WO2024082848A1 - 客流预测方法、装置和系统 - Google Patents

客流预测方法、装置和系统 Download PDF

Info

Publication number
WO2024082848A1
WO2024082848A1 PCT/CN2023/116913 CN2023116913W WO2024082848A1 WO 2024082848 A1 WO2024082848 A1 WO 2024082848A1 CN 2023116913 W CN2023116913 W CN 2023116913W WO 2024082848 A1 WO2024082848 A1 WO 2024082848A1
Authority
WO
WIPO (PCT)
Prior art keywords
preset
nodes
personnel
node
passenger flow
Prior art date
Application number
PCT/CN2023/116913
Other languages
English (en)
French (fr)
Inventor
罗静
孔祥斌
刘阳
李洪研
李懿祖
刘媛媛
王雪嵩
Original Assignee
通号通信信息集团有限公司
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 通号通信信息集团有限公司 filed Critical 通号通信信息集团有限公司
Publication of WO2024082848A1 publication Critical patent/WO2024082848A1/zh

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q10/00—Administration; Management
    • G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/50—Context or environment of the image
    • G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
    • G06V20/53—Recognition of crowd images, e.g. recognition of crowd congestion

Definitions

  • the present disclosure relates to the technical field of passenger flow prediction, and in particular to a passenger flow prediction method, device and system.
  • Cameras are usually installed in station waiting halls or subway platforms so that the flow of people in different areas can be monitored using the cameras.
  • the embodiments of the present disclosure provide a passenger flow prediction method and device, system, electronic device, and computer-readable storage medium, which can accurately predict personnel flow information in a preset area to ensure the safety of the preset area.
  • an embodiment of the present disclosure provides a passenger flow prediction method, the method comprising: obtaining the connectivity relationship between each preset node and its adjacent nodes within a preset area, the preset area comprising multiple preset nodes and connecting channels therebetween, the connectivity relationship being used to characterize the flow direction of the connecting channels; obtaining personnel density information corresponding to multiple preset nodes; and predicting personnel flow information within the preset area based on the personnel density information corresponding to multiple preset nodes and the connectivity relationship between each preset node and its adjacent nodes.
  • the method further includes:
  • an adjacency topology model is constructed.
  • the adjacency topology model is represented in the form of an N-order square matrix, where N is an integer greater than 1.
  • obtaining personnel density information corresponding to a plurality of preset nodes includes:
  • Acquire multiple frames of images to be processed wherein the images to be processed are images obtained by using at least one image acquisition device set in each preset node to acquire images of people in each preset node;
  • Multiple frames of images to be processed are analyzed according to a deep learning algorithm to determine the personnel density information corresponding to at least one preset node.
  • the method before obtaining the personnel density information corresponding to the plurality of preset nodes, the method further includes:
  • the personnel flow information in the multiple connection channels is counted according to the passenger flow counting algorithm to obtain passenger flow statistical information corresponding to the multiple connection channels.
  • passenger flow statistics include: the number of people passing through the connecting passage;
  • the population flow information in the preset area is predicted, including:
  • a weight coefficient matrix corresponding to each connection channel is determined
  • the personnel flow information in the preset area is predicted to obtain the regional flow prediction results.
  • the passenger flow statistics information further comprises: channel statistics error information;
  • the method Before obtaining the regional flow prediction result, the method also includes:
  • the linear estimation model is adjusted according to the channel statistical error information and the preset passenger flow total amount information of the preset node to obtain an updated linear estimation model.
  • the method further includes:
  • the personnel density information and/or passenger flow statistics information are updated at preset intervals.
  • each preset node includes: an entrance and an exit;
  • Get the personnel density information corresponding to multiple preset nodes including:
  • the population density information corresponding to the preset node is determined; wherein the preset node includes: at least one of: an entry node, a security check node, a ticket checking node, a waiting node and a boarding platform node.
  • an embodiment of the present disclosure provides a passenger flow prediction device, which includes: a first acquisition module, configured to obtain the connectivity relationship between each preset node and its adjacent nodes in a preset area, the preset area includes multiple preset nodes and connecting channels therebetween, and the connectivity relationship is used to characterize the flow direction of the connecting channels; a second acquisition module, configured to obtain personnel density information corresponding to multiple preset nodes; a prediction module, configured to predict personnel flow information in the preset area based on the personnel density information corresponding to multiple preset nodes and passenger flow statistical information in multiple connecting channels.
  • an embodiment of the present disclosure provides a passenger flow prediction system, which includes: a first image acquisition device arranged in a plurality of preset nodes, a second image acquisition device arranged in a connecting channel between the preset nodes, and a passenger flow prediction device;
  • the first image acquisition device is configured to acquire images of personnel in the preset nodes corresponding to the first image acquisition device, obtain a plurality of frames of images to be processed, and send the plurality of frames of images to be processed to the passenger flow prediction device, so that the passenger flow prediction device analyzes the plurality of frames of images to be processed according to a deep learning algorithm, and determines personnel density information corresponding to at least one preset node;
  • the second image acquisition device is configured to acquire images of personnel movement in the connecting channel between the preset nodes, obtain channel images, analyze the channel images, obtain passenger flow statistical information corresponding to the connecting channel, and send the passenger flow statistical information to the passenger flow prediction device, so that the passenger flow prediction device performs statistics on personnel flow information in a plurality of connecting channels according
  • an embodiment of the present disclosure provides an electronic device, including a memory, a processor, and a memory stored in the memory
  • a computer program that can be run on a processor, and when the processor executes the computer program, any passenger flow prediction method of the embodiments of the present disclosure is implemented.
  • an embodiment of the present disclosure provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any passenger flow prediction method of the embodiment of the present disclosure.
  • the connectivity relationship between each preset node and its adjacent nodes in the preset area it is possible to clarify whether there is a connection channel between multiple preset nodes in the preset area, as well as the flow direction of the connection channel, which is convenient for subsequent processing; by obtaining the personnel density information corresponding to multiple preset nodes in the preset area, it is possible to clarify the personnel density in different preset nodes, which is convenient for controlling the number of people in different preset nodes, wherein the preset area includes multiple preset nodes and the connection channels therebetween; based on the personnel density information corresponding to multiple preset nodes and the connectivity relationship between each preset node and its adjacent nodes, the personnel flow information in the preset area is predicted to achieve accurate prediction of the personnel flow information in the preset area, thereby achieving scheduling and management of the personnel flow between different preset nodes, and improving the safety in the preset area.
  • FIG1 is a flow chart showing a passenger flow prediction method provided by an embodiment of the present disclosure.
  • FIG2 is a flow chart showing a passenger flow prediction method provided by an embodiment of the present disclosure.
  • FIG3 shows a statistical diagram of the density of personnel within a preset node provided by an embodiment of the present disclosure.
  • FIG4 shows a block diagram of a passenger flow prediction device provided by an embodiment of the present disclosure.
  • FIG5 shows a block diagram of a passenger flow prediction system provided by an embodiment of the present disclosure.
  • FIG6 is a flow chart showing a working method of a passenger flow prediction system provided by an embodiment of the present disclosure.
  • FIG. 7 shows a schematic diagram of an adjacency topology model provided by an embodiment of the present disclosure.
  • FIG8 shows a block diagram of a composition of an electronic device provided by an embodiment of the present disclosure.
  • FIG. 9 shows a block diagram of a composition of an electronic device provided by an embodiment of the present disclosure.
  • Cameras are usually installed in station waiting rooms or subway platforms to monitor the flow of people in different areas.
  • the deployment location of the camera and the distance from the people will affect the monitoring effect, and there will be a problem of blind spots in monitoring, which makes it impossible to accurately predict the flow direction and flow of people in the area, reducing the safety of crowded places such as station waiting rooms and subway platforms.
  • the present disclosure provides a passenger flow prediction method.
  • the passenger flow prediction method of the present disclosure can be It is executed by a corresponding passenger flow prediction device, which can be implemented in software and/or hardware and can generally be integrated into an electronic device.
  • FIG1 is a flow chart of a passenger flow prediction method provided by an embodiment of the present disclosure. As shown in FIG1 , the passenger flow prediction method includes but is not limited to the following steps:
  • Step S101 obtaining the connectivity relationship between each preset node and its adjacent nodes in a preset area.
  • the preset area includes a plurality of preset nodes and connection channels therebetween.
  • a preset node has at least one adjacent node to determine whether multiple preset nodes are connected, thereby determining the connectivity relationship between the multiple preset nodes, and the connectivity relationship is used to characterize the flow direction of the connection channel.
  • Step S102 obtaining personnel density information corresponding to a plurality of preset nodes.
  • the personnel density information may include: the number of personnel existing in the preset node in real time, the total number of personnel that the preset node can accommodate, and the ratio of the number of personnel existing in the preset node in real time to the total number of personnel that the preset node can accommodate.
  • the personnel density information can be used to characterize the real-time personnel situation in the preset node, so as to control the personnel in the preset node and ensure the safety of the personnel.
  • Step S103 predicting the flow of people in the preset area based on the population density information corresponding to the plurality of preset nodes and the connectivity relationship between each preset node and its adjacent nodes.
  • the personnel density information includes the number of personnel and personnel distribution area information included in each preset node, which can clearly indicate how many personnel are in the preset node.
  • the connectivity relationships between different preset nodes are different, and the connectivity relationships corresponding to multiple preset nodes can be used to characterize the flow direction of the connection channels between the preset nodes, thereby clarifying from which preset node the personnel may flow to which preset node.
  • the flow direction of people between the preset nodes in the preset area and the flow of people can be predicted, and the flowing personnel can be managed and scheduled to improve the safety in the preset area.
  • the connectivity relationship between each preset node and its adjacent nodes in the preset area it can be clarified whether there is a connection channel between multiple preset nodes in the preset area, as well as the flow direction of the connection channel, which is convenient for subsequent processing; by obtaining the personnel density information corresponding to multiple preset nodes in the preset area, it is possible to clarify the personnel density in different preset nodes, which is convenient for controlling the number of people in different preset nodes, wherein the preset area includes multiple preset nodes and the connection channels therebetween; based on the personnel density information corresponding to multiple preset nodes and the connectivity relationship between each preset node and its adjacent nodes, the personnel flow information in the preset area is predicted to achieve accurate prediction of the personnel flow information in the preset area, thereby achieving scheduling and management of the personnel flow between different preset nodes, and improving the safety in the preset area.
  • FIG2 is a flow chart of a passenger flow prediction method provided by an embodiment of the present disclosure.
  • the passenger flow prediction method can be applied to a passenger flow prediction device. As shown in FIG2 , the passenger flow prediction method includes but is not limited to the following steps:
  • Step S201 obtaining the connectivity relationship between each preset node and its adjacent nodes.
  • step S201 in this embodiment is the same as step S101 in the previous embodiment, and will not be described in detail here.
  • Step S202 construct an adjacency topology model based on a plurality of preset nodes and a connectivity relationship between at least one preset node and its adjacent nodes.
  • the adjacency topology model is a model in which a plurality of preset nodes are nodes and a connection channel between two adjacent preset nodes is an edge.
  • the adjacency topology model includes 5 preset nodes (a first preset node, a second preset node, a third preset node, a fourth preset node, and a fifth preset node), wherein the first preset node is connected to the second preset node and the third preset node, respectively, while the fourth preset node and the fifth preset node are not connected.
  • 5 preset nodes a first preset node, a second preset node, a third preset node, a fourth preset node, and a fifth preset node
  • the constructed adjacency topology model may include a connectivity value between the first preset node and the second preset node (for example, set to 1), a connectivity value between the first preset node and the third preset node (for example, set to 1), and a connectivity value between the fourth preset node and the fifth preset node (for example, set to 0). Based on the determined adjacency topology model, the connectivity relationship between each preset node can be intuitively known.
  • the adjacency topology model is represented in the form of an N-order square matrix, where N is an integer greater than 1.
  • the N-order square matrix is a matrix of N rows and N columns.
  • the value of the element in the 1st row and 1st column of the 3rd-order square matrix indicates that the 1st preset node is connected to itself (that is, the value is 1);
  • the value of the element in the 2nd row and 1st column indicates whether the 2nd preset node is connected to the 1st preset node (that is, if connected, the value is 1; otherwise, the value is 0);
  • the value of each element in the 3rd-order square matrix can be known, so as to intuitively obtain the connectivity between each preset node in the adjacency topology model.
  • the preset nodes include: at least one of: an entry node, a security check node, a ticket checking node, a waiting node and a boarding platform node.
  • different preset nodes can be processed in multiple dimensions, so that the passenger flow prediction method is applicable to more usage scenarios and the application scope of the passenger flow prediction method is expanded.
  • Step S203 obtaining personnel density information corresponding to multiple preset nodes.
  • Step S204 predicting the flow of people in the preset area based on the population density information corresponding to the plurality of preset nodes and the connectivity relationship between each preset node and its adjacent nodes.
  • steps S203 to S204 in this embodiment are the same as steps S102 to S103 in the previous embodiment, and will not be described in detail herein.
