CN115936762A - Scenic spot pedestrian flow prediction method, device, equipment and storage medium - Google Patents

Scenic spot pedestrian flow prediction method, device, equipment and storage medium Download PDF

Info

Publication number
CN115936762A
CN115936762A CN202211453713.4A CN202211453713A CN115936762A CN 115936762 A CN115936762 A CN 115936762A CN 202211453713 A CN202211453713 A CN 202211453713A CN 115936762 A CN115936762 A CN 115936762A
Authority
CN
China
Prior art keywords
target
time
scenic spot
historical
data
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202211453713.4A
Other languages
Chinese (zh)
Inventor
王冀琛
魏巍
马梦珍
刘伟
王昊然
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
China United Network Communications Group Co Ltd
Unicom Digital Technology Co Ltd
Original Assignee
China United Network Communications Group Co Ltd
Unicom Digital Technology Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by China United Network Communications Group Co Ltd, Unicom Digital Technology Co Ltd filed Critical China United Network Communications Group Co Ltd
Priority to CN202211453713.4A priority Critical patent/CN115936762A/en
Publication of CN115936762A publication Critical patent/CN115936762A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
    • Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
    • Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation

Landscapes

  • Traffic Control Systems (AREA)

Abstract

The application provides a method, a device, equipment and a storage medium for predicting the pedestrian flow of a scenic region, which relate to the technical field of smart cities.

