulas are all dimensionless and numerically calculated, and the preset parameters in the formulas are set by technicians in this field according to actual conditions.

[0038] This embodiment also provides an electronic device, Figure 3 is a structural block diagram of an electronic device according to an embodiment of the present invention, such as Figure 3 As shown, the electronic device includes: a memory 301 and a processor 302, wherein the memory 301 stores a computer program; when the computer program is executed by the processor 302, the processor 302 executes the update process of the crowdsourcing collection and freshness maintenance system of the self-updating spatial database as in any of the above embodiments.

[0039] Memory 301 may be an electronic memory such as flash memory, EEPROM (Electrically Erasable Programmable Read-Only Memory), EPROM, hard disk, or ROM. Memory 301 has storage space 303 for program code 313 for executing any of the method steps described above. For example, storage space 303 for program code may include program code 313 for implementing various steps in the method described above. These program codes can be read from or written to one or more computer program products. These computer program products include program code carriers such as hard disks, compact disks (CDs), memory cards, or floppy disks. The program codes may be compressed, for example, in a suitable format. When executed by a processing device, these codes cause the processing device to execute the various steps in the method described above. These program codes can be read from or written to one or more computer program products. These computer program products include program code carriers such as hard disks, compact disks (CDs), memory cards, or floppy disks. The program codes may be compressed, for example, in a suitable format. When these codes are executed by a computing and processing device, they cause the computing and processing device to execute the various steps in the update process of the crowdsourcing collection and freshness maintenance system for a self-updating spatial database described above.

[0040] The above-described embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented by software, the above-described embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the processes or functions described in the embodiments of the present application are wholly or partially generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center through a wired (for example, infrared, wireless, microwave, etc.) manner. The computer-readable storage medium can be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. containing one or more available medium collections. The available medium can be a magnetic medium (for example, a floppy disk, a hard disk, a magnetic tape), an optical medium (for example, a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state disk.

[0041] Those skilled in the art can realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be realized in electronic hardware or a combination of computer software and electronic hardware. Whether the functions are realized in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0042] Those skilled in the art can clearly understand that, for the convenience and brevity of the description, the specific working processes of the above-described system, device and unit can refer to the corresponding processes in the foregoing method embodiments, which will not be described here.

[0043] In several embodiments provided in the present application, it should be understood that the disclosed system, device and method can be implemented in other ways. For example, the above-described device embodiments are only schematic, for example, the division of the units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the units shown or discussed can be indirect coupling or communication connection through some interface, device or unit, and can be electrical, mechanical or other forms.

[0044] The units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, that is, may be located in one place, or may also be distributed to multiple network units. Part or all of the units can be selected to achieve the purpose of the embodiment of the present application according to actual needs.

[0045] The above is merely a specific implementation of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think of changes or replacements within the technical range disclosed by the present application, which should be covered within the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A crowdsourcing collection and freshness maintenance system for self-updating spatial databases, characterized by: The system includes: A crowdsourcing collection unit is used to collect the data to be updated uploaded by legally authenticated selected participants based on the crowdsourcing platform, wherein the data to be updated includes the location information of the selected participants, road condition information, and building image information; A data matching unit is used to calculate the location information of the selected participant and the location matching feature value of each candidate update node in the spatial database, and select the candidate update node corresponding to the largest location matching feature value as the update node; a road analysis unit, configured to determine historical road condition information of an update node based on the road condition information and the update node corresponding to the road condition information, perform road trajectory tracking on the historical road condition information and the road condition information respectively to obtain a historical road trajectory and a road trajectory, calculate a trajectory change characteristic value between the historical road trajectory and the road trajectory, and determine whether to update the historical road condition information based on the trajectory change characteristic value; a building analysis unit for extracting features from the building image information and the historical building image information in the spatial database, respectively, to obtain updated image features and historical image features, interactively comparing the updated image features with the historical image features to obtain a building change index, and determining whether to update the historical building image information based on the building change index; The data updating unit is used to update the historical road condition information and historical building image information in the spatial database.

2. The crowdsourcing collection and freshness maintenance system for the self-updating spatial database according to claim 1 is characterized in that: Calculate the location information of the selected participants and the location matching feature values ​​of each candidate update node in the spatial database. The following processes are included: Get the coordinates of the selected participant's mobile terminal based on the location information of the selected participant ; Based on coordinates Use the inverse tangent function to calculate the selected participant's point to the update node to be selected Azimuth : , the calculated azimuth Convert to degrees , where the node to be updated is Coordinates ; Based on coordinates Use the Euclidean distance formula to calculate the selected participants and the candidate update nodes distance : ; Calculate the location information of the selected participants and the nodes to be updated Position proximity ; The azimuth Corresponding degrees ,distance Proximity to location Remove the dimension and substitute its value into the position matching eigenvalue calculation formula to obtain the position matching eigenvalue WZT. The position matching eigenvalue calculation formula is: ; Among them, the value of c is , is the number of nodes to be updated in the spatial database.