  • the connectivity relationship between the multiple preset nodes can be intuitively clarified, which is convenient for subsequent processing; and by obtaining the personnel density information corresponding to the multiple preset nodes in the preset area, the personnel density situation in different preset nodes can be clarified, which is convenient for controlling the number of people in different preset nodes; based on the personnel density information corresponding to the multiple preset nodes and the passenger flow statistics information in the multiple connecting channels, the personnel flow information in the preset area is predicted to achieve accurate prediction of the personnel flow information in the preset area, thereby realizing the scheduling and management of the personnel flow between different preset nodes, and improving the safety in the preset area.
  • the obtaining of the personnel density corresponding to the plurality of preset nodes in step S102 or step S203 The information can be realized in the following manner: obtaining multiple frames of images to be processed; analyzing the multiple frames of images to be processed according to a deep learning algorithm to determine the personnel density information corresponding to at least one preset node.
  • the image to be processed is an image obtained by capturing an image of a person in each preset node using at least one image capture device arranged in each preset node.
  • At least one image acquisition device may be provided in each preset node to facilitate accurate acquisition of images of persons in the preset node.
  • the preset node is a ticketing node
  • FIG3 shows a statistical diagram of the density of people in the preset node provided by an embodiment of the present disclosure.
  • the image on the left side of FIG3 is used to represent the image captured by the camera installed in the ticketing node, wherein it can be known that there are multiple people in the ticketing node, wherein different people correspond to different distribution areas.
  • the personnel density statistical algorithm the image on the left side of FIG3 is analyzed, and the personnel distribution information shown on the right side of FIG3 can be obtained (e.g., based on the position information corresponding to the head, the number of people in the ticketing node is counted to obtain personnel density information, etc.).
  • the number of heads shown on the right side of FIG. 3 may be counted to obtain the number of persons existing in the ticketing node.
  • the number of people in the preset node and the personnel density information can be obtained, and the personnel density in the preset node can be intelligently counted to achieve real-time monitoring and management of the people in the preset node, thereby improving the safety of the people in the preset node.
  • the preset nodes include a ticket selling node and a ticket checking node. At least one image acquisition device arranged in the ticket selling node is used to acquire images of persons in the ticket selling node, and at least one image acquisition device arranged in the ticket checking node is used to acquire images of persons in the ticket checking node.
  • At least one image acquisition device arranged in the ticket selling node is used to acquire images of persons in the ticket selling node
  • at least one image acquisition device arranged in the ticket checking node is used to acquire images of persons in the ticket checking node.
  • multiple frames of images to be processed corresponding to the ticket selling node and multiple frames of images to be processed corresponding to the ticket checking node are analyzed to determine the personnel density information corresponding to the ticket selling node and/or the personnel density information corresponding to the ticket checking node, so as to facilitate the comprehensive scheduling of personnel in these two nodes.
  • each preset node includes: an entrance and an exit; the obtaining of the personnel density information corresponding to the plurality of preset nodes in step S102 or step S203 may be implemented in the following manner:
  • First communication interaction information passing through the entrance of each preset node is obtained; second communication interaction information passing through the exit of each preset node is obtained; and personnel density information corresponding to the preset node is determined based on the first communication interaction information and the second communication interaction information.
  • the first communication interaction information is information between the devices carried by each person passing through the entrance and the preset devices set at the entrance
  • the second communication interaction information is information between the devices carried by each person passing through the exit and the preset devices set at the exit.
  • the preset device may be a detection device set at an entrance or exit.
  • the preset device is a gate that can swipe a card.
  • a person passes through the entrance or exit, he or she interacts with the gate through the card to be verified (such as a ticket, a card representing personal identity information, etc.) and obtains the first communication interaction information and the second communication interaction information. interest.
  • the first communication interaction information and the second communication interaction information are summarized to the server through the gates at the entrance and exit, so that the server can count the number of people passing through the entrance and exit, and verify their identities, etc., so as to know the number of people entering and leaving the preset node, and then, when the real-time number of people in the preset node is clear, the population density information corresponding to the preset node can be determined.
  • the personnel density information in each preset node can be monitored in real time, so that when an abnormal situation occurs (such as the number of people in the preset node exceeds the preset threshold, etc.), the passenger flow prediction device can promptly manage and control the personnel in the preset node to improve the safety of the personnel.
  • an abnormal situation such as the number of people in the preset node exceeds the preset threshold, etc.
  • the method before executing step S102 or step S203 to obtain the personnel density information corresponding to the plurality of preset nodes, the method further includes:
  • the passenger flow statistics algorithm can be a statistical algorithm determined based on the inflow and outflow of personnel, so as to facilitate the statistics of personnel flow information in multiple connection channels, so that the passenger flow statistics information corresponding to the multiple connection channels obtained can reflect the personnel flow situation in multiple connection channels (such as the movement direction and number of personnel in a certain connection channel).
  • passenger flow statistics information includes: the number of people passing through the connecting channel and the flow direction of the connecting channel.
  • Step S103 or step S204 predicts the flow of people in a preset area based on the population density information corresponding to the plurality of preset nodes and the connectivity relationship between each preset node and its adjacent nodes, which can be implemented in the following manner:
  • the flow direction of each connecting channel is determined; according to the flow direction of each connecting channel, the weight coefficient matrix corresponding to each connecting channel is determined; according to the weight coefficient matrix, the personnel density information and the number of personnel passing through the connecting channel, the linear estimation model is determined; according to the linear estimation model, the personnel flow information in the preset area is predicted to obtain the regional flow prediction result.
  • the weight coefficient matrix may include multiple weight values, which are used to represent the proportion of movement in a preset direction within the connection channel (for example, in the connection channel between the first preset node and the second preset node, the proportion may represent the proportion of people moving from the first preset node to the second preset node, or the proportion of people moving from the second preset node to the first preset node, etc.).
  • each preset node through the connectivity relationship between each preset node and its adjacent nodes, it is possible to clarify which adjacent nodes a preset node has a connectivity relationship with, thereby determining to which adjacent node the personnel in the preset node may flow, and further determining the flow direction of the connecting channels between the preset nodes.
  • an offline estimation model is established to clarify the personnel flow situation in different preset nodes and their corresponding connecting channels, and then the personnel flow information in the preset area is predicted through the linear estimation model, so as to predict the personnel flow situation between multiple preset nodes in the preset area (i.e., the regional flow prediction result).
  • the regional traffic forecast results can be used to obtain information such as the number of people that may exist in the preset area in the future and the direction of personnel flow, so as to facilitate timely control of personnel in the preset area. For example, when it is determined that there are too many people in the preset area, the flow of personnel can be accelerated by adding detection checkpoints to ensure the safety of personnel in the preset area; or, when it is determined that there are too few people in the preset area, the number of people entering the entrance can be increased to avoid idle equipment or management personnel and improve the equipment utilization efficiency in the preset area.
  • the passenger flow statistics information also includes: channel statistical error information. Before predicting the personnel flow information in a preset area based on the linear estimation model and obtaining the regional flow prediction result, the method also includes: adjusting the linear estimation model based on the channel statistical error information and the preset passenger flow total amount information of the preset node to obtain an updated linear estimation model.
  • the channel statistical error information is used to characterize the error information that may exist when counting the people in the connection channel.
  • the channel statistical error information may include the identifier of the connection channel and the statistical error value corresponding to the connection channel (for example, the statistical error value may be 2, 3, etc.).
  • the linear estimation model is adjusted by using the channel statistical error information and the preset passenger flow total information of the preset node so that the linear estimation model can be closer to the actual personnel statistics in the area, and then when the updated linear estimation model is used to predict the personnel in the preset area, the accuracy of the prediction results can be improved.
  • the method after executing step S103 or step S204 to predict the personnel flow information in a preset area based on the personnel density information corresponding to multiple preset nodes and the connectivity relationship between each preset node and its adjacent nodes, the method also includes: updating the personnel density information and passenger flow statistics information at every preset time interval.
  • the preset time lengths may include: 1 hour, 2 hours, etc.
  • the above preset time lengths are only examples and can be specifically set according to actual needs. Other unspecified preset time lengths are also within the protection scope of this disclosure and will not be repeated here.
  • the population density within the preset node and the passenger flow statistics in the connection channel between the preset node and its adjacent nodes can be timely known, thereby realizing timely monitoring and management of the personnel situation in different areas.
  • the embodiments of the present disclosure provide a passenger flow prediction device, which is a corresponding device for implementing the passenger flow prediction method provided in the above embodiments of the present disclosure, and the device can be implemented in software and/or hardware, and can generally be integrated into an electronic device.
  • Fig. 4 shows a block diagram of a passenger flow prediction device provided by an embodiment of the present disclosure.
  • the passenger flow prediction device 400 includes but is not limited to the following modules.
  • the first acquisition module 410 is configured to acquire the connectivity relationship between each preset node and its adjacent nodes in a preset area.
  • the preset area includes a plurality of preset nodes and connection channels therebetween.
  • the connectivity relationship is used to characterize the flow direction of the connection channels.
  • the second acquisition module 420 is configured to acquire personnel density information corresponding to a plurality of preset nodes.
  • the prediction module 430 is configured to calculate the number of personnel density information corresponding to the plurality of preset nodes and the number of connection channels. passenger flow statistics information and predict personnel flow information in the preset area.
  • the device in this embodiment can execute any passenger flow prediction method in the embodiments of the present disclosure, and its specific implementation is not limited to the above embodiments, and other undescribed embodiments are also within the protection scope of this device.
  • the connectivity relationship between each preset node and its adjacent nodes in the preset area is obtained by the first acquisition module, so it can be clarified whether there is a connection channel between multiple preset nodes in the preset area, and the flow direction of the connection channel, which is convenient for subsequent processing;
  • the second acquisition module is used to obtain the personnel density information corresponding to multiple preset nodes in the preset area, so it can be clarified that the personnel density situation in different preset nodes, which is convenient for controlling the number of people in different preset nodes, wherein the preset area includes multiple preset nodes and the connection channels therebetween;
  • the prediction module is used to predict the personnel flow information in the preset area based on the personnel density information corresponding to multiple preset nodes and the connectivity relationship between each preset node and its adjacent nodes, so as to realize accurate prediction of the personnel flow information in the preset area, thereby realizing the scheduling and management of the personnel flow between different preset nodes, and improving the safety in the preset area.
  • modules involved in this embodiment are logic modules.
  • a logic unit can be a physical unit, a part of a physical unit, or a combination of multiple physical units.
  • this embodiment does not introduce units that are not closely related to solving the technical problems proposed by the present disclosure, but this does not mean that there are no other units in this embodiment.
  • Fig. 5 shows a block diagram of a passenger flow prediction system provided by an embodiment of the present disclosure. As shown in Fig. 5, the passenger flow prediction system includes but is not limited to the following devices.
  • a first image acquisition device 510 disposed in a plurality of preset nodes, a second image acquisition device 520 disposed in a connection channel between the preset nodes, and a passenger flow prediction device 530;
  • the first image acquisition device 510 is configured to acquire images of persons within a preset node corresponding to the first image acquisition device 510, obtain multiple frames of images to be processed, and send the multiple frames of images to be processed to the passenger flow prediction device 530, so that the passenger flow prediction device 530 analyzes the multiple frames of images to be processed according to a deep learning algorithm to determine the personnel density information corresponding to at least one preset node.
  • the second image acquisition device 520 is configured to acquire images of personnel movement in the connection channel between preset nodes, obtain channel images, analyze the channel images, obtain passenger flow statistical information corresponding to the connection channel, and send the passenger flow statistical information to the passenger flow prediction device 530, so that the passenger flow prediction device 530 can perform statistics on the personnel flow information in multiple connection channels according to the passenger flow statistical algorithm, and obtain the passenger flow statistical information corresponding to the connection channel.
  • the passenger flow prediction device 530 is configured to execute any passenger flow prediction method in the embodiments of the present disclosure.
  • Fig. 6 shows a schematic flow chart of a working method of a passenger flow prediction system provided by an embodiment of the present disclosure. As shown in Fig. 6, the working method of the passenger flow prediction system includes but is not limited to the following steps.
  • Step S601 construct an adjacency topology model according to a plurality of preset nodes in a preset area and the connectivity relationship between each preset node and its adjacent nodes.