Description

Scenic spot pedestrian flow prediction method, device, equipment and storage medium
Technical Field
The application relates to the technical field of smart cities, in particular to a scenic spot pedestrian flow prediction method, a scenic spot pedestrian flow prediction device, scenic spot pedestrian flow prediction equipment and a storage medium.
Background
In recent years, with the adjustment of industrial structures and the advance of urbanization, people have more and more great demands on tourism, which leads to explosive people flow in part of scenic spots, and further brings potential problems of destruction of ecological environment of scenic spots, accelerated erosion of cultural relics, shortening of service life of facilities, poor tourism experience of tourists and the like. Therefore, accurate scenic spot pedestrian volume prediction can assist a scenic spot manager to make scenic spot shunting in advance, and the reception capacity of the scenic spot is improved; the tourist can also be helped to stagger the peak of the pedestrian volume in scenic spots, and the satisfaction degree of tourism is improved.
In the related art, the traditional machine learning method is usually adopted to predict the scenic spot people flow. Specifically, the back propagation neural network is trained through historical data related to the scenic spot in a historical preset time period, so that the back propagation neural network can perform big data analysis processing on the historical data and real-time data related to the scenic spot in a future preset time period, and people flow data of the scenic spot in the future preset time period are obtained through prediction, wherein the future preset time period corresponds to the historical preset time period, but the prediction precision is low.
Disclosure of Invention
The application provides a scenic spot pedestrian flow prediction method, a scenic spot pedestrian flow prediction device, scenic spot pedestrian flow prediction equipment and a storage medium, which are used for solving the problem of low scenic spot pedestrian flow prediction precision in the related technology.
In a first aspect, the present application provides a method for predicting a pedestrian volume in a scenic spot, including:
acquiring historical data and external related data of a target time interval corresponding to a target scene, wherein the target time interval is determined according to time to be predicted, the historical data comprises historical pedestrian volume data of the target scene in the historical target time interval under multi-dimensional time granularity and historical traffic flow data of traffic objects in a preset range of the target scene, the external related data comprises weather data of the target scene in the time to be predicted, per capita production total data of the area where the target scene is located and key event data, the multi-dimensional time granularity comprises year, quarter, month, week, day, hour and preset time intervals, and the traffic objects comprise stations, checkpoints and landmarks; determining a spatial dependency relationship between the scenic spot pedestrian volume of the target scenic spot and the traffic flow data of the traffic object according to the historical pedestrian volume data and the historical traffic flow data; determining the time dependency relationship between the scenic spot pedestrian volume of the target scenic spot and historical target time periods under the multi-dimensional time granularity according to the space dependency relationship and the historical pedestrian volume data; and predicting the pedestrian volume of the target scenic spot at the time to be predicted according to the time dependence relationship and the external related data.
In one possible implementation, determining a spatial dependency relationship between a scenic spot pedestrian volume of a target scenic spot and traffic flow data of a traffic object according to historical pedestrian volume data and historical traffic flow data includes: determining the distance between a target scenic spot and a traffic object; acquiring target historical traffic flow data corresponding to the target traffic object with the corresponding distance smaller than the distance threshold value from the historical traffic flow data; and determining the spatial dependence relationship between the scenic spot pedestrian volume of the target scenic spot and the traffic flow data of the target traffic object according to the historical traffic flow data of the target and the historical pedestrian volume data.
In one possible implementation manner, determining a spatial dependency relationship between a scenic spot pedestrian volume of a target scenic spot and traffic flow data of a target traffic object according to target historical traffic flow data and historical pedestrian volume data includes: and inputting the historical traffic flow data and the historical pedestrian flow data of the target into the spatial relationship model to obtain a spatial dependency relationship between the scenic spot pedestrian flow of the target scenic spot and the traffic flow data of the target traffic object, which are output by the spatial relationship model.
In one possible implementation, the spatial relationship model includes a plurality of convolutional layers; a flow control layer is arranged between the adjacent convolution layers and is used for adjusting the spatial dependence relationship by adopting a flow control mechanism; inputting the historical traffic flow data and the historical pedestrian flow data of the target into a spatial relationship model to obtain a spatial dependency relationship between the scenic spot pedestrian flow of the target scenic spot and the traffic flow data of the target traffic object, which is output by the spatial relationship model, and the spatial dependency relationship comprises the following steps: inputting target historical traffic flow data and historical pedestrian flow data into the 1 st convolutional layer for spatial feature extraction to obtain spatial features output by the 1 st convolutional layer, wherein different spatial features have spatial dependency relations; adjusting the local spatial dependency relationship according to the spatial characteristics output by the ith convolutional layer through a flow control layer, wherein i sequentially takes 1,2, … … and N-1,N as the number of convolutional layers; and inputting the local spatial dependency relationship into the (i + 1) th convolution layer for spatial feature extraction to obtain spatial features output by the (i + 1) th convolution layer, wherein the spatial dependency relationship between the spatial features output by the Nth convolution layer is the spatial dependency relationship between the scenic spot pedestrian flow of the target scenic spot and the traffic flow data of the target traffic object.
In one possible implementation, determining a time dependency relationship between historical target time periods of scenic spot pedestrian volume of a target scenic spot at a multi-dimensional time granularity according to a spatial dependency relationship and historical pedestrian volume data includes: and inputting the spatial dependency relationship and historical people flow data into the time relationship model to obtain the time dependency relationship between historical target time periods of the scenic spot people flow of the target scenic spot output by the time relationship model under the multi-dimensional time granularity.
In one possible implementation, the time relationship model includes a bidirectional long-short term memory network and an attention module, and the attention module is configured to adjust the time dependency relationship using an attention mechanism; inputting the spatial dependency relationship and historical people flow data into a time relationship model to obtain the time dependency relationship between historical target time periods of the scenic spot people flow of a target scenic spot output by the time relationship model under the multi-dimensional time granularity, and the method comprises the following steps: inputting the historical pedestrian flow data of the space dependency relationship and the dimension time granularity into a bidirectional long and short term memory network aiming at each dimension time granularity in the multi-dimension time granularity to obtain time characteristics output by the bidirectional long and short term memory network, wherein the time characteristics have a time dependency relationship; and adjusting the time dependency corresponding to the time characteristics of the multidimensional time granularity by the attention module to obtain the time dependency of the scenic spot pedestrian volume of the target scenic spot in the historical target time period under the multidimensional time granularity.
In one possible implementation manner, predicting the pedestrian volume of the target scenic spot at the time to be predicted according to the time dependency relationship and the external related data includes: and carrying out fusion processing on the time dependence relation and the external related data to obtain the pedestrian volume of the target scenic spot at the time to be predicted.
In a second aspect, the present application provides a device for predicting a flow rate of people in a scenic spot, comprising:
the system comprises an acquisition module, a prediction module and a prediction module, wherein the acquisition module is used for acquiring historical data and external related data of a target time interval corresponding to a target scenic spot, the target time interval is determined according to time to be predicted, the historical data comprises historical pedestrian volume data of the target scenic spot in the historical target time interval and historical traffic flow data of traffic objects in a preset range of the target scenic spot under multi-dimensional time granularity, the external related data comprises weather data of the target scenic spot in the time to be predicted, per capita total production data and key event data of the area where the target scenic spot is located, the multi-dimensional time granularity comprises year, quarter, month, week, day, hour and preset time intervals, and the traffic objects comprise stations, checkpoints and landmarks; the first determination module is used for determining a spatial dependency relationship between the scenic spot pedestrian volume of the target scenic spot and the traffic flow data of the traffic object according to the historical pedestrian volume data and the historical traffic flow data; the second determination module is used for determining the time dependency relationship between the scenic spot pedestrian volume of the target scenic spot and historical target time periods under the multi-dimensional time granularity according to the space dependency relationship and the historical pedestrian volume data; and the prediction module is used for predicting the pedestrian volume of the target scenic spot at the time to be predicted according to the time dependence relationship and the external related data.
In a possible implementation manner, the first determining module is specifically configured to: determining the distance between a target scenic spot and a traffic object; acquiring target historical traffic flow data corresponding to the target traffic object with the corresponding distance smaller than the distance threshold value from the historical traffic flow data; and determining the spatial dependence relationship between the scenic spot pedestrian volume of the target scenic spot and the traffic flow data of the target traffic object according to the historical traffic flow data of the target and the historical pedestrian volume data.
In one possible implementation manner, the first determining module may be further configured to: and inputting the historical traffic flow data and the historical pedestrian flow data of the target into the spatial relationship model to obtain the spatial dependence relationship between the scenic spot pedestrian flow of the target scenic spot and the traffic flow data of the target traffic object, which is output by the spatial relationship model.
In one possible implementation, the spatial relationship model includes a plurality of convolution layers; a flow control layer is arranged between the adjacent convolution layers and is used for adjusting the spatial dependence relationship by adopting a flow control mechanism; the first determining module may be further configured to: inputting target historical traffic flow data and historical pedestrian flow data into the 1 st convolutional layer for spatial feature extraction to obtain spatial features output by the 1 st convolutional layer, wherein different spatial features have spatial dependency relations; adjusting the local spatial dependency relationship according to the spatial characteristics output by the ith convolutional layer through a flow control layer, wherein i sequentially takes 1,2, … … and N-1,N as the number of convolutional layers; and inputting the local spatial dependency relationship into the (i + 1) th convolution layer for spatial feature extraction to obtain the spatial feature output by the (i + 1) th convolution layer, wherein the spatial dependency relationship between the spatial features output by the (N) th convolution layer is the spatial dependency relationship between the scenic spot pedestrian flow of the target scenic spot and the traffic flow data of the target traffic object.
In a possible implementation manner, the second determining module is specifically configured to: and inputting the spatial dependency relationship and historical people flow data into the time relationship model to obtain the time dependency relationship between historical target time periods of the scenic spot people flow of the target scenic spot output by the time relationship model under the multi-dimensional time granularity.