3. The crowdsourcing collection and freshness maintenance system for the self-updating spatial database according to claim 2 is characterized in that: Calculate the location information of the selected participants and the nodes to be updated Position proximity The specific process includes the following: Based on the location information of the selected participant, the IP address information, node code information and geographic location information of the selected participant are selected. The CRF model is used to extract entity references from the IP address information, node code information and geographic location information, and the entity references are converted into vector matrices based on the bag-of-words model. Calculate vector matrix and nodes to be updated The modularity vector between the corresponding node vector matrices : ； Where K is the number of vectors in the vector matrix, represents the adjacency matrix of the vector matrix and the node vector matrix, Represents the node number of the adjacency matrix, 2P represents the adjacency matrix The sum of all edges in , represents the degree of i, represents the degree of j, represents the community in which node i exists, represents the community in which node j exists, Is the conditional restriction coefficient, if the vector matrix and the node to be updated are The node encoding lengths of the corresponding node vector matrices are different, then Is 0, if the vector matrix and the node to be updated The node encoding lengths of the corresponding node vector matrices are the same, then is 1; Modularity vector The modulus value is recorded as the location information of the selected participant and the update node to be selected Position proximity .

4. The crowdsourcing collection and freshness maintenance system for the self-updating spatial database according to claim 1 is characterized in that: Determine the specific historical road condition information of the update node based on the road condition information and the update node corresponding to the road condition information The following processes are included: The road condition information and the update nodes corresponding to the road condition information are input into the deep convolutional neural network model, and the historical road condition information of the update nodes is output.

5. The crowdsourcing collection and freshness maintenance system for the self-updating spatial database according to claim 1 is characterized in that: Tracking the historical road condition information and the road condition information separately to obtain the historical road trajectory and the road trajectory specifically includes the following processes: Step 1: Based on the historical road condition information or road condition information, obtain the corresponding historical road route or road route, set Set the same sampling points for the combined parameters, historical road routes or road routes, and The sampling points are combined according to the sampling order of the sampling interval value to obtain several combination groups; Step 2: The first sampling point and the second sampling point of each combination group form a reference vector m, and the remaining z-th sampling point and the z-1-th sampling point form a vector ,in, , respectively calculated and Angle between , where the angle The calculation formula is as follows: ; Step 3: Angle with threshold For comparison, if the zth point Greater than the threshold angle , then it will be arranged in the first The sampling points in front of the z-th sampling point are subdivided into a group, and steps 2 and 3 are repeated for the sampling points arranged in front of the z-th point to complete the subdivision of the combined group, and the sampling points in each subdivided group are connected to obtain the historical road trajectory and road trajectory.

6. The crowdsourcing collection and freshness maintenance system for the self-updating spatial database according to claim 1 is characterized in that: Calculating the trajectory change feature value between the historical road trajectory and the road trajectory specifically includes the following process: The curve lengths of the overlapping trajectory segments of the historical road trajectory and the road trajectory are counted, and the curve lengths of the trajectory segments corresponding to the historical road trajectory and the road trajectory are counted separately to determine the largest trajectory curve length between the two. The ratio between the overlapping trajectory segment curve length and the largest trajectory curve length is calculated and recorded as the trajectory change characteristic value.

7. The crowdsourcing collection and freshness maintenance system for the self-updating spatial database according to claim 1 is characterized in that: The process of determining whether to update historical road condition information based on trajectory change characteristic values ​​specifically includes the following steps: A trajectory change characteristic value threshold is loaded, and it is determined whether the trajectory change characteristic value exceeds the trajectory change characteristic value threshold. If so, it is determined that the historical road condition information needs to be updated; if not, it is determined that the historical road condition information does not need to be updated, wherein the trajectory change characteristic value threshold is set by the system and stored in the system.

8. The crowdsourcing collection and freshness maintenance system for the self-updating spatial database according to claim 1 is characterized in that: Extract features from building image information and historical building image information in the spatial database to obtain updated image features and historical image features. The following processes are included: The building image information and historical building image information are respectively input into the Transformer encoding layer for feature extraction, and the updated image features are obtained respectively. and historical image features ,in, are the lengths of the features, is the dimension of the feature, and R is a set of real numbers.

9. The crowdsourcing collection and freshness maintenance system for the self-updating spatial database according to claim 8 is characterized in that: The updated image features and historical image features are interactively compared to obtain the building change index. The following processes are included: based on Generate graph structure , where V is the node set representing the target building feature information in the updated image feature, and the node set , is the number of nodes, E is the edge set representing the characteristic information of the target building in the historical image features, , the number of edges is A, where two adjacent nodes Constituent edge ; Edge centrality is defined as the average of the centrality scores of two adjacent nodes as follows: ;in, For the edge The centrality is used to express the influence of the importance between two connected nodes. represents the node centrality measure; Calculate the probability of each edge based on the centrality of each edge: Set To reduce the impact of dense nodes, after normalization, the interaction probability between the two is calculated : ； in, is a hyperparameter whose purpose is to control the overall probability of interaction; yes The maximum value of yes The average value of , used to limit the maximum value of probability; Calculate the building change index: ; Among them, the greater the interaction probability, the more similar the updated image features are to the historical image features, and the smaller the change in the building.

10. The crowdsourcing collection and freshness maintenance system for self-updating spatial database according to claim 1 is characterized in that: The process of determining whether to update the historical building image information based on the building change index specifically includes the following steps: Load the building change index threshold and determine whether the building change index exceeds the building change index threshold. If so, determine that the historical building image information needs to be updated. If not, determine that the historical building image information does not need to be updated. The building change index threshold is set by the system and stored in the system.

Images