  • the adjacency topology model can be represented in the form of an N-order square matrix, where N is an integer greater than 1.
  • an N-order square matrix M is a matrix with N rows and N columns.
  • Each element in the N-order square matrix M can be expressed as M (i, j) .
  • the value of M (i, j) indicates whether the i-th preset node is connected to the j-th preset node.
  • i and j are both integers greater than or equal to 1 and less than or equal to N.
  • FIG7 shows a schematic diagram of an adjacency topology model provided by an embodiment of the present disclosure.
  • the preset area e.g., a station
  • the preset area includes the following preset nodes: station entrance 1, security checkpoint 2, ticket gate 3, first waiting hall 4, second waiting hall 5, and platform 6.
  • FIG7 shows the connection relationship between the preset nodes.
  • a 6*6 matrix is constructed, and the matrix is used as the adjacency topology model.
  • the constructed adjacency topology model is expressed as follows using formula (1):
  • M (1, 1) 1, which means that entrance 1 is connected to entrance 1
  • Step S602 Obtain personnel density information corresponding to a plurality of preset nodes in a preset area.
  • images of people in the preset node i.e., multiple frames of images to be processed
  • the image acquisition device installed in the preset node such as ticket gate 3, first waiting hall 4, etc.
  • the multiple frames of images to be processed are analyzed according to the deep learning algorithm to determine the personnel density information corresponding to each preset node.
  • Step S603 obtaining passenger flow statistics information in a corridor or aisle where a connection relationship exists between two adjacent preset nodes.
  • the passenger flow statistics information between the security checkpoint 2 and the ticket gate 3 can be obtained.
  • the collected images can be analyzed to obtain passenger flow statistics information (i.e., information such as the number of people and the direction of people flow) passing through the corridor within a preset time period.
  • the passenger flow statistics information between other adjacent preset nodes is the same as the above method, and the number of people in the connecting channel between the i-th preset node and the j-th preset node can be recorded as a ij for subsequent use.
  • Step S604 determining a weight coefficient matrix passing through each connection channel according to the flow direction of each connection channel.
  • the directional flow direction between different preset nodes can be clarified.
  • people in entrance 1 can only flow to security checkpoint 2, so that the ratio of the number of mobile people in the connecting channel between entrance 1 and security checkpoint 2 to the total number of people in entrance 1 can be clarified, and the ratio value is used as the weight coefficient; based on different connecting channels, the corresponding weight coefficients are different.
  • the weight coefficient matrix W includes multiple elements (e.g., the element is represented by wij , that is, the proportion of people flowing from the i-th preset node to the j-th preset node, for example, ), and 0 ⁇ w ij ⁇ 1.
  • the value of p i corresponding to the image collected by the camera installed in the i-th preset node can be updated to obtain the number of people flowing from the i-th preset node to its adjacent preset node.
  • the value of w ij can be updated every preset time interval (such as 1 hour, 2 hours, etc.) to make the prediction result more accurate.
  • Step S605 determining a linear estimation model according to the weight coefficient matrix, the personnel density information and the number of personnel passing through the connection channel.
  • the linear estimation model is used to characterize the change value of the number of personnel at different preset nodes.
  • the linear estimation model can be represented by formula (2).
  • mi represents the change value of the number of personnel in the i-th preset node; Indicates the number of people flowing into the i-th preset node; Indicates the number of people flowing out of the i-th node.
  • K represents the number of outflow nodes in the preset nodes
  • L represents the number of inflow nodes in the preset nodes, K and L are both integers greater than or equal to 0;
  • k is an integer greater than or equal to 0 and less than or equal to K;
  • l is an integer greater than or equal to 0 and less than or equal to L.
  • i is an integer greater than or equal to 1 and less than or equal to the number of preset nodes in the preset area.
  • w ki represents the proportion of people flowing from the kth preset node to the ith preset node within time t 1 ;
  • a ki represents the number of people in the connection channel between the kth preset node and the ith preset node within time t 1 (that is, the number of people existing in the connection channel in real time);
  • p k represents the total number of people in the kth preset node within time t 1 .
  • Step S606 predicting the personnel flow information in the preset area based on the linear estimation model to obtain the regional flow prediction result.
  • the regional traffic prediction result can be obtained by using a linear estimation model to minimize the difference between the actual change in the number of people at a preset node and the estimated value.
  • the objective function i.e., regional flow prediction result
  • the objective function can be expressed by the following formula (3):
  • min ⁇ 1 means that for a preset node i, taking 1 as the total amount of personnel outflow, the minimum value of the change value corresponding to the preset node i is determined, so as to obtain the flow prediction result of the preset area including multiple preset nodes.
  • I represents the number of preset nodes included in the preset area, I is an integer greater than or equal to 1, and i is a large An integer greater than or equal to 1 and less than or equal to 1.
  • the following constraints are also required: That is, for a preset node i, the total outflow of personnel is 1.
  • M represents the maximum number of personnel changes in the preset node i, M is an integer greater than or equal to 1, and m is an integer greater than or equal to 0 and less than or equal to M.
  • the above prediction process can also be constrained based on channel statistical error information within a connection channel (e.g., a connection channel from the i-th preset node to the j-th preset node, etc.), such as using formula (4) to represent the channel statistical error information.
  • a connection channel e.g., a connection channel from the i-th preset node to the j-th preset node, etc.
  • e ij represents the number of people flowing from the i-th preset node to the j-th preset node (that is, the number of people who have flowed into the j-th preset node);
  • ⁇ represents the statistical error value, for example, the value of ⁇ can be 2, 3, etc.
  • the linear estimation model is adjusted according to the channel statistical error information and the preset passenger flow total information of the preset node to obtain an updated linear estimation model.
  • the updated linear estimation model can be expressed by formula (5):
  • e ij represents the number of people flowing from the i-th preset node to the j-th preset node (that is, the number of people who have flowed into the j-th preset node).
  • the number of personnel flowing from the i-th preset node to the j-th preset node can be further obtained: that is, w ij a ij p i .
  • an adjacency topology model is constructed by connecting multiple preset nodes in a preset area and the connectivity between the nodes and the adjacent nodes, so as to predict the flow of people in the preset area based on the adjacency topology model, and facilitate the management of people in the preset area; and, through a deep learning algorithm, multiple frames of images to be processed collected by cameras in different preset nodes are analyzed to determine the population density information corresponding to each preset node, and the population density in the preset nodes can be intelligently counted to realize real-time monitoring and supervision of the people in the preset nodes; based on the population density information corresponding to multiple preset nodes and the passenger flow statistics information in multiple connecting channels, the population flow information in the preset area is predicted to realize accurate prediction of the flow information of people in the preset area, thereby realizing the scheduling and management of the flow of people between different preset nodes, and improving the safety in the preset area.
  • FIG8 shows a block diagram of a composition of an electronic device provided by an embodiment of the present disclosure.
  • an embodiment of the present disclosure provides an electronic device, the electronic device comprising: at least one processor 801; at least one memory 802, and one or more I/O interfaces 803 connected between the processor 801 and the memory 802; wherein the memory 802 stores one or more computer programs that can be executed by the at least one processor 801. Program, one or more computer programs are executed by at least one processor 801, so that at least one processor 801 can execute the above-mentioned passenger flow prediction method.
  • At least one processor 801 is capable of obtaining the connectivity relationship between each preset node and its adjacent nodes within a preset area, where the preset area includes multiple preset nodes and connecting channels therebetween, and the connectivity relationship is used to characterize the flow direction of the connecting channels; obtaining the population density information corresponding to multiple preset nodes; and predicting the population flow information within the preset area based on the population density information corresponding to multiple preset nodes and the connectivity relationship between each preset node and its adjacent nodes.
  • At least one processor 801 can implement any passenger flow prediction method in the embodiments of the present disclosure.
  • FIG. 9 shows a block diagram of a composition of an electronic device provided by an embodiment of the present disclosure.
  • an embodiment of the present disclosure provides an electronic device, which includes multiple processing cores 901 and an on-chip network 902 , wherein the multiple processing cores 901 are all connected to the on-chip network 902 , and the on-chip network 902 is used to exchange data between the multiple processing cores and external data.
  • One or more instructions are stored in one or more processing cores 901, and the one or more instructions are executed by one or more processing cores 901, so that one or more processing cores 901 can execute the above-mentioned passenger flow prediction method.
  • one or more processing cores 901 are capable of obtaining the connectivity relationship between each preset node and its adjacent nodes within a preset area, where the preset area includes multiple preset nodes and connection channels therebetween, and the connectivity relationship is used to characterize the flow direction of the connection channels; obtaining the personnel density information corresponding to multiple preset nodes; and predicting the personnel flow information within the preset area based on the personnel density information corresponding to multiple preset nodes and the connectivity relationship between each preset node and its adjacent nodes.
  • At least one processor 801 can implement any passenger flow prediction method in the embodiments of the present disclosure.
  • the present disclosure also provides a computer-readable storage medium on which a computer program is stored, wherein the computer program implements the above-mentioned passenger flow prediction method when executed by a processor/processing core.
  • the computer-readable storage medium may be a volatile or non-volatile computer-readable storage medium.
  • the embodiments of the present disclosure also provide a computer program product, including a computer-readable code, or a non-volatile computer-readable storage medium carrying the computer-readable code.
  • a computer program product including a computer-readable code, or a non-volatile computer-readable storage medium carrying the computer-readable code.
  • computer storage media includes volatile media implemented in any method or technology for storage of information such as computer-readable program instructions, data structures, program modules or other data. and non-volatile, removable and non-removable media.
  • Computer storage media include, but are not limited to, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), static random access memory (SRAM), flash memory or other memory technology, portable compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical disk storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other medium that can be used to store the desired information and can be accessed by a computer.
  • communication media typically contain computer-readable program instructions, data structures, program modules, or other data in a modulated data signal such as a carrier wave or other transmission mechanism, and may include any information delivery media.
  • the computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to each computing/processing device, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and/or a wireless network.
  • the network can include copper transmission cables, optical fiber transmissions, wireless transmissions, routers, firewalls, switches, gateway computers, and/or edge servers.
  • the network adapter card or network interface in each computing/processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in the computer-readable storage medium in each computing/processing device.
  • the computer program instructions for performing the operation of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-related instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages, such as Smalltalk, C++, etc., and conventional procedural programming languages, such as "C" language or similar programming languages.
  • Computer-readable program instructions may be executed entirely on a user's computer, partially on a user's computer, as an independent software package, partially on a user's computer, partially on a remote computer, or entirely on a remote computer or server.
  • the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., using an Internet service provider to connect via the Internet).
  • an electronic circuit such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), may be personalized by utilizing the state information of the computer-readable program instructions, and the electronic circuit may execute the computer-readable program instructions, thereby realizing various aspects of the present disclosure.
  • the computer program product described herein may be implemented in hardware, software, or a combination thereof.
  • the computer program product is embodied as a computer storage medium, and in another optional embodiment, the computer program product is embodied as a software product, such as a software development kit (SDK), etc.
  • SDK software development kit
  • These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer or other programmable data processing device to produce a machine such that when these instructions are executed by the processor of the computer or other programmable data processing device, the functions specified in one or more blocks in the flowchart and/or block diagram are realized.
  • These computer-readable program instructions may also be stored in a computer-readable storage medium, which enables a computer, a programmable data processing device and/or other device to work in a specific manner, so that a computer-readable medium storing instructions includes a product, which includes instructions for implementing various aspects of the functions/actions specified in one or more blocks in the flowchart and/or block diagram.
  • Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device so that a series of operating steps are performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to implement the functions/actions specified in one or more boxes in the flowchart and/or block diagram.
  • each square box in the flow chart or block diagram can represent a part of a module, program segment or instruction, and a part of a module, program segment or instruction includes one or more executable instructions for realizing the specified logical function.
  • the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two continuous square boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved.
  • each square box in the block diagram and/or flow chart, and the combination of the square boxes in the block diagram and/or flow chart can be implemented with a dedicated hardware-based system that performs the specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Strategic Management (AREA)
  • Theoretical Computer Science (AREA)
  • Human Resources & Organizations (AREA)
  • Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Development Economics (AREA)
  • Multimedia (AREA)
  • Entrepreneurship & Innovation (AREA)
  • Marketing (AREA)
  • Operations Research (AREA)
  • Quality & Reliability (AREA)
  • Tourism & Hospitality (AREA)
  • General Business, Economics & Management (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