In one possible implementation, the time relationship model includes a bidirectional long-short term memory network and an attention module, and the attention module is configured to adjust the time dependency relationship using an attention mechanism; the second determination module may be further operable to: inputting the historical pedestrian flow data of the space dependency relationship and the dimension time granularity into a bidirectional long and short term memory network aiming at each dimension time granularity in the multi-dimension time granularity to obtain time characteristics output by the bidirectional long and short term memory network, wherein the time characteristics have a time dependency relationship; and adjusting the time dependency corresponding to the time characteristics of the multidimensional time granularity by the attention module to obtain the time dependency of the scenic spot pedestrian volume of the target scenic spot in the historical target time period under the multidimensional time granularity.
In one possible implementation, the prediction module is specifically configured to: and carrying out fusion processing on the time dependence relation and the external related data to obtain the pedestrian volume of the target scenic spot at the time to be predicted.
In a third aspect, the present application provides an electronic device, comprising:
at least one processor;
and a memory coupled to the at least one processor;
wherein the memory is configured to store instructions executable by the at least one processor to cause the at least one processor to perform the method for predicting scenic spot traffic provided by the first aspect.
In a fourth aspect, the present application provides a computer-readable storage medium, in which computer-executable instructions are stored, and the computer-executable instructions are executed by a processor to implement the method for predicting the flow of people in the scenic spot provided in the first aspect.
In a fifth aspect, the present application provides a program product comprising computer executable instructions. When the computer executes the instructions, the method for predicting the scenic spot pedestrian volume provided by the first aspect is realized.
The application provides a method, a device, equipment and a storage medium for predicting the pedestrian flow of a scenic region, historical pedestrian flow data of the target scenic region in a historical target time period, historical traffic flow data of traffic objects in a preset range of the target scenic region, weather data of the target scenic region in time to be predicted, total production value data of all people in the region where the target scenic region is located and key event data, which are divided according to the granularity of year, quarter, month, week, day, hour and preset time interval time, of the target time period corresponding to the target scenic region are obtained, the spatial dependence relationship between the pedestrian flow of the scenic region of the target scenic region and the traffic flow data of the traffic objects is determined based on the historical pedestrian flow data and the historical traffic flow data, the time dependence relationship between the pedestrian flow of the scenic region of the target scenic region and the traffic flow data of the traffic objects is further determined by combining the historical pedestrian flow data on the basis of the spatial dependence relationship, the historical pedestrian flow data is combined with the historical pedestrian flow data on the historical target time interval time of the multidimensional time, the temporal and spatial dependence of the pedestrian flow of the target scenic region is fully mined, and the accuracy of the predicted pedestrian flow of the target scenic region is improved by further combining the external weather data, the total production data of the predicted pedestrian flow of the target scenic region is further on the basis of the temporal dependence of the temporal and the temporal dependence of the predicted time. In addition, the historical pedestrian flow data and the historical traffic flow data are divided in a multi-dimension mode in time granularity, so that the long-term pedestrian flow and the short-term pedestrian flow in the scenic region can be accurately predicted.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the application and, together with the description, serve to explain the principles of the application.
Fig. 1 is a schematic structural diagram of an application scenario provided in an embodiment of the present application;
fig. 2 is a flowchart of a method for predicting scenic spot traffic according to an embodiment of the present disclosure;
fig. 3 is a flowchart of a method for predicting scenic spot traffic according to another embodiment of the present disclosure;
fig. 4 is a flowchart of a method for predicting scenic spot traffic according to another embodiment of the present application;
FIG. 5 is a schematic structural diagram of a graph structure provided in an embodiment of the present application;
fig. 6 is a flowchart of a method for predicting scenic spot traffic according to another embodiment of the present application;
fig. 7 is a schematic structural diagram of a device for predicting scenic spot pedestrian volume according to an embodiment of the present application;
fig. 8 is a schematic structural diagram of an electronic device according to an embodiment of the present application.
With the above figures, there are shown specific embodiments of the present application, which will be described in more detail below. The drawings and written description are not intended to limit the scope of the inventive concepts in any manner, but rather to illustrate the concepts of the application by those skilled in the art with reference to specific embodiments.
Detailed Description
Reference will now be made in detail to the exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, like numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary examples do not represent all implementations consistent with the present application. Rather, they are merely examples of apparatus and methods consistent with certain aspects of the present application, as detailed in the appended claims.
In the related art, for the prediction of the scenic spot pedestrian volume, on one hand, a traditional statistical method is adopted, for example, historical annual average pedestrian volume data of previous years of the scenic spot are collected, and the moving average of the annual average pedestrian volume data is used as the predicted pedestrian volume of the current year, so that the method cannot process complex nonlinear scenic spot pedestrian volume data and cannot accurately model the real situation that the scenic spot pedestrian volume is influenced by uncertain factors such as environment and the like; on the other hand, the traditional machine learning method is adopted to predict the scenic spot pedestrian volume, and the specific implementation manner is as described in the background art, the method lacks the extraction of the characteristics of the scenic spot pedestrian volume data combining time and space, cannot fully mine the time-space correlation of the scenic spot pedestrian volume data, and has low prediction precision.
Aiming at the problems in the related technology, the embodiment of the application fully excavates the spatial dependency relationship between the scenic spot and subway stations, bus stations, landmarks and the like around the scenic spot, and the time dependency relationship of the scenic spot pedestrian volume under the multi-dimensional time granularity of year, quarter, month, week, day, hour, time interval and the like, and predicts the pedestrian volume of the scenic spot in the time to be predicted on the basis of the combination of the spatial dependency relationship and the time dependency relationship, so that the prediction precision of the pedestrian volume of the scenic spot is improved. In addition, by considering the time dependence of the scenic spot pedestrian volume under the multidimensional time granularity, the long-term and short-term scenic spot pedestrian volume can be accurately predicted.
For ease of understanding, first, a brief description is made of an application scenario related to the embodiments of the present application.
Fig. 1 is a schematic structural diagram of an application scenario provided in an embodiment of the present application. As shown in fig. 1, an application scenario provided by the embodiment of the present application includes a traffic object 11, a scenic spot 12, and a server 13. The traffic object 11 may be a bus station, a subway station, a landmark, or the like, and specifically, the traffic object may be a part of traffic objects distributed around the scenic spot 12, or may be all traffic objects located in the same city as the scenic spot 12. The embodiment of the application does not limit the number of the traffic objects and scenic spots. Specifically, the application scenario provided in the embodiment of the present application is described with a scenic spot as an example.
The execution subject of the embodiment of the present application may be the server 13. Specifically, the server 13 acquires historical traffic flow data and historical people flow data for predicting traffic objects corresponding to the scenic spot 12 at the time to be predicted from the system corresponding to the traffic objects 11 and the scenic spot 12; acquiring weather data of the scenic spot 12 in the time to be predicted from a weather monitoring platform corresponding to the scenic spot 12; acquiring the total production value data of everyone in the area of the scenic spot 12 and the key event data from the related website, analyzing and processing the acquired data based on big data to obtain the spatial dependency relationship between the scenic spot 12 and the traffic object 11 and the time dependency relationship of the pedestrian volume of the scenic spot 12 in different time dimensions, and accurately predicting the pedestrian volume of the scenic spot 12 at the time to be predicted by combining the spatial dependency relationship and the time dependency relationship. For example, the system corresponding to the traffic object 11 may be a subway service system, a bus station platform management system, and the like, and the system corresponding to the scenic spot 12 may be a scenic spot management system, and the like.
The following describes in detail a method for predicting the traffic of people in a scenic spot according to an embodiment of the present application, taking the above application scenario as an example.
Fig. 2 is a flowchart of a method for predicting scenic spot traffic according to an embodiment of the present disclosure. As shown in fig. 2, the method for predicting the flow of people in the scenic spot comprises the following steps:
s201, historical data and external related data of a target time interval corresponding to a target scene are obtained, wherein the target time interval is determined according to time to be predicted, the historical data comprises historical pedestrian volume data of the target scene in the historical target time interval under multi-dimensional time granularity and historical traffic flow data of traffic objects in a preset range of the target scene, the external related data comprises weather data of the target scene in the time to be predicted, per capita production total data of the region where the target scene is located and key event data, the multi-dimensional time granularity comprises year, quarter, month, week, day, hour and preset time intervals, and the traffic objects comprise stations, checkpoints and landmarks.
Alternatively, the time to be predicted may be a specific time point or a time period. For example, when the time to be predicted is 3 pm on 11/3/2022, the target time period may be a time period from 2 pm 45 to 15 pm, or from 2 pm half to 3 pm half, and the time interval between the predicted time and the start time point or the end time point corresponding to the target time period is not limited in the embodiment of the present application.
It is understood that the scenic spot traffic may exhibit periodicity-like characteristics over a year period, a quarter period, a month period, a week period, a day period, an hour period, and a time interval period, and thus, the multidimensional time granularity provided by the embodiments of the present application includes year, quarter, month, week, day, hour, and preset time intervals. For example, the class periodicity corresponding to the annual time granularity can be reflected in the graduation season of each year, and the scenic spot traffic is more; the similar periodicity corresponding to the seasonal time granularity can reflect that the plant garden has more people flow in spring and autumn; the class periodicity corresponding to the month time granularity can reflect the class periodicity change of the human flow caused by cycle factors such as climate change, vacation arrangement and the like; the class periodicity corresponding to the week time granularity can be embodied in the class periodicity change of the human traffic on weekends and weekdays; the class periodicity corresponding to the day time granularity can be reflected in different travel and rest rules of different tourists; the class periodicity corresponding to the hour time granularity can be reflected in the difference of physical strength, personal habits and the like of different tourists; the class periodicity corresponding to the time interval time granularity can reflect the influence of the pedestrian volume corresponding to traffic objects such as landmarks, bus stations, subway stations and the like near the scenic spot on the pedestrian volume of the scenic spot.