本公开涉及一种客流预测方法、装置和系统，涉及客流预测技术领域。客流预测方法包括：获取预设区域内的各个预设节点与其相邻节点之间的连通关系，预设区域包括多个预设节点及其之间的连接通道，连通关系用于表征连接通道的流通方向；获取多个预设节点对应的人员密度信息；依据多个预设节点对应的人员密度信息、各个预设节点与其相邻节点之间的连通关系，预测预设区域内的人员流动信息。

Description

客流预测方法、装置和系统 技术领域
本公开涉及客流预测技术领域，具体涉及一种客流预测方法、装置和系统。
背景技术
在车站候车厅或地铁站台内，通常会安装有摄像头，以便于使用该摄像头对不同区域内的人员流动情况进行监控。
发明内容
本公开实施例提供一种客流预测方法及装置、系统、电子设备、计算机可读存储介质，其可以准确预测某预设区域内的人员流动信息，以保证预设区域的安全性。
第一方面，本公开实施例提供一种客流预测方法，方法包括：获取预设区域内的各个预设节点与其相邻节点之间的连通关系，预设区域包括多个预设节点及其之间的连接通道，连通关系用于表征连接通道的流通方向；获取多个预设节点对应的人员密度信息；依据多个预设节点对应的人员密度信息、各个预设节点与其相邻节点之间的连通关系，预测预设区域内的人员流动信息。
在一些实施例中，获取预设区域内的各个预设节点与其相邻节点之间的连通关系之后，获取多个预设节点对应的人员密度信息之前，方法还包括：
依据多个预设节点、以及至少一个预设节点与其相邻节点之间的连通关系，构建邻接拓扑模型，邻接拓扑模型采用N阶方阵的形式表示，N为大于1的整数。
在一些实施例中，获取多个预设节点对应的人员密度信息，包括：
获取多帧待处理图像，其中，待处理图像为采用设置于每个预设节点中的至少一个图像采集设备对每个预设节点内的人员的图像进行采集所获得的图像；
依据深度学习算法对多帧待处理图像进行分析，确定至少一个预设节点对应的人员密度信息。
在一些实施例中，获取多个预设节点对应的人员密度信息之前，方法还包括：
确定多个连接通道内是否存在监控摄像头；
在确定连接通道内存在监控摄像头的情况下，依据客流统计算法，分别对多个连接通道内的人员流动信息进行统计，获得多个连接通道对应的客流统计信息。
在一些实施例中，客流统计信息包括：经过连接通道的人员数量；
依据多个预设节点对应的人员密度信息、各个预设节点与其相邻节点之间的连通关系，预测预设区域内的人员流动信息，包括：
依据各个预设节点与其相邻节点之间的连通关系，确定各个连接通道的流通方向；
依据各个连接通道的流通方向，确定与各个连接通道对应的权重系数矩阵；
依据权重系数矩阵、人员密度信息和经过连接通道的人员数量，确定线性估计模型；
依据线性估计模型，对预设区域内的人员流动信息进行预测，获得区域流量预测结果。
在一些实施例中，客流统计信息，方法还包括：通道统计误差信息；
依据线性估计模型，对预设区域内的人员流动信息进行预测，获得区域流量预测结果之前，方法还包括：
依据通道统计误差信息和预设节点的预设客流总量信息对线性估计模型进行调整，获得更新后的线性估计模型。
在一些实施例中，依据多个预设节点对应的人员密度信息、各个预设节点与其相邻节点之间的连通关系，预测预设区域内的人员流动信息之后，方法还包括：
每间隔预设时长，更新人员密度信息和/或客流统计信息。
在一些实施例中，每个预设节点包括：入口和出口；
获取多个预设节点对应的人员密度信息，包括：
获取经过每个预设节点的入口的第一通信交互信息，第一通信交互信息为经过入口的各个人员携带的设备与设置于入口处的预设设备之间的信息，
获取经过每个预设节点的出口的第二通信交互信息，第二通信交互信息为经过出口的各个人员携带的设备与设置于出口处的预设设备之间的信息；
依据第一通信交互信息和第二通信交互信息，确定预设节点对应的人员密度信息；其中，预设节点包括：进站节点、安全检查节点、检票节点、候车节点和乘车站台节点中的至少一种。
第二方面，本公开实施例提供一种客流预测装置，其包括：第一获取模块，被配置为获取预设区域内的各个预设节点与其相邻节点之间的连通关系，预设区域包括多个预设节点及其之间的连接通道，连通关系用于表征连接通道的流通方向；第二获取模块，被配置为获取多个预设节点对应的人员密度信息；预测模块，被配置为依据多个预设节点对应的人员密度信息和多个连接通道内的客流统计信息，预测预设区域内的人员流动信息。
第三方面，本公开实施例提供一种客流预测系统，其包括：设置于多个预设节点内的第一图像采集设备、设置于预设节点之间的连接通道内的第二图像采集设备和客流预测装置；第一图像采集设备，被配置为对与第一图像采集设备对应的预设节点内的人员的图像进行采集，获得多帧待处理图像，并向客流预测装置发送多帧待处理图像，以供客流预测装置依据深度学习算法对多帧待处理图像进行分析，确定至少一个预设节点对应的人员密度信息；第二图像采集设备，被配置为对预设节点之间的连接通道内的人员运动图像进行采集，获得通道图像，并对通道图像进行分析，获得与连接通道对应的客流统计信息，并向客流预测装置发送客流统计信息，以供客流预测装置依据客流统计算法，分别对多个连接通道内的人员流动信息进行统计，获得连接通道对应的客流统计信息；客流预测装置，被配置为执行本公开中任意一种客流预测方法。
第四方面，本公开实施例提供一种电子设备，包括存储器、处理器及存储在存储器 上并可在处理器上运行的计算机程序，处理器执行计算机程序时实现本公开实施例任意一种客流预测方法。
第五方面，本公开实施例提供一种计算机可读存储介质，其上存储有计算机程序，该计算机程序被处理器执行时实现本公开实施例任意一种客流预测方法。
在本公开实施例中，通过获取预设区域内的各个预设节点与其相邻节点之间的连通关系，能明确预设区域内的多个预设节点之间是否存在连接通道，以及该连接通道的流通方向，便于后续处理；获取预设区域内的多个预设节点对应的人员密度信息，能够明确不同的预设节点内的人员密度情况，便于对不同的预设节点内的人员数量进行控制，其中的预设区域包括多个预设节点及其之间的连接通道；依据多个预设节点对应的人员密度信息、各个预设节点与其相邻节点之间的连通关系，预测预设区域内的人员流动信息，以实现对预设区域内的人员的流动信息进行准确预测，从而实现对不同的预设节点之间的人员流向进行调度和管理，提升预设区域内的安全性。
应当理解的是，以上的一般描述和后文的细节描述仅是示例性和解释性的，而非限制本公开。根据下面参考附图对示例性实施例的详细说明，本公开的其它特征及方面将变得清楚。
附图说明
图1示出本公开实施例提供的一种客流预测方法的流程示意图。
图2示出本公开实施例提供的一种客流预测方法的流程示意图。
图3示出本公开实施例提供的对预设节点内的人员密度的统计示意图。
图4示出本公开实施例提供的一种客流预测装置的组成方框图。
图5示出本公开实施例提供的一种客流预测系统的组成方框图。
图6示出本公开实施例提供的一种客流预测系统的工作方法的流程示意图。
图7示出本公开实施例提供的一种邻接拓扑模型的示意图。
图8示出本公开实施例提供的一种电子设备的组成方框图。
图9示出本公开实施例提供的一种电子设备的组成方框图。
具体实施方式
下面结合附图和实施例对本公开作进一步的详细说明。可以理解的是，此处所描述的具体实施例仅仅用于解释本公开，而非对本公开的限定。另外还需要说明的是，为了便于描述，附图中仅示出了与本公开相关的部分而非全部结构。
在车站候车厅或地铁站台内，通常会安装有摄像头，以便于使用该摄像头对不同区域内的人员流动情况进行监控。但是，摄像头的部署位置以及距离人员的远近都会影响监控效果，且会存在监控盲区的问题，无法实现对区域内的人员的流向和流量进行准确预测，降低了站台候车厅、地铁站台内等人员密集场所的安全性。
第一方面，本公开实施例提供一种客流预测方法。本公开实施例的客流预测方法可 由相应的客流预测装置执行，该客流预测装置可采用软件和/或硬件的方式实现，并一般可集成于电子设备中。
图1示出本公开实施例提供的一种客流预测方法的流程示意图。如图1所示，该客流预测方法包括但不限于如下步骤：
步骤S101，获取预设区域内的各个预设节点与其相邻节点之间的连通关系。
其中，预设区域包括多个预设节点及其之间的连接通道。
需要说明的是，一个预设节点至少有一个相邻节点，以确定多个预设节点之间是否是连通的，从而确定多个预设节点之间的连通关系，该连通关系用于表征连接通道的流通方向。
步骤S102，获取多个预设节点对应的人员密度信息。
人员密度信息可以包括：预设节点内实时存在的人员的数量、预设节点可容纳的人员总数、以及预设节点内实时存在的人员的数量与该预设节点可容纳的人员总数的比值等信息。
通过人员密度信息能够表征该预设节点内的实时人员情况，以便于对预设节点的人员进行控制，保证人员的安全性。
步骤S103，依据多个预设节点对应的人员密度信息、各个预设节点与其相邻节点之间的连通关系，预测预设区域内的人员流动信息。
其中，人员密度信息包括每个预设节点中包括的人员数量和人员分布区域信息等，能够明确在预设节点内具体有多少人员。
而不同的预设节点之间的连通关系不同，可以通过多个预设节点对应的连通关系能够表征各个预设节点之间的连接通道的流通方向，从而明确人员可能从哪个预设节点流动到哪个预设节点。
进一步地，还可以基于不同预设节点在不同时段内的人员数量，对预设区域内的各个预设节点之间的人员的流动方向，以及人员的流动数量进行预测，并对流动的人员进行管理和调度，提升预设区域内的安全性。
在本实施例中，通过获取预设区域内的各个预设节点与其相邻节点之间的连通关系，能明确预设区域内的多个预设节点之间是否存在连接通道，以及该连接通道的流通方向，便于后续处理；获取预设区域内的多个预设节点对应的人员密度信息，能够明确不同的预设节点内的人员密度情况，便于对不同的预设节点内的人员数量进行控制，其中的预设区域包括多个预设节点及其之间的连接通道；依据多个预设节点对应的人员密度信息、各个预设节点与其相邻节点之间的连通关系，预测预设区域内的人员流动信息，以实现对预设区域内的人员的流动信息进行准确预测，从而实现对不同的预设节点之间的人员流向进行调度和管理，提升预设区域内的安全性。
图2示出本公开实施例提供的一种客流预测方法的流程示意图。该客流预测方法可应用于客流预测装置。如图2所示，该客流预测方法包括但不限于如下步骤：
步骤S201，获取每个预设节点与其相邻节点之间的连通关系。
需要说明的是，本实施例中的步骤S201与上一实施例中的步骤S101相同，在此不再赘述。
步骤S202，依据多个预设节点、以及至少一个预设节点与其相邻节点之间的连通关系，构建邻接拓扑模型。
其中，邻接拓扑模型是以多个预设节点为节点，以相邻两个预设节点之间的连接通道为边的模型。
例如，邻接拓扑模型包括5个预设节点(第一预设节点、第二预设节点、第三预设节点、第四预设节点、以及第五预设节点)，其中，第一预设节点分别与第二预设节点和第三预设节点之间是连通的，而第四预设节点和第五预设节点之间是不连通的，则构建的邻接拓扑模型可以包括第一预设节点和第二预设节点之间的连通值(例如，设定为1)、第一预设节点和第三预设节点之间的连通值(例如，设定为1)、第四预设节点和第五预设节点之间的连通值(例如，设定为0)，则可根据确定的邻接拓扑模型，直观的获知各个预设节点之间的连通关系。
例如，邻接拓扑模型采用N阶方阵的形式表示，N为大于1的整数。如，N阶方阵为N行N列的矩阵。当N等于3时，该3阶方阵中的第1行第1列的元素的值表示第1个预设节点与其自身是连通的(即，该值为1)；第2行第1列的元素的值表示第2个预设节点与第1个预设节点之间是否是连通的(即，若是连通的，则该值为1；否则，该值为0)；……、以此类推，可获知该3阶方阵中的各个元素的值，从而直观的获得邻接拓扑模型中各个预设节点之间的连通性。
例如，预设节点包括：进站节点、安全检查节点、检票节点、候车节点和乘车站台节点中的至少一种。
通过不同类型的预设节点，能够多维度的对不同的预设节点进行处理，使客流预测方法适用于更多的使用场景中，扩展客流预测方法的应用范围。
步骤S203，获取多个预设节点对应的人员密度信息。
步骤S204，依据多个预设节点对应的人员密度信息、各个预设节点与其相邻节点之间的连通关系，预测预设区域内的人员流动信息。
需要说明的是，本实施例中的步骤S203～步骤S204与上一实施例中的步骤S102～步骤S103相同，在此不再赘述。
在本实施例中，通过依据多个预设节点、以及每个预设节点与其相邻节点之间的连通关系，构建邻接拓扑模型，能够直观的明确多个预设节点之间的连通关系，便于后续处理；并通过获取预设区域内的多个预设节点对应的人员密度信息，能够明确不同的预设节点内的人员密度情况，便于对不同的预设节点内的人员数量进行控制；依据多个预设节点对应的人员密度信息和多个连接通道内的客流统计信息，预测预设区域内的人员流动信息，以实现对预设区域内的人员的流动信息进行准确预测，从而实现对不同的预设节点之间的人员流向进行调度和管理，提升预设区域内的安全性。
在一些具体实现中，步骤S102或步骤S203中的获取多个预设节点对应的人员密度 信息，可以采用如下方式实现：获取多帧待处理图像；依据深度学习算法对多帧待处理图像进行分析，确定至少一个预设节点对应的人员密度信息。
其中，待处理图像为采用设置于每个预设节点中的至少一个图像采集设备对每个预设节点内的人员的图像进行采集所获得的图像。
在一些实施例中，每个预设节点中可以设置至少一个图像采集设备，以便于对该预设节点内的人员的图像进行准确的采集。
例如，预设节点为售票节点，图3示出本公开实施例提供的对预设节点内的人员密度的统计示意图。如图3所示，图3左侧的图像用于表征售票节点中安装的摄像头所采集到的图像，其中，可获知售票节点中存在多个人员，其中，不同的人员对应的分布区域不同。通过采用人员密度统计算法，对图3左侧的图像进行分析，可获得图3右侧所示的人员分布信息(如，基于人头对应的位置信息，对售票节点内存在的人员进行数量统计，获得人员密度信息等)。
例如，可对图3右侧所显示的人头的数量进行统计，从而获得在售票节点中存在的人员的数量。
通过基于深度学习算法所获得的人员密度统计算法，对待处理图像进行分析处理，可获得在预设节点内的人员的数量，以及人员密度信息，智能化的对预设节点内的人员密度进行统计，以实现对预设节点内的人员进行实时监控和管理，提升预设节点内的人员安全性。
又例如，预设节点包括售票节点和检票节点，则分别采用设置于售票节点中的至少一个图像采集设备对该售票节点内的人员的图像进行采集，以及采用设置于检票节点中的至少一个图像采集设备对该检票节点内的人员的图像进行采集，可获得售票节点对应的多帧待处理图像、以及检票节点对应的多帧待处理图像。
进一步地，依据深度学习算法对售票节点对应的多帧待处理图像和检票节点对应的多帧待处理图像进行分析，确定售票节点对应的人员密度信息和/或检票节点对应的人员密度信息，便于对这两个节点中的人员进行综合的调度。
在一些具体实现中，每个预设节点包括：入口和出口；步骤S102或步骤S203中的获取多个预设节点对应的人员密度信息，可以采用如下方式实现：
获取经过每个预设节点的入口的第一通信交互信息；获取经过每个预设节点的出口的第二通信交互信息；依据第一通信交互信息和第二通信交互信息，确定预设节点对应的人员密度信息。
其中，第一通信交互信息为经过入口的各个人员携带的设备与设置于入口处的预设设备之间的信息，第二通信交互信息为经过出口的各个人员携带的设备与设置于出口处的预设设备之间的信息。
其中，预设设备可以为设置于入口或出口处的检测设备。例如，预设设备为可进行刷卡的闸机，人员在通过入口或出口处时，通过其携带的待验证卡片(如，车票、表征人员身份信息的卡片等)与闸机进行交互，可获得第一通信交互信息和第二通信交互信 息。
进一步地，通过入口和出口的闸机将上述第一通信交互信息和第二通信交互信息汇总到服务器，以使服务器可以对经过入口和出口的人员进行数量的统计，以及身份的验证等，从而可获知进出预设节点的人员数量，进而在明确预设节点内的实时人员数量的情况下，可确定该预设节点对应的人员密度信息。
通过上述操作，能够使每个预设节点内的人员密度信息都可被实时监控，以便于客流预测装置在出现异常情况(如，预设节点中人员数量超过预设阈值等)时，能够及时对预设节点内的人员进行管控，提升人员的安全性。
在一些具体实现中，在执行步骤S102或步骤S203中的获取多个预设节点对应的人员密度信息之前，方法还包括：
确定多个连接通道内是否存在监控摄像头；在确定连接通道内存在监控摄像头的情况下，依据客流统计算法，分别对多个连接通道内的人员流动信息进行统计，获得多个连接通道对应的客流统计信息。
其中，客流统计算法可以是基于人员的流入和流出情况确定的统计算法，以便于对多个连接通道内的人员流动信息进行统计，使获得的多个连接通道对应的客流统计信息能够体现多个连接通道内的人员流动情况(如，某个连接通道内的人员的运动方向以及人员数量等信息)。
例如，客流统计信息包括：经过连接通道的人员数量和连接通道的流通方向。
步骤S103或步骤S204中的依据多个预设节点对应的人员密度信息、各个预设节点与其相邻节点之间的连通关系，预测预设区域内的人员流动信息，可以采用如下方式实现：
依据各个预设节点与其相邻节点之间的连通关系，确定各个连接通道的流通方向；依据各个连接通道的流通方向，确定与各个连接通道对应的权重系数矩阵；依据权重系数矩阵、人员密度信息和经过连接通道的人员数量，确定线性估计模型；依据线性估计模型，对预设区域内的人员流动信息进行预测，获得区域流量预测结果。
其中，权重系数矩阵可以包括多个权重值，该权重值用于表征连接通道内的向预设方向运动的比重值(如，第1个预设节点和第2个预设节点之间的连接通道内，该比重值可表征从第1个预设节点运动到第2个预设节点的人员的比例，或，表征从第2个预设节点运动到第1个预设节点的人员的比例等)。
需要说明的是，通过各个预设节点与其相邻节点之间的连通关系，可以明确某个预设节点具体与哪些相邻节点存在连通关系，从而确定该预设节点中的人员可能流动到哪个相邻节点中，进而确定各个预设节点之间的连通通道的流通方向。
通过将权重系数矩阵、人员密度信息和经过连接通道的人员数量进行综合分析，建立线下估计模型，以明确不同预设节点以及其对应的连接通道内的人员流向情况，进而通过该线性估计模型对预设区域内的人员流动信息进行预测，以便于预知预设区域内的多个预设节点之间的人员流向情况(即，区域流量预测结果)。
通过该区域流量预测结果能够获知预设区域在未来一段时间内可能存在的人员数量以及人员流动方向等信息，便于对该预设区域内的人员进行及时管控。如，当确定预设区域内人员过多时，可通过增加检测关口，以加快人员的流动速度，保证预设区域内的人员安全性；或，当确定预设区域内人员过少时，可放开入口处的人员进入数量，以避免设备或管理人员的闲置，提升预设区域内的设备利用效率。
在一些具体实现中，客流统计信息，还包括：通道统计误差信息。依据线性估计模型，对预设区域内的人员流动信息进行预测，获得区域流量预测结果之前，方法还包括：依据通道统计误差信息和预设节点的预设客流总量信息对线性估计模型进行调整，获得更新后的线性估计模型。
其中，通道统计误差信息用于表征对连接通道内的人员进行统计时，可能存在的差错信息，如，该通道统计误差信息可以包括连接通道的标识、该连接通道对应的统计误差值(如，该统计误差值可以为2、3等值)。
通过通道统计误差信息对和预设节点的预设客流总量信息对线性估计模型进行调整，以使线性估计模型可以更贴近实际的区域内的人员统计情况，进而使用更新后的线性估计模型对预设区域中的人员进行预测时，能够提升预测结果的准确性。