Optionally, the historical people flow data may be historical people flow data corresponding to different time granularities, and the historical people flow data may correspond to different people flow data corresponding to different time granularities. Illustratively, when the time to be predicted is 3 pm on 11/3/2022, then the historical traffic data obtained according to the granularity of year, quarter, month, week, day, hour and preset time interval are traffic data of 2 pm 45 to 15 pm on 11/3/2021, 8/3/pm 2 pm 45 to 15 pm, traffic data of 2 pm 45 to 3 pm on 10/3/2022, 3/pm 2 pm 45 to 3 pm 15 on 10/3/2022, 27/10/27/pm 2 pm 45 to 3 pm 15, traffic data of 9 am 45 to 10 am 15 on 11/3/m 2022/3/m, and traffic data of 2 pm 15 to 3 pm 45 on 11/3/2/3/m 2022, respectively, wherein the selected hour dimension is 5 hours and the preset time interval is 30 minutes. It can be understood that, according to the difference of the time to be predicted, the dimensionality of the historical people flow data corresponding to different time granularities which can be obtained is also different, for the prediction of the short-term people flow, the time dimensionality can obtain the historical data corresponding to the time granularity which takes the preset time interval as the maximum, and for the prediction of the long-term people flow, the time dimensionality can obtain the historical data corresponding to the time granularity which takes the year as the maximum.
Optionally, the historical traffic flow data includes historical pedestrian flow data and historical road vehicle flow data corresponding to the traffic object. The historical pedestrian flow data and the historical road traffic flow data may also be historical pedestrian flow data and historical road traffic flow data corresponding to different dimensionality time granularities, and the specific division of the time granularity is the same as that described above, and is not repeated here.
The historical pedestrian flow data and the historical traffic flow data include both inflow data and outflow data.
Optionally, the weather data may also include historical weather data for a historical target period.
For example, the critical event data may be data of events and some important activities in the area of the target scene.
S202, according to the historical people flow data and the historical traffic flow data, the spatial dependence relationship between the scenic spot people flow of the target scenic spot and the traffic flow data of the traffic object is determined.
Optionally, the spatial dependency relationship between the scenic spot pedestrian volume of the target scenic spot and the traffic flow data of the traffic object is used to reflect the importance degree of the traffic object on the influence of the scenic spot pedestrian volume, and specifically, the spatial dependency relationship may be represented by a weight. For example, in a historical target time period, if the pedestrian flow and/or the road and vehicle flow of the corresponding subway station near the target scenic spot is more or less, and the pedestrian flow of the target scenic spot is less or more in the time period, it is considered that the influence of the subway station on the pedestrian flow of the scenic spot in the time period is greater, which indicates that the spatial dependency between the target scenic spot and the subway station is greater.
In some embodiments, the spatial dependency between the scenic spot traffic volume of the target scenic spot and the traffic flow data of the traffic object may be determined according to a linear relationship between the historical people flow data and the historical traffic flow data. Illustratively, when the historical pedestrian flow data and the historical traffic flow data are obviously positively or negatively correlated, the corresponding spatial dependency relationship is larger, and when no obvious linear correlation exists between the historical pedestrian flow data and the historical traffic flow data, the corresponding spatial dependency relationship is smaller.
S203, according to the spatial dependency relationship and the historical people flow data, determining the time dependency relationship between historical target time periods of the scenic spot people flow of the target scenic spot under the multidimensional time granularity.
Optionally, the time dependency is used to reflect the importance degree of the corresponding time period under different time granularities on the time to be predicted. Specifically, the temporal dependency may be represented by a weight. For example, the weights for different time granularities may be adjusted and determined using an attention mechanism. Attention mechanism is a mechanism that selectively uses features of the input sequence such that features with a higher relevance to the task are specifically considered and features with a lower relevance are ignored.
In some embodiments, based on an attention mechanism, a correlation degree between time intervals between the time to be predicted and time granularities corresponding to historical people flow data is extracted, and a time dependency relationship between historical target time periods of the scenic spot people flow of the target scenic spot under the multidimensional time granularity is determined.
And S204, predicting the pedestrian volume of the target scenic spot at the time to be predicted according to the time dependence relationship and the external related data.
In some embodiments, the external related data may be normalized into a two-dimensional vector, and the historical pedestrian volume data corresponding to the time dependency relationship may be combined with the external related data after the normalization processing, and the pedestrian volume of the target scenic spot at the time to be predicted may be output through the full connection layer of the neural network. Illustratively, the neural network may be a bidirectional long-short term memory network.
In the embodiment of the application, historical people flow data of a target scene in the historical target time period, historical traffic flow data of a traffic object in a preset range of the target scene, weather data of the target scene in time to be predicted, total production value data of people in the area where the target scene is located and key event data of the target scene, which are divided according to the granularity of time at preset time intervals and correspond to the target time period, are obtained, a spatial dependency relationship between the scene people flow of the target scene and the traffic flow data of the traffic object is determined based on the historical people flow data and the historical traffic flow data, and further, on the basis of the spatial dependency relationship, the time dependency relationship between the scene people flow of the target scene in the historical target time period under the multi-dimensional time granularity is determined by combining the historical people flow data, the space-time dependency of the people flow of the target scene is fully mined, on the basis of the space-time dependency, the influence of relevant data such as external weather data, total production value data of people and key event data on the scene flow of the target scene is further combined on the time, and the time dependency of the target scene flow to be predicted, and the accuracy of the target scene to be predicted is improved. In addition, the historical pedestrian flow data and the historical traffic flow data are divided in a multi-dimension mode in time granularity, so that the long-term pedestrian flow and the short-term pedestrian flow in the scenic region can be accurately predicted.
In some embodiments, according to the time dependency relationship and the external related data, the implementation manner of predicting the pedestrian volume of the target scenic spot at the time to be predicted may be: and performing fusion processing on the time dependency relationship and the external related data to obtain the pedestrian volume of the target scenic spot at the time to be predicted. Illustratively, the fusion processing can be performed on the time dependency relationship and the external related data based on the neural network.
Based on the above embodiment, step S202 is described in detail below with reference to fig. 3.
Fig. 3 is a flowchart of a method for predicting scenic spot traffic according to another embodiment of the present disclosure. As shown in fig. 3, determining the spatial dependency relationship between the scenic spot pedestrian volume of the target scenic spot and the traffic flow data of the traffic object according to the historical pedestrian volume data and the historical traffic flow data may specifically include the following steps:
s301, determining the distance between the target scenic spot and the traffic object.
Optionally, the server may also obtain the distance between the target scenic spot and the traffic object. For example, the distance may be an actual distance between the target scenic spot and each traffic object based on different travel manners, or may be a vehicle speed, and the like.
S302, obtaining target historical traffic flow data corresponding to the target traffic object with the corresponding distance smaller than the distance threshold value from the historical traffic flow data.
Optionally, the historical traffic flow data may also include distances between the target scene and the traffic objects. It can be understood that, when the distance between the target scenic spot and the traffic object is relatively long, the traffic flow data corresponding to the traffic object has little influence on the flow of people in the target scenic spot within a certain period of time, and the spatial dependence relationship between the corresponding traffic object and the target scenic spot is also weak and can be ignored.
And S303, determining a spatial dependence relationship between the scenic spot pedestrian volume of the target scenic spot and the traffic flow data of the target traffic object according to the historical traffic flow data and the historical pedestrian volume data of the target.
The specific method for determining the spatial dependency relationship is similar to that described above, and is not described herein again.
In the embodiment of the application, the distance between the target scenic spot and the traffic object is determined, the target historical traffic flow data corresponding to the target traffic object with the corresponding distance smaller than the distance threshold value is obtained from the historical traffic flow data, and the spatial dependence relationship between the scenic spot pedestrian volume of the target scenic spot and the traffic flow data of the target traffic object is determined according to the target historical traffic flow data and the historical pedestrian volume data, so that the accuracy of the target traffic object on the spatial dependence relationship of the target scenic spot is ensured, and the prediction precision of the scenic spot pedestrian volume is improved.
In some embodiments, the spatial dependency relationship between the scenic spot pedestrian volume of the target scenic spot and the traffic flow data of the target traffic object may be obtained based on a spatial relationship model, and specifically, determining the spatial dependency relationship between the scenic spot pedestrian volume of the target scenic spot and the traffic flow data of the target traffic object according to the target historical traffic flow data and the historical pedestrian volume data may further include the following steps: and inputting the historical traffic flow data and the historical pedestrian flow data of the target into the spatial relationship model to obtain the spatial dependence relationship between the scenic spot pedestrian flow of the target scenic spot and the traffic flow data of the target traffic object, which is output by the spatial relationship model.
Optionally, the spatial relationship model may include a plurality of convolutional layers, and a flow control layer is disposed between adjacent convolutional layers, where the flow control layer is configured to adjust the spatial dependency relationship by using a flow control mechanism. The flow control mechanism has the action principle that historical traffic flow data corresponding to traffic objects in the surrounding neighborhood of a target scenic spot at a certain moment are considered, and then the weight between the traffic objects in the target scenic spot and the surrounding neighborhood is dynamically adjusted to obtain a dynamic spatial dependency relationship.
A specific implementation manner of inputting the target historical traffic flow data and the historical pedestrian flow data into the spatial relationship model to obtain the spatial dependency relationship between the scenic spot pedestrian flow of the target scenic spot and the traffic flow data of the target traffic object, which is output by the spatial relationship model, in the above embodiment is described in detail with reference to fig. 4.
Fig. 4 is a flowchart of a method for predicting scenic spot traffic according to another embodiment of the present disclosure. As shown in fig. 4, inputting the target historical traffic flow data and the historical pedestrian flow data into the spatial relationship model to obtain a spatial dependency relationship between the scenic spot pedestrian flow of the target scenic spot and the traffic flow data of the target traffic object, which is output by the spatial relationship model, and specifically, the method may further include the following steps:
s401, inputting the target historical traffic flow data and the historical pedestrian flow data into the 1 st convolutional layer for spatial feature extraction, and obtaining the spatial features output by the 1 st convolutional layer, wherein different spatial features have spatial dependency relations.
In one possible implementation, the target historical traffic flow data and the historical pedestrian flow data may be input into the 1 st convolutional layer in a matrix form, and the 1 st convolutional layer performs convolution processing of downsampling on the target historical traffic flow data and the historical pedestrian flow data to obtain the spatial features output by the 1 st convolutional layer. Wherein, the weights corresponding to the spatial dependency relationship between the spatial features of the 1 st convolutional layer output are the same. For example, a city in which the target scenic spot is located may be divided into M × N grids according to the target scenic spot and the traffic object, each grid carries corresponding target historical traffic flow data or historical pedestrian flow data, the grids are divided according to time granularity, and the grids are respectively input to the first convolution layer according to different time granularities. For example, the input data may be represented by the following formulas according to different time granularities:
Figure BDA0003952614130000121
Figure BDA0003952614130000122
Figure BDA0003952614130000123
Figure BDA0003952614130000124
Figure BDA0003952614130000125
/>
Figure BDA0003952614130000126
Figure BDA0003952614130000127
wherein i represents a target sceneRegion, t represents the time to be predicted, F i，t Representing input data, y representing a time of year granularity, q representing a time of quarter granularity, m representing a time of month granularity, w representing a time of week granularity, d representing a time of day granularity, h representing an hour time granularity, and τ representing a preset time interval granularity.
Illustratively, the 1 st convolutional layer may be represented by the following formula:
Figure BDA0003952614130000131
where k denotes the order of the convolutional layers, where ReLU denotes the activation function,
Figure BDA0003952614130000132
the convolutional neural network corresponding to the convolutional layer is a parameter to be learned, and represents the convolution operation. Illustratively, the convolutional neural network may be a locally spatially dependent convolutional neural network.
In another possible implementation manner, the target historical traffic flow data and the historical pedestrian flow data may be input into the 1 st convolutional layer in a graph structure form, and the 1 st convolutional layer filters out nodes corresponding to traffic objects with weak influence on the pedestrian flow of the target scenic region in the undirected graph according to the distance between the target scenic region and the traffic objects, so as to obtain the spatial features output by the 1 st convolutional layer. Wherein, the weights corresponding to the spatial dependency relationship between the spatial features of the 1 st convolutional layer output are the same. Fig. 5 is a schematic structural diagram of a graph structure provided in an embodiment of the present application. As shown in fig. 5, wherein the black circles represent target scenic spots and the gray circles represent traffic objects. Specifically, the nodes corresponding to different traffic objects may be nodes directly connected to the target scenic spot, or may be nodes connected to the traffic objects.
S402, adjusting the local spatial dependency relationship according to the spatial characteristics output by the ith convolution layer through a flow control layer, wherein i sequentially takes 1,2, … … and N-1,N as the number of convolution layers; and inputting the local spatial dependency relationship into the (i + 1) th convolution layer for spatial feature extraction to obtain the spatial feature output by the (i + 1) th convolution layer.
And the spatial dependency relationship between the spatial features output by the Nth convolution layer is the spatial dependency relationship between the scenic spot pedestrian volume of the target scenic spot and the traffic flow data of the target traffic object.
Optionally, the flow control layer adjusts the local spatial dependency relationship by using a flow control mechanism based on the spatial feature output by the ith convolutional layer. Specifically, the local spatial dependency relationship is used for reflecting the importance degree of the traffic object on the influence of the traffic object on the flow of people in the scenic region. The local spatial dependency relationship is adjusted according to the historical people flow data and the historical traffic flow data, and the specific mode is similar to the above, and is not described again here.
In the embodiment of the application, historical traffic flow data and historical pedestrian flow data of the target are processed by utilizing the plurality of convolutional layers in the spatial relationship model and the flow control layers between the adjacent convolutional layers to obtain the spatial dependence relationship between the scenic spot pedestrian flow of the target scenic spot and the traffic flow data of the target traffic object, so that the accuracy of the spatial dependence relationship between the scenic spot pedestrian flow of the target scenic spot and the traffic flow data of the target traffic object is ensured, and the prediction precision is improved.
In some embodiments, the time dependency relationship between the historical target time periods of the scenic spot pedestrian volume of the target scenic spot at the multidimensional time granularity may be obtained based on a time relationship model, and specifically, determining the time dependency relationship between the historical target time periods of the scenic spot pedestrian volume of the target scenic spot at the multidimensional time granularity according to the spatial dependency relationship and the historical pedestrian volume data may further include the following steps: and inputting the spatial dependency relationship and historical people flow data into the time relationship model to obtain the time dependency relationship between historical target time periods of the scenic spot people flow of the target scenic spot output by the time relationship model under the multi-dimensional time granularity.
Optionally, the temporal relationship model may include a bidirectional long-short term memory network and an attention module, wherein the attention module is configured to adjust the temporal dependencies using an attention mechanism. A specific implementation manner of inputting the spatial dependency relationship and the historical people flow data into the time relationship model to obtain the time dependency relationship between the historical target time periods of the scenic spot people flow of the target scenic spot output by the time relationship model under the multidimensional time granularity in the above embodiment is described in detail below with reference to fig. 6.
Fig. 6 is a flowchart of a method for predicting scenic spot traffic according to yet another embodiment of the present application. As shown in fig. 6, inputting the spatial dependency relationship and the historical pedestrian volume data into the time relationship model to obtain the time dependency relationship between historical target time periods of the scenic spot pedestrian volume of the target scenic spot output by the time relationship model under the multidimensional time granularity, may specifically include the following steps:
s601, aiming at each dimension time granularity in the multi-dimension time granularities, inputting the historical pedestrian flow data of the space dependency relationship and the dimension time granularity into the bidirectional long-short term memory network to obtain the time characteristics output by the bidirectional long-short term memory network, wherein the time characteristics have a time dependency relationship.
First, a brief description will be given of a bidirectional long-term and short-term memory network. The bidirectional long and short term memory network (BLSTM) has a forward propagating long and short term memory network and a backward propagating long and short term memory network, which are used for inputting input data according to a forward sequence and a backward sequence, respectively, extracting features of the input data, and combining output features of the forward propagating long and short term memory network and the backward propagating long and short term memory network to obtain final output features of the BLSTM. Illustratively, when the pedestrian volume data input to BLSTM is pedestrian volume data from 2 pm 45 to 3 pm 15 on 3 pm on 10/2022, the input data in the forward order is pedestrian volume data in the order from 2 pm 45 to 3 pm 15, and the input data in the reverse order is pedestrian volume data in the order from 3 pm 15 to 2 pm 45.
Compared with the traditional long-short term memory network with only positive sequence input in time sequence, the bidirectional long-short term memory network can fully mine the periodic pattern of the scenic spot pedestrian flow from two angles of time forward and time reverse, and the influence on the prediction accuracy of the scenic spot pedestrian flow caused by the fact that some characteristics of the scenic spot pedestrian flow are filtered by network units or partial characteristics cannot effectively pass through the network units is avoided. For example, in practice, the scenic spot traffic does not have sudden increase or sudden decrease in a short time, and the sudden increase or sudden decrease in the traffic in a certain period is formed by convergence and divergence of the traffic of adjacent traffic objects, so that the data with sudden increase or sudden decrease in the target period can be smoothed by using the bidirectional long-short term memory network to improve the final prediction accuracy.
In a possible implementation manner, the embodiments of the present application adopt
Figure BDA0003952614130000151
Represents the positive sequence input of the time t to be predicted in the target scenic spot i>
Figure BDA0003952614130000152
Input in reverse order, F, representing the time t to be predicted in the target scene i i，t Representing the characteristic of the target scenic spot i at the time t to be predicted, and the corresponding characteristic x output by the bidirectional long-short term memory network i，t Can be expressed by the following formula:
Figure BDA0003952614130000153
optionally, for the historical pedestrian volume data with the multi-dimensional time granularity provided in the embodiment of the present application, after the BLSTM performs feature extraction, the output time features of the historical pedestrian volume data can be sequentially represented by the following formula:
Figure BDA0003952614130000154
Figure BDA0003952614130000155
Figure BDA0003952614130000156
/>
Figure BDA0003952614130000157
Figure BDA0003952614130000158
Figure BDA0003952614130000159
wherein i represents a target scenic region, t represents time to be predicted, y represents year time granularity, q represents quarter time granularity, m represents month time granularity, w represents week time granularity, d represents day time granularity, h represents hour time granularity, and eta represents a time interval corresponding to the time to be predicted and the target time period.
S602, adjusting the time dependency corresponding to the time characteristics of the multi-dimensional time granularity through the attention module to obtain the time dependency of the scenic spot pedestrian volume of the target scenic spot in the historical target time period under the multi-dimensional time granularity.
Optionally, the weights of the temporal features of the multi-dimensional temporal granularity are dynamically adjusted based on an attention mechanism in the attention module. For example, the weight corresponding to the time characteristic of the above-mentioned multidimensional time granularity can be expressed by the following formula:
Figure BDA0003952614130000161
Figure BDA0003952614130000162
Figure BDA0003952614130000163
Figure BDA0003952614130000164
Figure BDA0003952614130000165
Figure BDA0003952614130000166
where Γ represents the target period, the weight
Figure BDA0003952614130000167
The method is used for measuring the importance degree of h hour, d days, w weeks, m months, q seasons and y years to the predicted time t.
Specifically, the definition of the weight can be expressed by the following formula:
Figure BDA0003952614130000168
Figure BDA0003952614130000169
Figure BDA00039526141300001610
Figure BDA00039526141300001611
/>
Figure BDA00039526141300001612
Figure BDA00039526141300001613
wherein the score function is defined as:
Figure BDA00039526141300001614
Figure BDA0003952614130000171
Figure BDA0003952614130000172
Figure BDA0003952614130000173
Figure BDA0003952614130000174
Figure BDA0003952614130000175
wherein, W γ1 ～W γ6 、W β1 ～W β6 、b β1 ～b β6 、V 1 ～V 6 Parameters to be learned for the time-relation model, V T Representing the transpose of V. And dynamically adjusting the weight corresponding to each time granularity based on the attention mechanism to optimize the model parameters corresponding to the time relation model.