在一些具体实现中，在执行步骤S103或步骤S204中的依据多个预设节点对应的人员密度信息、各个预设节点与其相邻节点之间的连通关系，预测预设区域内的人员流动信息之后，方法还包括：每间隔预设时长，更新人员密度信息和客流统计信息。
其中，预设时长可以包括：1个小时、2个小时等时长，以上对于预设时长仅是举例说明，可根据实际需要进行具体设定，其他未说明的预设时长也在本公开的保护范围之内，在此不再赘述。
通过每间隔预设时长，对人员密度信息和客流统计信息进行更新，能够及时获知预设节点内的人员密度，以及该预设节点与其相邻节点之间的连接通道中的客流统计信息，实现对不同区域内的人员情况进行及时的监控和管理。
应当理解，以上实施例还可与本公开实施例的其它任意方式结合使用。以上实施例只是本公开的一个具体例子，而不是对本公开保护范围的限定。
第二方面，本公开实施例提供一种客流预测装置。其是实现本公开上述实施例提供的客流预测方法的相应装置，该装置可采用软件和/或硬件的方式实现，并一般可集成于电子设备中。
图4示出本公开实施例提供的一种客流预测装置的组成方框图。如图4所示，该客流预测装置400包括但不限于如下模块。
第一获取模块410，被配置为获取预设区域内的各个预设节点与其相邻节点之间的连通关系，预设区域包括多个预设节点及其之间的连接通道，连通关系用于表征连接通道的流通方向。
第二获取模块420，被配置为获取多个预设节点对应的人员密度信息。
预测模块430，被配置为依据多个预设节点对应的人员密度信息和多个连接通道内 的客流统计信息，预测预设区域内的人员流动信息。
本实施方式中的装置可以执行本公开实施例中任一种客流预测方法，其具体实施不局限于以上实施例，其他未说明的实施例也在本装置的保护范围之内。
在本实施方式中，通过第一获取模块获取预设区域内的各个预设节点与其相邻节点之间的连通关系，能明确预设区域内的多个预设节点之间是否存在连接通道，以及该连接通道的流通方向，便于后续处理；使用第二获取模块获取预设区域内的多个预设节点对应的人员密度信息，能够明确不同的预设节点内的人员密度情况，便于对不同的预设节点内的人员数量进行控制，其中的预设区域包括多个预设节点及其之间的连接通道；使用预测模块依据多个预设节点对应的人员密度信息、各个预设节点与其相邻节点之间的连通关系，预测预设区域内的人员流动信息，以实现对预设区域内的人员的流动信息进行准确预测，从而实现对不同的预设节点之间的人员流向进行调度和管理，提升预设区域内的安全性。
值得一提的是，本实施方式中所涉及到的各模块均为逻辑模块，在实际应用中，一个逻辑单元可以是一个物理单元，也可以是一个物理单元的一部分，还可以以多个物理单元的组合实现。此外，为了突出本公开的创新部分，本实施方式中并没有将与解决本公开所提出的技术问题关系不太密切的单元引入，但这并不表明本实施方式中不存在其它的单元。
图5示出本公开实施例提供的一种客流预测系统的组成方框图。如图5所示，客流预测系统包括但不限于如下设备。
设置于多个预设节点内的第一图像采集设备510、设置于预设节点之间的连接通道内的第二图像采集设备520和客流预测装置530；
第一图像采集设备510，被配置为对与第一图像采集设备510对应的预设节点内的人员的图像进行采集，获得多帧待处理图像，并向客流预测装置530发送多帧待处理图像，以供客流预测装置530依据深度学习算法对多帧待处理图像进行分析，确定至少一个预设节点对应的人员密度信息。
第二图像采集设备520，被配置为对预设节点之间的连接通道内的人员运动图像进行采集，获得通道图像，并对通道图像进行分析，获得与连接通道对应的客流统计信息，并向客流预测装置530发送客流统计信息，以供客流预测装置530依据客流统计算法，分别对多个连接通道内的人员流动信息进行统计，获得连接通道对应的客流统计信息。
客流预测装置530，被配置为执行本公开实施例中任一种客流预测方法。
图6示出本公开实施例提供的一种客流预测系统的工作方法的流程示意图。如图6所示，客流预测系统的工作方法包括但不限于如下步骤。
步骤S601，根据预设区域内的多个预设节点以及每个预设节点与其相邻节点之间的连通关系，构建邻接拓扑模型。
其中，邻接拓扑模型可以采用N阶方阵的形式表示，N为大于1的整数。
例如，N阶方阵M为N行N列的矩阵，在N阶方阵M中的每个元素可以表示为M(i，j)， 其中，M(i，j)的值表示第i个预设节点与第j个预设节点之间是否连通。i和j均为大于或等于1，且小于或等于N的整数。
图7示出本公开实施例提供的一种邻接拓扑模型的示意图。如图7所示，预设区域(例如，车站)内包括如下预设节点：进站口1、安检处2、检票口3、第一候车厅4、第二候车厅5和站台6。图7中展示了各个预设节点之间的连接关系，为了方便观测，构建6*6的矩阵，以该矩阵为邻接拓扑模型。
例如，构建的邻接拓扑模型采用公式(1)表示如下：
其中，M(1，1)＝1，即表示从进站口1到进站口1是连通的；M(1，2)＝1，即表示从进站口1到安检处2之间是连通的(即从进站口1到安检处2之间存在走廊或过道，以便于人员从进站口1行进到安检处2处)；……；依次类推，M(6，6)＝1，即表示站台6到站台6之间是连通的。
步骤S602，获取预设区域内的多个预设节点对应的人员密度信息。
例如，对每个预设节点都进行如下操作：从预设节点(如，检票口3、第一候车厅4等)中安装的图像采集设备中，获取该预设节点内的人员的图像(即，多帧待处理图像)，依据深度学习算法对多帧待处理图像进行分析，确定每个预设节点对应的人员密度信息。
步骤S603，获取相邻两个预设节点之间存在连通关系的走廊或过道内的客流统计信息。
例如，通过对安检处2到检票口3之间的过道内的人员数量和人员流动方向进行统计，从而获得安检处2与检票口3之间的客流统计信息。如，可通过在安检处2到检票口3之间的走廊或过道内安装的摄像头，并采用该摄像头进行图像的采集，从而对采集到的图像进行分析，可获得预设时长内经过该走廊的客流统计信息(即，人员数量和人员流动方向等信息)。
需要说明的是，其他相邻预设节点之间的客流统计信息与上述方法相同，可将从第i个预设节点与第j个预设节点之间的连接通道的人员数量记作aij,方便后续使用。
步骤S604，依据各个连接通道的流通方向，确定经过各个连接通道的权重系数矩阵。
其中，基于各个连接通道的流通方向，能够明确不同的预设节点之间的定向的流动方向，例如，进站口1中的人员只能流向安检处2，从而能够明确进站口1与安检处2之间的连接通道内的流动人员的数量占进站口1内的人员总数的比例值，从而将该比例值作为权重系数；基于不同的连接通道，对应的权重系数不同。
例如，假设在时间t1内，第i个预设节点的总人数是pi，第j个预设节点的总人数是pj；在时间t2内，第i个预设节点的总人数是p′i，第j个预设节点的总人数是p′j；设置权重系数矩阵为W，其中，该权重系数矩阵W包括多个元素(如，元素表示为wij，即从第i个预设节点流向第j个预设节点的人数比例，例如，)，并且，0≤wij≤1。
需要说明的是，每间隔预设时长，都可以根据第i个预设节点中安装的摄像头所采集到的图像对应的pi的值，进行更新，以便于获得第i个预设节点流向其相邻的预设节点的人数。
例如，由于在火车站，客车站或地铁站等场合，在每天的不同时段内的人员流向和人员流动数量都不同，可每间隔预设时长(如，1个小时、2个小时等)，都重新更新一次wij的值，以使预测结果更准确。
步骤S605，依据权重系数矩阵、人员密度信息和经过连接通道的人员数量，确定线性估计模型。
其中，线性估计模型用于表征不同预设节点的人员数量的变化值。例如，可采用公式(2)表示该线性估计模型。
其中，mi表示第i个预设节点的人员数量的变化值；表示流入第i个预设节点的人员数量；表示流出第i个节点的人员数量。其中，K表示预设节点中的流出节点的数量，L表示预设节点中的流入节点的数量，K、L均为大于或等于0的整数；k为大于或等于0，且小于或等于K的整数；l为大于或等于0，且小于或等于L的整数。i为大于或等于1，且，小于或等于预设区域内的预设节点的数量的整数。
wki表示在时间t1内，从第k个预设节点流入第i个预设节点的人数比例；aki表示在时间t1内，从第k个预设节点与第i个预设节点之间的连接通道的人员数量(即，在该连接通道内实时存在的人员的数量)；pk表示在时间t1内，第k个预设节点的总人数。
wil表示从第i个预设节点流入第l个预设节点的人数比例；ail表示从第i个预设节点与第l个预设节点之间的连接通道的人员数量(即，在该连接通道内实时存在的人员的数量)；pi表示在时间t1内，第i个预设节点的总人数。
步骤S606，依据线性估计模型，对预设区域内的人员流动信息进行预测，获得区域流量预测结果。
例如，通过线性估计模型使预设节点的人员数量实际变化值与估计值之差尽量小的方式，获得区域流量预测结果。
如，可采用如下公式(3)表示目标函数(即区域流量预测结果)：
其中，min‖‖1表示对于一个预设节点i而言，以1作为人员流出的总量，确定该预设节点i对应的变化值中的最小值，从而获得包括多个预设节点的预设区域的流量预测结果。其中，I表示预设区域内包括的预设节点的数量，I为大于或等于1的整数，i为大 于或等于1，且，小于或等于I的整数。
在采用上述公式(3)来确定区域流量预测结果时，还需要使用如下约束条件：即表示对于一个预设节点i而言，其人员流出的总量为1。其中，M表示预设节点i内的人员变化次数的最大值，M为大于或等于1的整数，m为大于或等于0，且，小于或等于M的整数。
在一些具体实现中，还可以基于连接通道(如，从第i个预设节点到第j个预设节点之间的连接通道等)内的通道统计误差信息，对上述预测过程进行约束，如采用公式(4)表示该通道统计误差信息。