In the embodiment of the application, the spatial dependency and the historical people flow data are processed by utilizing the bidirectional long-short term memory network and the attention mechanism in the time relation model, so that the time dependency of the scenic spot people flow of the target scenic spot in the historical target time period under the multi-dimensional time granularity is obtained, the accuracy of the time dependency of the scenic spot people flow of the target scenic spot in the historical target time period under the multi-dimensional time granularity is ensured, and the prediction precision is improved.
It can be understood that the spatial relationship model and the temporal relationship model for obtaining the spatial dependency relationship and the temporal dependency relationship may also be obtained by performing iterative training on historical data of a target scenic spot corresponding to the target time period and external related data in the above embodiment, and specifically, the determination of the optimal model parameter may be optimized by a Root Mean Square Error (RMSE) and an average Absolute Percentage Error (MAPE).
Wherein, RMSE and MAPE can be represented by the following formulas, respectively:
Figure BDA0003952614130000176
Figure BDA0003952614130000177
wherein f is i，t+1 Representing the real human flow at time t,
Figure BDA0003952614130000178
indicating the predicted traffic at time t.
The prediction accuracy can be optimized by adopting a mode of combining RMSE and MAPE for optimization. Specifically, the RMSE is sensitive to a predicted outlier, and if a certain prediction result is an outlier in the test process, the RMSE fluctuates greatly; MAPE may take into account deviations between predicted and actual values, as well as ratios between error values and actual values. Illustratively, for a certain scenic spot, the real traffic is 40 to 100 thousands, 40 thousands are predicted to be 20 thousands in the training process, and 100 thousands are predicted to be 120 thousands. And the prediction accuracy reaches the optimum only when the RMSE and MAPE trained by the model are both minimum.
The following are embodiments of the apparatus of the present application that may be used to perform embodiments of the method of the present application. For details which are not disclosed in the embodiments of the apparatus of the present application, reference is made to the embodiments of the method of the present application.
Fig. 7 is a schematic structural diagram of a device for predicting scenic spot pedestrian volume according to an embodiment of the present application. As shown in fig. 7, the scenic spot traffic prediction apparatus 70 includes: an acquisition module 710, a first determination module 720, a second determination module 730, and a prediction module 740.
The acquisition module 710 is configured to acquire historical data of a target time interval corresponding to a target scenic spot and external related data, where the target time interval is determined according to time to be predicted, the historical data includes historical pedestrian volume data of the target scenic spot in the historical target time interval and historical traffic flow data of traffic objects in a preset range of the target scenic spot under multidimensional time granularity, the external related data includes weather data of the target scenic spot in the time to be predicted, per capita total production data of the area where the target scenic spot is located, and key event data, the multidimensional time granularity includes year, quarter, month, week, day, hour, and preset time intervals, and the traffic objects include stations, gates, and landmarks; a first determining module 720, configured to determine a spatial dependency relationship between the scenic spot pedestrian volume of the target scenic spot and the traffic flow data of the traffic object according to the historical pedestrian volume data and the historical traffic flow data; the second determining module 730, configured to determine, according to the spatial dependency relationship and the historical pedestrian volume data, a temporal dependency relationship between historical target time periods of the scenic spot pedestrian volumes of the target scenic spot at the multidimensional time granularity; and the prediction module 740 is configured to predict the pedestrian volume of the target scenic spot at the time to be predicted according to the time dependency relationship and the external related data.
In a possible implementation manner, the first determining module 720 is specifically configured to: determining the distance between a target scenic spot and a traffic object; acquiring target historical traffic flow data corresponding to the target traffic object with the corresponding distance smaller than the distance threshold value from the historical traffic flow data; and determining the spatial dependence relationship between the scenic spot pedestrian volume of the target scenic spot and the traffic flow data of the target traffic object according to the historical traffic flow data of the target and the historical pedestrian volume data.
In one possible implementation, the first determining module 720 may further be configured to: and inputting the historical traffic flow data and the historical pedestrian flow data of the target into the spatial relationship model to obtain the spatial dependence relationship between the scenic spot pedestrian flow of the target scenic spot and the traffic flow data of the target traffic object, which is output by the spatial relationship model.
In one possible implementation, the spatial relationship model includes a plurality of convolutional layers; a flow control layer is arranged between the adjacent convolution layers and is used for adjusting the spatial dependence relationship by adopting a flow control mechanism; the first determination module 720 may also be configured to: inputting target historical traffic flow data and historical pedestrian flow data into the 1 st convolutional layer for spatial feature extraction to obtain spatial features output by the 1 st convolutional layer, wherein different spatial features have spatial dependency relations; adjusting the local spatial dependency relationship according to the spatial characteristics output by the ith convolutional layer through a flow control layer, wherein i sequentially takes 1,2, … … and N-1,N as the number of convolutional layers; and inputting the local spatial dependency relationship into the (i + 1) th convolution layer for spatial feature extraction to obtain the spatial feature output by the (i + 1) th convolution layer, wherein the spatial dependency relationship between the spatial features output by the (N) th convolution layer is the spatial dependency relationship between the scenic spot pedestrian flow of the target scenic spot and the traffic flow data of the target traffic object.
In a possible implementation manner, the second determining module 730 is specifically configured to: and inputting the spatial dependency relationship and historical pedestrian flow data into the time relationship model to obtain the time dependency relationship between historical target time periods of the scenic spot pedestrian flow of the target scenic spot output by the time relationship model under the multi-dimensional time granularity.
In one possible implementation, the time relationship model includes a bidirectional long-short term memory network and an attention module, and the attention module is configured to adjust the time dependency relationship using an attention mechanism; the second determination module 730 may be further configured to: inputting the historical pedestrian flow data of the space dependency relationship and the dimension time granularity into a bidirectional long and short term memory network aiming at each dimension time granularity in the multi-dimension time granularity to obtain time characteristics output by the bidirectional long and short term memory network, wherein the time characteristics have a time dependency relationship; and adjusting the time dependency corresponding to the time characteristics of the multidimensional time granularity by the attention module to obtain the time dependency of the scenic spot pedestrian volume of the target scenic spot in the historical target time period under the multidimensional time granularity.
In a possible implementation manner, the prediction module 740 is specifically configured to: and carrying out fusion processing on the time dependence relation and the external related data to obtain the pedestrian volume of the target scenic spot at the time to be predicted.
The apparatus provided in the embodiment of the present application may be configured to perform the method steps in the foregoing method embodiments, and the implementation principle and technical effects are similar, which are not described herein again.
It should be noted that the division of the modules of the above apparatus is only a logical division, and the actual implementation may be wholly or partially integrated into one physical entity, or may be physically separated. And these modules can be realized in the form of software called by processing element; or may be implemented entirely in hardware; and part of the modules can be realized in the form of calling software by the processing element, and part of the modules can be realized in the form of hardware. For example, the processing module may be a processing element that is separately configured, or may be integrated into a chip of the apparatus, or may be stored in a memory of the apparatus in the form of program code, and a processing element of the apparatus calls and executes a function of the processing module. Other modules are implemented similarly. In addition, all or part of the modules can be integrated together or can be independently realized. The processing element here may be an integrated circuit with signal processing capabilities. In implementation, each step of the above method or each module above may be implemented by an integrated logic circuit of hardware in a processor element or an instruction in the form of software.
Fig. 8 is a schematic structural diagram of an electronic device according to an embodiment of the present application. As shown in fig. 8, the electronic apparatus 80 includes: at least one processor 810, a memory 820, a communication interface 830, and a system bus 840. The memory 820 and the communication interface 830 are connected to the processor 810 through the system bus 840 and complete mutual communication, the memory 820 is used for storing instructions, the communication interface 830 is used for communicating with other devices, and the processor 810 is used for calling the instructions in the memory to execute the scheme of the above-mentioned detection method embodiment of the skid size deviation abnormality.
The system bus 840 shown in fig. 8 may be a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, or the like. The system bus 840 may be divided into an address bus, a data bus, a control bus, and the like. For ease of illustration, only one thick line is shown, but this is not intended to represent only one bus or type of bus.
The communication interface 830 is used to enable communication between the database access device and other devices (e.g., clients, read-write libraries, and read-only libraries).
The Memory 820 may include a Random Access Memory (RAM), and may further include a non-volatile Memory (non-volatile Memory), such as at least one disk Memory.
Processor 810 may be a general-purpose Processor including a central processing unit, a Network Processor (NP), etc.; the Integrated Circuit may also be a Digital Signal Processor (DSP), an Application Specific Integrated Circuit (ASIC), a Field Programmable Gate Array (FPGA) or other Programmable logic device, discrete Gate or transistor logic device, or discrete hardware components.
The embodiment of the present application further provides a computer-readable storage medium, where computer-executable instructions are stored in the computer-readable storage medium, and when the computer-executable instructions are executed by a processor, the method steps in the foregoing method embodiments are implemented, and the specific implementation manner and the technical effect are similar, and are not described herein again.
The embodiment of the application also provides a program product, and the program product comprises computer execution instructions. When the computer executes the instructions, the method steps in the above method embodiments are implemented in a similar manner and with similar technical effects, which are not described herein again.
Other embodiments of the invention will be apparent to those skilled in the art from consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of the invention following, in general, the principles of the application and including such departures from the present disclosure as come within known or customary practice within the art to which the invention pertains. It is intended that the specification and examples be considered as exemplary only, with a true scope and spirit of the application being indicated by the following claims.
It will be understood that the present application is not limited to the precise arrangements described above and shown in the drawings and that various modifications and changes may be made without departing from the scope thereof. The scope of the application is limited only by the appended claims.
It should be noted that, the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) referred to in the present application are information and data authorized by the user or fully authorized by each party, and the collection, use and processing of the related data need to comply with the relevant laws and regulations and standards of the relevant country and region, and are provided with corresponding operation entrances for the user to choose authorization or denial.

Claims (10)

1. A method for predicting the pedestrian volume in a scenic spot is characterized by comprising the following steps:
acquiring historical data and external related data of a target time interval corresponding to a target scene, wherein the target time interval is determined according to time to be predicted, the historical data comprises historical pedestrian volume data of the target scene in the historical target time interval under multi-dimensional time granularity and historical traffic flow data of traffic objects in a preset range of the target scene, the external related data comprises weather data of the target scene in the time to be predicted, per capita production total data of a region where the target scene is located and key event data, the multi-dimensional time granularity comprises year, quarter, month, week, day, hour and preset time intervals, and the traffic objects comprise stations, gates and landmarks;
according to the historical people flow data and the historical traffic flow data, determining a spatial dependency relationship between the scenic spot people flow of the target scenic spot and the traffic flow data of the traffic object;
according to the spatial dependency relationship and the historical people flow data, determining the time dependency relationship between the historical target time periods of the scenic spot people flow of the target scenic spot under the multidimensional time granularity;
and predicting the pedestrian flow of the target scenic spot at the time to be predicted according to the time dependency relationship and the external related data.
2. The prediction method according to claim 1, wherein the determining a spatial dependency relationship between the scenic spot pedestrian volume of the target scenic spot and the traffic flow data of the traffic object according to the historical pedestrian volume data and the historical traffic flow data comprises:
determining a distance between the target scenic region and the traffic object;
acquiring target historical traffic flow data corresponding to the target traffic object with the corresponding distance smaller than the distance threshold value from the historical traffic flow data;
and determining the spatial dependence relationship between the scenic spot pedestrian volume of the target scenic spot and the traffic flow data of the target traffic object according to the historical traffic flow data of the target and the historical pedestrian volume data.
3. The prediction method according to claim 2, wherein the determining a spatial dependency between the scenic spot traffic volume of the target scenic spot and the traffic flow data of the target traffic object according to the target historical traffic flow data and the historical traffic flow data comprises:
inputting the target historical traffic flow data and the historical pedestrian flow data into a spatial relationship model to obtain a spatial dependency relationship between the scenic spot pedestrian flow of the target scenic spot and the traffic flow data of the target traffic object, which is output by the spatial relationship model.
4. The prediction method of claim 3, wherein the spatial relationship model comprises a plurality of convolutional layers; a flow control layer is arranged between the adjacent convolution layers and is used for adjusting the spatial dependence relationship by adopting a flow control mechanism;
inputting the historical traffic flow data of the target and the historical pedestrian flow data into a spatial relationship model to obtain a spatial dependency relationship between the scenic spot pedestrian flow of the target scenic spot and the traffic flow data of the target traffic object, which are output by the spatial relationship model, and the spatial dependency relationship comprises the following steps:
inputting the target historical traffic flow data and the historical pedestrian flow data into a 1 st convolutional layer for spatial feature extraction to obtain spatial features output by the 1 st convolutional layer, wherein different spatial features have spatial dependency relations;
adjusting the local spatial dependency relationship according to the spatial characteristics output by the ith convolutional layer through the flow control layer, wherein i sequentially takes 1,2, … … and N-1,N as the number of convolutional layers; and inputting the local spatial dependency relationship into the (i + 1) th convolution layer for spatial feature extraction to obtain the spatial feature output by the (i + 1) th convolution layer, wherein the spatial dependency relationship between the spatial features output by the (N) th convolution layer is the spatial dependency relationship between the scenic spot pedestrian flow of the target scenic spot and the traffic flow data of the target traffic object.
5. The prediction method according to any one of claims 1 to 4, wherein the determining a temporal dependency relationship between the historical target time periods of the scenic spot traffic of the target scenic spot at the multi-dimensional time granularity according to the spatial dependency relationship and the historical traffic data comprises:
and inputting the spatial dependency relationship and the historical people flow data into a time relationship model to obtain the time dependency relationship between historical target time periods of the scenic spot people flow of the target scenic spot under the multi-dimensional time granularity, which is output by the time relationship model.
6. The prediction method of claim 5, wherein the temporal relationship model comprises a bidirectional long-short term memory network and an attention module for adapting temporal dependencies using an attention mechanism;
inputting the spatial dependency relationship and the historical people flow data into a time relationship model to obtain a time dependency relationship between historical target time periods of the scenic spot people flow of the target scenic spot output by the time relationship model under the multi-dimensional time granularity, wherein the time dependency relationship comprises:
inputting the spatial dependency relationship and historical people flow data of the dimensional time granularity into a bidirectional long and short term memory network aiming at each dimensional time granularity in the multi-dimensional time granularity to obtain time characteristics output by the bidirectional long and short term memory network, wherein the time characteristics have a time dependency relationship;
and adjusting the time dependency corresponding to the time characteristics of the multidimensional time granularity by the attention module to obtain the time dependency of the scenic spot pedestrian volume of the target scenic spot in the historical target time period under the multidimensional time granularity.
7. The prediction method according to any one of claims 1 to 4, wherein the predicting the pedestrian volume of the target scenic spot at the time to be predicted according to the time dependency relationship and the external relevant data comprises:
and carrying out fusion processing on the time dependency relationship and the external related data to obtain the pedestrian volume of the target scenic spot at the time to be predicted.
8. An apparatus for predicting a traffic of persons in a scenic spot, comprising:
the system comprises an acquisition module, a prediction module and a prediction module, wherein the acquisition module is used for acquiring historical data and external related data of a target time interval corresponding to a target scenic spot, the target time interval is determined according to time to be predicted, the historical data comprises historical pedestrian volume data of the target scenic spot in the historical target time interval under multidimensional time granularity and historical traffic flow data of traffic objects in a preset range of the target scenic spot, the external related data comprises weather data of the target scenic spot in the time to be predicted, per capita production total data of the region of the target scenic spot and key event data, the multidimensional time granularity comprises year, quarter, month, week, day, hour and preset time intervals, and the traffic objects comprise stations, gates and landmarks;
the first determination module is used for determining the spatial dependence relationship between the scenic spot pedestrian volume of the target scenic spot and the traffic flow data of the traffic object according to the historical pedestrian volume data and the historical traffic flow data;
a second determining module, configured to determine, according to the spatial dependency relationship and the historical pedestrian volume data, a temporal dependency relationship between scenic spot pedestrian volumes of the target scenic spot in historical target time periods under the multidimensional time granularity;
and the prediction module is used for predicting the pedestrian volume of the target scenic spot at the time to be predicted according to the time dependency relationship and the external related data.
9. An electronic device, comprising:
at least one processor;
and a memory coupled to the at least one processor;
wherein the memory is configured to store instructions executable by the at least one processor to enable the at least one processor to perform a method of predicting scenic spot traffic as claimed in any one of claims 1 to 7.
10. A computer-readable storage medium having computer-executable instructions stored thereon, the computer-executable instructions when executed being configured to implement a method of predicting scenic spot traffic as claimed in any one of claims 1 to 7.
CN202211453713.4A 2022-11-21 2022-11-21 Scenic spot pedestrian flow prediction method, device, equipment and storage medium Pending CN115936762A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202211453713.4A CN115936762A (en) 2022-11-21 2022-11-21 Scenic spot pedestrian flow prediction method, device, equipment and storage medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202211453713.4A CN115936762A (en) 2022-11-21 2022-11-21 Scenic spot pedestrian flow prediction method, device, equipment and storage medium