-δ≤wijaijpi-eij≤δ     (4)
其中，eij表征从第i个预设节点流向第j个预设节点的人员的数量(即，已经流入第j个预设节点的人员的数量)；δ表示统计误差值，例如，δ的取值可以为2、3等。
进一步地，依据通道统计误差信息和预设节点的预设客流总量信息对线性估计模型进行调整，获得更新后的线性估计模型。例如，更新后的线性估计模型可采用公式(5)表示：
其中，eij表征从第i个预设节点流向第j个预设节点的人员的数量(即，已经流入第j个预设节点的人员的数量)。
需要说明的是，在公式(5)中，除了wij之外的其他变量均是已知数，可以通过对公式(5)进行线性求解的方式，获得最优的wij，即从第i个预设节点流向第j个预设节点的人员数量的比例。
其中，还可以进一步获得从第i个预设节点流向第j个预设节点的人员数量：即wijaijpi。
在本实施例中，通过在预设区域内的多个预设节点及其与相邻节点之间的连通关系，构建邻接拓扑模型，以便于基于该邻接拓扑模型对预设区域内的人员流动情况进行预测，便于对预设区域进行人员的管理；并且，通过深度学习算法对不同的预设节点内的摄像头所采集的多帧待处理图像进行分析，确定每个预设节点对应的人员密度信息，可智能化对预设节点内的人员密度进行统计，以实现对预设节点内的人员进行实时监测和监控；依据多个预设节点对应的人员密度信息和多个连接通道内的客流统计信息，预测预设区域内的人员流动信息，以实现对预设区域内的人员的流动信息进行准确预测，从而实现对不同的预设节点之间的人员流向进行调度和管理，提升预设区域内的安全性。
图8示出本公开实施例提供的一种电子设备的组成方框图。
参照图8，本公开实施例提供了一种电子设备，该电子设备包括：至少一个处理器801；至少一个存储器802，以及一个或多个I/O接口803，连接在处理器801与存储器802之间；其中，存储器802存储有可被至少一个处理器801执行的一个或多个计算机 程序，一个或多个计算机程序被至少一个处理器801执行，以使至少一个处理器801能够执行上述的客流预测方法。
在一些实施例中，至少一个处理器801能够获取预设区域内的各个预设节点与其相邻节点之间的连通关系，预设区域包括多个预设节点及其之间的连接通道，连通关系用于表征连接通道的流通方向；获取多个预设节点对应的人员密度信息；依据多个预设节点对应的人员密度信息、各个预设节点与其相邻节点之间的连通关系，预测预设区域内的人员流动信息。
需要说明的是，至少一个处理器801能够实现本公开实施例中任一种客流预测方法。
图9示出本公开实施例提供的一种电子设备的组成方框图。
参照图9，本公开实施例提供了一种电子设备，该电子设备包括多个处理核901以及片上网络902，其中，多个处理核901均与片上网络902连接，片上网络902用于交互多个处理核间的数据和外部数据。
其中，一个或多个处理核901中存储有一个或多个指令，一个或多个指令被一个或多个处理核901执行，以使一个或多个处理核901能够执行上述的客流预测方法。
在一些实施例中，一个或多个处理核901能够获取预设区域内的各个预设节点与其相邻节点之间的连通关系，预设区域包括多个预设节点及其之间的连接通道，连通关系用于表征连接通道的流通方向；获取多个预设节点对应的人员密度信息；依据多个预设节点对应的人员密度信息、各个预设节点与其相邻节点之间的连通关系，预测预设区域内的人员流动信息。
需要说明的是，至少一个处理器801能够实现本公开实施例中任一种客流预测方法。
本公开实施例还提供了一种计算机可读存储介质，其上存储有计算机程序，其中，计算机程序在被处理器/处理核执行时实现上述的客流预测方法。计算机可读存储介质可以是易失性或非易失性计算机可读存储介质。
本公开实施例还提供了一种计算机程序产品，包括计算机可读代码，或者承载有计算机可读代码的非易失性计算机可读存储介质，当计算机可读代码在电子设备的处理器中运行时，电子设备中的处理器执行上述客流预测方法。
本领域普通技术人员可以理解，上文中所公开方法中的全部或某些步骤、系统、装置中的功能模块/单元可以被实施为软件、固件、硬件及其适当的组合。在硬件实施方式中，在以上描述中提及的功能模块/单元之间的划分不一定对应于物理组件的划分；例如，一个物理组件可以具有多个功能，或者一个功能或步骤可以由若干物理组件合作执行。某些物理组件或所有物理组件可以被实施为由处理器，如中央处理器、数字信号处理器或微处理器执行的软件，或者被实施为硬件，或者被实施为集成电路，如专用集成电路。这样的软件可以分布在计算机可读存储介质上，计算机可读存储介质可以包括计算机存储介质(或非暂时性介质)和通信介质(或暂时性介质)。
如本领域普通技术人员公知的，术语计算机存储介质包括在用于存储信息(诸如计算机可读程序指令、数据结构、程序模块或其他数据)的任何方法或技术中实施的易失 性和非易失性、可移除和不可移除介质。计算机存储介质包括但不限于随机存取存储器(RAM)、只读存储器(ROM)、可擦式可编程只读存储器(EPROM)、静态随机存取存储器(SRAM)、闪存或其他存储器技术、便携式压缩盘只读存储器(CD-ROM)、数字多功能盘(DVD)或其他光盘存储、磁盒、磁带、磁盘存储或其他磁存储装置、或者可以用于存储期望的信息并且可以被计算机访问的任何其他的介质。此外，本领域普通技术人员公知的是，通信介质通常包含计算机可读程序指令、数据结构、程序模块或者诸如载波或其他传输机制之类的调制数据信号中的其他数据，并且可包括任何信息递送介质。
这里所描述的计算机可读程序指令可以从计算机可读存储介质下载到各个计算/处理设备，或者通过网络、例如因特网、局域网、广域网和/或无线网下载到外部计算机或外部存储设备。网络可以包括铜传输电缆、光纤传输、无线传输、路由器、防火墙、交换机、网关计算机和/或边缘服务器。每个计算/处理设备中的网络适配卡或者网络接口从网络接收计算机可读程序指令，并转发该计算机可读程序指令，以供存储在各个计算/处理设备中的计算机可读存储介质中。
用于执行本公开操作的计算机程序指令可以是汇编指令、指令集架构(ISA)指令、机器指令、机器相关指令、微代码、固件指令、状态设置数据、或者以一种或多种编程语言的任意组合编写的源代码或目标代码，编程语言包括面向对象的编程语言—诸如Smalltalk、C++等，以及常规的过程式编程语言—诸如“C”语言或类似的编程语言。计算机可读程序指令可以完全地在用户计算机上执行、部分地在用户计算机上执行、作为一个独立的软件包执行、部分在用户计算机上部分在远程计算机上执行、或者完全在远程计算机或服务器上执行。在涉及远程计算机的情形中，远程计算机可以通过任意种类的网络—包括局域网(LAN)或广域网(WAN)—连接到用户计算机，或者，可以连接到外部计算机(例如利用因特网服务提供商来通过因特网连接)。在一些实施例中，通过利用计算机可读程序指令的状态信息来个性化定制电子电路，例如可编程逻辑电路、现场可编程门阵列(FPGA)或可编程逻辑阵列(PLA)，该电子电路可以执行计算机可读程序指令，从而实现本公开的各个方面。
这里所描述的计算机程序产品可以具体通过硬件、软件或其结合的方式实现。在一个可选实施例中，计算机程序产品具体体现为计算机存储介质，在另一个可选实施例中，计算机程序产品具体体现为软件产品，例如软件开发包(Software Development Kit，SDK)等等。
这里参照根据本公开实施例的方法、装置(系统)和计算机程序产品的流程图和/或框图描述了本公开的各个方面。应当理解，流程图和/或框图的每个方框以及流程图和/或框图中各方框的组合，都可以由计算机可读程序指令实现。
这些计算机可读程序指令可以提供给通用计算机、专用计算机或其它可编程数据处理装置的处理器，从而生产出一种机器，使得这些指令在通过计算机或其它可编程数据处理装置的处理器执行时，产生了实现流程图和/或框图中的一个或多个方框中规定的功 能/动作的装置。也可以把这些计算机可读程序指令存储在计算机可读存储介质中，这些指令使得计算机、可编程数据处理装置和/或其他设备以特定方式工作，从而，存储有指令的计算机可读介质则包括一个制造品，其包括实现流程图和/或框图中的一个或多个方框中规定的功能/动作的各个方面的指令。
也可以把计算机可读程序指令加载到计算机、其它可编程数据处理装置、或其它设备上，使得在计算机、其它可编程数据处理装置或其它设备上执行一系列操作步骤，以产生计算机实现的过程，从而使得在计算机、其它可编程数据处理装置、或其它设备上执行的指令实现流程图和/或框图中的一个或多个方框中规定的功能/动作。
附图中的流程图和框图显示了根据本公开的多个实施例的系统、方法和计算机程序产品的可能实现的体系架构、功能和操作。在这点上，流程图或框图中的每个方框可以代表一个模块、程序段或指令的一部分，模块、程序段或指令的一部分包含一个或多个用于实现规定的逻辑功能的可执行指令。在有些作为替换的实现中，方框中所标注的功能也可以以不同于附图中所标注的顺序发生。例如，两个连续的方框实际上可以基本并行地执行，它们有时也可以按相反的顺序执行，这依所涉及的功能而定。也要注意的是，框图和/或流程图中的每个方框、以及框图和/或流程图中的方框的组合，可以用执行规定的功能或动作的专用的基于硬件的系统来实现，或者可以用专用硬件与计算机指令的组合来实现。
本文已经公开了示例实施例，并且虽然采用了具体术语，但它们仅用于并仅应当被解释为一般说明性含义，并且不用于限制的目的。在一些实例中，对本领域技术人员显而易见的是，除非另外明确指出，否则可单独使用与特定实施例相结合描述的特征、特性和/或元素，或可与其他实施例相结合描述的特征、特性和/或元件组合使用。因此，本领域技术人员将理解，在不脱离由所附的权利要求阐明的本公开的范围的情况下，可进行各种形式和细节上的改变。

Claims (10)

  1. 一种客流预测方法，其中，所述方法包括：
    获取预设区域内的各个预设节点与其相邻节点之间的连通关系，所述预设区域包括多个所述预设节点及其之间的连接通道，所述连通关系用于表征所述连接通道的流通方向；
    获取多个所述预设节点对应的人员密度信息；
    依据多个所述预设节点对应的人员密度信息、各个所述预设节点与其相邻节点之间的连通关系，预测所述预设区域内的人员流动信息。
  2. 根据权利要求1所述的方法，其中，所述获取预设区域内的各个预设节点与其相邻节点之间的连通关系之后，所述获取多个所述预设节点对应的人员密度信息之前，所述方法还包括：
    依据多个所述预设节点、以及至少一个所述预设节点与其相邻节点之间的连通关系，构建邻接拓扑模型，所述邻接拓扑模型采用N阶方阵的形式表示，N为大于1的整数。
  3. 根据权利要求1所述的方法，其中，所述获取多个所述预设节点对应的人员密度信息，包括：
    获取多帧待处理图像，其中，所述待处理图像为采用设置于每个所述预设节点中的至少一个图像采集设备对每个所述预设节点内的人员的图像进行采集所获得的图像；
    依据深度学习算法对多帧所述待处理图像进行分析，确定至少一个所述预设节点对应的人员密度信息。
  4. 根据权利要求1所述的方法，其中，所述获取多个所述预设节点对应的人员密度信息之前，所述方法还包括：
    确定多个所述连接通道内是否存在监控摄像头；
    在确定所述连接通道内存在所述监控摄像头的情况下，依据客流统计算法，分别对多个所述连接通道内的人员流动信息进行统计，获得多个所述连接通道对应的客流统计信息。
  5. 根据权利要求4所述的方法，其中，所述客流统计信息包括：经过所述连接通道的人员数量；
    所述依据多个所述预设节点对应的人员密度信息、各个所述预设节点与其相邻节点之间的连通关系，预测所述预设区域内的人员流动信息，包括：
    依据各个所述预设节点与其相邻节点之间的连通关系，确定各个所述连接通道的流通方向；
    依据各个所述连接通道的流通方向，确定与各个所述连接通道对应的权重系数矩阵；
    依据所述权重系数矩阵、所述人员密度信息和经过所述连接通道的人员数量，确定线性估计模型；
    依据所述线性估计模型，对所述预设区域内的人员流动信息进行预测，获得区域流量预测结果。
  6. 根据权利要求5所述的方法，其中，所述客流统计信息，还包括：通道统计误差信息；
    所述依据所述线性估计模型，对所述预设区域内的人员流动信息进行预测，获得区域流量预测结果之前，所述方法还包括：
    依据所述通道统计误差信息和所述预设节点的预设客流总量信息对所述线性估计模型进行调整，获得更新后的线性估计模型。
  7. 根据权利要求4所述的方法，其中，所述依据多个所述预设节点对应的人员密度信息、各个所述预设节点与其相邻节点之间的连通关系，预测所述预设区域内的人员流动信息之后，所述方法还包括：
    每间隔预设时长，更新所述人员密度信息和/或所述客流统计信息。
  8. 根据权利要求1所述的方法，其中，每个所述预设节点包括：入口和出口；
    所述获取多个所述预设节点对应的人员密度信息，包括：
    获取经过每个所述预设节点的入口的第一通信交互信息，所述第一通信交互信息为经过所述入口的各个人员携带的设备与设置于所述入口处的预设设备之间的信息，
    获取经过每个所述预设节点的出口的第二通信交互信息，所述第二通信交互信息为经过所述出口的各个人员携带的设备与设置于所述出口处的预设设备之间的信息；
    依据所述第一通信交互信息和所述第二通信交互信息，确定所述预设节点对应的所述人员密度信息；
    其中，所述预设节点包括：进站节点、安全检查节点、检票节点、候车节点和乘车站台节点中的至少一种。
  9. 一种客流预测装置，其包括：
    第一获取模块，被配置为获取预设区域内的各个预设节点与其相邻节点之间的连通关系，所述预设区域包括多个预设节点及其之间的连接通道，所述连通关系用于表征所述连接通道的流通方向；
    第二获取模块，被配置为获取多个所述预设节点对应的人员密度信息；
    预测模块，被配置为依据多个所述预设节点对应的人员密度信息和多个所述连接通道内的客流统计信息，预测所述预设区域内的人员流动信息。
  10. 一种客流预测系统，其包括：设置于多个预设节点内的第一图像采集设备、设置于所述预设节点之间的连接通道内的第二图像采集设备和客流预测装置；
    所述第一图像采集设备，被配置为对与所述第一图像采集设备对应的预设节点内的人员的图像进行采集，获得多帧待处理图像，并向所述客流预测装置发送多帧所述待处理图像，以供所述客流预测装置依据深度学习算法对多帧所述待处理图像进行分析，确定至少一个所述预设节点对应的人员密度信息；
    所述第二图像采集设备，被配置为对所述预设节点之间的连接通道内的人员运动图像进行采集，获得通道图像，并对所述通道图像进行分析，获得与所述连接通道对应的客流统计信息，并向所述客流预测装置发送所述客流统计信息，以供所述客流预测装置依据客流统计算法，分别对多个所述连接通道内的人员流动信息进行统计，获得所述连接通道对应的客流统计信息；
    所述客流预测装置，被配置为执行如权利要求1至8中任一项所述的客流预测方法。
PCT/CN2023/116913 2022-10-19 2023-09-05 客流预测方法、装置和系统 WO2024082848A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202211278955.4A CN115358492B (zh) 2022-10-19 2022-10-19 客流预测方法、装置和系统
CN202211278955.4 2022-10-19

Publications (1)

Publication Number Publication Date
WO2024082848A1 true WO2024082848A1 (zh) 2024-04-25

Family

ID=84008767

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/116913 WO2024082848A1 (zh) 2022-10-19 2023-09-05 客流预测方法、装置和系统

Country Status (2)

Country Link
CN (1) CN115358492B (zh)
WO (1) WO2024082848A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115358492B (zh) * 2022-10-19 2023-03-24 通号通信信息集团有限公司 客流预测方法、装置和系统

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016180323A1 (zh) * 2015-05-12 2016-11-17 杭州海康威视数字技术股份有限公司 客流量的统计方法及装置
CN111611894A (zh) * 2020-05-14 2020-09-01 深圳中科慧据科技有限公司 人员轨迹预测方法、装置、计算机设备及存储介质
CN113077090A (zh) * 2021-04-09 2021-07-06 上海大学 客流预测方法、系统及计算机可读存储介质
CN115358492A (zh) * 2022-10-19 2022-11-18 通号通信信息集团有限公司 客流预测方法、装置和系统

Family Cites Families (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102496058B (zh) * 2011-11-11 2014-09-17 北京声迅电子股份有限公司 客流密度检测方法
CN108600965B (zh) * 2018-05-11 2020-05-22 南京邮电大学 一种基于客人位置信息的客流数据预测方法
US11032150B2 (en) * 2019-06-17 2021-06-08 International Business Machines Corporation Automatic prediction of behavior and topology of a network using limited information
CN110414715B (zh) * 2019-06-28 2023-06-09 武汉大学 一种基于社团检测的客流量预警方法
CN113269347B (zh) * 2021-03-31 2023-05-30 安徽农业大学 一种基于随机森林的高校快递网络节点流量预测方法
CN112988847B (zh) * 2021-04-20 2021-08-03 广东智九信息科技有限公司 一种基于大数据的景区人数预测系统以及方法

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2016180323A1 (zh) * 2015-05-12 2016-11-17 杭州海康威视数字技术股份有限公司 客流量的统计方法及装置
CN111611894A (zh) * 2020-05-14 2020-09-01 深圳中科慧据科技有限公司 人员轨迹预测方法、装置、计算机设备及存储介质
CN113077090A (zh) * 2021-04-09 2021-07-06 上海大学 客流预测方法、系统及计算机可读存储介质
CN115358492A (zh) * 2022-10-19 2022-11-18 通号通信信息集团有限公司 客流预测方法、装置和系统

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
CHEN, WEIYA ET AL.: "Dynamic Monitoring And Control System For Metro Station Passenger Organization Based on Multi-Source Data", URBAN PUBLIC TRANSPORT, 15 April 2017 (2017-04-15) *
MA, PEI: "The Calculation Evaluation of Large Passenger Flow Transport Capacity Theory And Simulation of Rail Transit Station", SHANXI ARCHITECTURE, 1 March 2018 (2018-03-01), pages 132 - 133 *

Also Published As

Publication number Publication date
CN115358492A (zh) 2022-11-18
CN115358492B (zh) 2023-03-24

Similar Documents

Publication Publication Date Title
Zhu et al. A probabilistic passenger-to-train assignment model based on automated data
WO2024082848A1 (zh) 客流预测方法、装置和系统
US10491501B2 (en) Traffic-adaptive network control systems and methods
Jenelius Data-driven metro train crowding prediction based on real-time load data
Sangoboye et al. Plcount: A probabilistic fusion algorithm for accurately estimating occupancy from 3d camera counts
US20140309977A1 (en) Performing-time-series based predictions with projection thresholds using secondary time-series-based information stream
US10832416B2 (en) Crowd flow rate estimation
Zhang et al. Dynamic orienteering on a network of queues
WO2019105269A1 (zh) 作业人员评价系统、作业人员评价装置及评价方法
US9836979B2 (en) Method and system for latent demand modeling for a transportation system
CN104899577B (zh) 建筑物内的人员数量确定方法和人群疏散方法
Mohottige et al. Role of campus WiFi infrastructure for occupancy monitoring in a large university
US10586130B2 (en) Method, system and apparatus for providing access to videos
US8620675B2 (en) Method and apparatus for optimal service channel reconfiguration based on multi-agent simulation
JP2018103924A (ja) 混雑予測装置
EP3425606B1 (en) Traffic situation estimation system and traffic situation estimation method
Carrel et al. A framework for evaluating operations control on a metro line: integrating multiple perspectives and automatically collected train and passenger movement data
CN116486332A (zh) 客流量监测方法、装置、设备及存储介质
Sutjarittham et al. Measuring and modeling car park usage: Lessons learned from a campus field-trial
Mishalani et al. Evaluating real-time origin-destination flow estimation using remote sensing-based surveillance data
Ding et al. Parallel queues with discrete-choice arrival pattern: Empirical evidence and asymptotic characterization
De Vita et al. Traffic Condition Estimation at the Smart City Edge using Deep Learning: A Ro-Pax Terminal Case Study
WO2019111866A1 (ja) 分析装置、通信システム、データ処理方法、及び非一時的なコンピュータ可読媒体
Denman et al. Large scale monitoring of crowds and building utilisation: A new database and distributed approach
Tuncel et al. An Unsupervised Learning Approach for Robust Denied Boarding Probability Estimation Using Smart Card and Operation Data in Urban Railways