Publications (1)

Publication Number Publication Date
CN115936762A true CN115936762A (en) 2023-04-07

Family

ID=86655020

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202211453713.4A Pending CN115936762A (en) 2022-11-21 2022-11-21 Scenic spot pedestrian flow prediction method, device, equipment and storage medium

Country Status (1)

Country Link
CN (1) CN115936762A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117218604A (en) * 2023-11-07 2023-12-12 北京城建智控科技股份有限公司 Method and device for supplementing missing passenger flow information, electronic equipment and storage medium

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117218604A (en) * 2023-11-07 2023-12-12 北京城建智控科技股份有限公司 Method and device for supplementing missing passenger flow information, electronic equipment and storage medium
CN117218604B (en) * 2023-11-07 2024-04-12 北京城建智控科技股份有限公司 Method and device for supplementing missing passenger flow information, electronic equipment and storage medium

Similar Documents

Publication Publication Date Title
Xu et al. Real‐time regional seismic damage assessment framework based on long short‐term memory neural network
JP6770125B2 (en) Estimated damage prevention due to building renovation
CN109658695B (en) Multi-factor short-term traffic flow prediction method
Cimellaro et al. Seismic resilience of a hospital system
CN111091196B (en) Passenger flow data determination method and device, computer equipment and storage medium
Mesta et al. Urban growth modelling and social vulnerability assessment for a hazardous Kathmandu Valley
EP3736735A1 (en) Apparatus and method for providing application service using satellite image
CN110503267B (en) Urban financial invasion case prediction system and prediction method based on space-time scale self-adaptive model
Tomar et al. Active learning method for risk assessment of distributed infrastructure systems
CN111861028A (en) Method for predicting crime number based on spatio-temporal data fusion
CN115936762A (en) Scenic spot pedestrian flow prediction method, device, equipment and storage medium
EP3192061B1 (en) Measuring and diagnosing noise in urban environment
CN114936620B (en) Sea surface temperature numerical forecasting deviation correcting method based on attention mechanism
CN117349795B (en) Precipitation fusion method and system based on ANN and GWR coupling
Bodenmann et al. Dynamic post-earthquake updating of regional damage estimates using Gaussian processes
CN117131991A (en) Urban rainfall prediction method and platform based on hybrid neural network
CN108712317B (en) Urban crowd space-time dynamic sensing method and system based on mobile social network
Chen et al. Construction of a sediment disaster risk assessment model
CN115830865A (en) Vehicle flow prediction method and device based on adaptive hypergraph convolution neural network
CN115640756A (en) Parking demand prediction model and method based on multi-source data and application thereof
CN115310672A (en) City development prediction model construction method, city development prediction method and device
Xue et al. Urban population density estimation based on spatio‐temporal trajectories
CN110347938B (en) Geographic information processing method and device, electronic equipment and medium
CN115985086A (en) Traffic data completion method, system, terminal and storage medium
Bodenmann et al. Using gaussian process models for dynamic post-earthquake impact estimation with regional risk predictors

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination