CN112380925A

CN112380925A - Data processing method, device, equipment and storage medium

Info

Publication number: CN112380925A
Application number: CN202011165351.XA
Authority: CN
Inventors: 吴正中; 冯帆; 常海利; 汪永刚
Original assignee: Beijing Urban Construction Intelligent Control Technology Co ltd
Current assignee: Beijing Urban Construction Intelligent Control Technology Co ltd
Priority date: 2020-10-27
Filing date: 2020-10-27
Publication date: 2021-02-19

Abstract

An embodiment of the application provides a data processing method, a device, equipment and a storage medium, wherein the method comprises the following steps: acquiring historical passenger flow information of a position to be released of a medium and corresponding environment characteristic information; predicting a passenger flow prediction value of a future predetermined time period through a passenger flow prediction model based on historical passenger flow information and the environmental characteristic information; predicting a passenger flow residence time prediction value of the preset time period through a passenger flow residence time prediction model based on historical passenger flow residence time information and environmental characteristic information; and evaluating the media delivery value of the media delivery position in a preset time period based on the passenger flow predicted value and the passenger flow residence time predicted value, wherein the passenger flow prediction model and the passenger flow residence time prediction model are machine learning models. According to the technical scheme of the embodiment of the application, the media delivery value of the media to-be-delivered position in each preset time period is accurately evaluated, so that the media delivery value can be maximized.

Description

Data processing method, device, equipment and storage medium

Technical Field

The embodiment of the application relates to the technical field of machine learning, in particular to a data processing method, a device, equipment and a storage medium.

Background

In the rail transit field, such as the subway field, media are popular among advertisers due to the advantages of large audience quantity, rich expressive force and the like. Therefore, how to evaluate the media delivery value in the rail transit field becomes the focus of attention.

At present, in most of technical schemes for evaluating media delivery value, a predicted click rate corresponding to a media to be delivered is obtained through a click rate prediction model according to characteristic information of the media to be delivered and user information, and the media delivery value is determined according to the predicted click rate. However, this solution is not suitable for evaluating the media delivery value in the rail transit field because the user information needs to be acquired.

Therefore, how to accurately evaluate the media delivery value in the rail transit field becomes a technical problem to be solved urgently.

Disclosure of Invention

The embodiment of the application provides a data processing method, a data processing device, data processing equipment and a storage medium, which are used for solving the problem of accurately evaluating the media delivery value in the field of rail transit.

In a first aspect of the embodiments of the present application, a data processing method is provided, including:

obtaining historical passenger flow information of a position to be released of a medium and corresponding environment characteristic information, wherein the historical passenger flow information comprises historical passenger flow information and historical passenger flow residence time information, and the position to be released of the medium is a preset area of a target station;

predicting a passenger flow predicted value of a future predetermined time period through a passenger flow prediction model based on the historical passenger flow information and the environmental characteristic information;

predicting a passenger flow residence time prediction value of the preset time period through a passenger flow residence time prediction model based on the historical passenger flow residence time information and the environment characteristic information;

and evaluating the media delivery value of the position to be delivered with the media in the preset time period based on the passenger flow predicted value and the passenger flow residence time predicted value, wherein the passenger flow prediction model and the passenger flow residence time prediction model are machine learning models.

In some embodiments of the present application, based on the above scheme, the obtaining historical passenger flow information of the to-be-delivered-location of the media includes:

acquiring passenger flow images of the media to-be-released position history in each time period shot by a camera, and determining first history passenger flow of the media to-be-released position history in each time period according to the passenger flow images;

and/or the presence of a gas in the gas,

acquiring mobile terminal signal data of each time period of the history of the position to be released of the media, which is determined by a wireless local area network sniffing technology, and determining second history passenger flow of each time period of the history of the position to be released of the media according to the mobile terminal signal data;

and/or the presence of a gas in the gas,

and acquiring passenger flow data of each historical time period of the gate of the position to be released of the media, and determining the third historical passenger flow of each historical time period of the historical position to be released of the media according to the passenger flow data.

In some embodiments of the present application, based on the above scheme, the method further includes:

and performing weighted summation operation on the first history passenger flow volume, the second history passenger flow volume and the third history passenger flow volume to obtain the passenger flow volume of the media position to be released in each time period.

In some embodiments of the present application, based on the above solution, the predicting, by a passenger flow prediction model, a passenger flow prediction value of a predetermined time period in the future based on the historical passenger flow information and the environmental characteristic information includes:

training the passenger flow prediction model based on the historical passenger flow information and the environmental characteristic information, wherein the environmental characteristic information comprises one or more of date, weather and traffic congestion degree corresponding to the historical passenger flow information;

acquiring environmental characteristic information corresponding to the preset time period;

and predicting the passenger flow prediction value of the preset time period through the trained passenger flow prediction model based on the environmental characteristic information corresponding to the preset time period.

In some embodiments of the present application, based on the above scheme, the training the passenger flow prediction model based on the historical passenger flow information and the environmental characteristic information includes:

generating a corresponding environment feature vector based on the environment feature information;

taking the passenger flow corresponding to the historical passenger flow information as a passenger flow label of the environment characteristic vector;

and training the passenger flow prediction model based on the environment feature vector and the passenger flow label.

In some embodiments of the present application, based on the above scheme, the obtaining historical passenger flow residence time information of the location to be dropped includes:

acquiring mobile terminal signal data of each time period of the historical position to be launched of the media, which is determined by a wireless local area network sniffing technology;

acquiring train running time information, corresponding train full load rate and passenger flow volume data of each time period of the historical position of the media to be released;

and determining the passenger flow residence time of each time period of the media to-be-released position history based on the mobile terminal signal data, the train running time information, the train full load rate and the passenger flow data.

In some embodiments of the application, based on the above scheme, the determining a passenger flow residence time of each time period of the media to-be-delivered location history based on the mobile terminal signal data, the train operation time information, the train full load rate, and the passenger flow volume data includes:

determining a mobile terminal and corresponding residence time of each time period of the historical position of the media to be released based on the signal data of the mobile terminal, and determining first passenger flow residence time of each time period of the position of the media to be released based on the mobile terminal and the corresponding residence time;

determining second passenger flow residence time of each time period of the media to-be-released position history based on train running time information of each time period of the media to-be-released position history, corresponding train full load rate and the passenger flow data;

and performing weighted operation on the first passenger flow residence time and the second passenger flow residence time, and determining the passenger flow residence time of each time period of the history of the positions where the media are to be released.

In some embodiments of the present application, based on the above solution, the predicting a predicted value of the passenger flow residence time in the predetermined time period by a passenger flow residence time prediction model based on the historical passenger flow residence time information and the environmental characteristic information includes:

training the passenger flow residence time prediction model based on the historical passenger flow residence time information and the environmental characteristic information, wherein the environmental characteristic information comprises one or more of date, weather and traffic congestion degree corresponding to the historical passenger flow information;

and predicting the passenger flow residence time predicted value of the preset time period through the trained passenger flow residence time prediction model based on the environmental characteristic information corresponding to the preset time period.

In some embodiments of the present application, based on the above scheme, the training the passenger flow residence time prediction model based on the historical passenger flow residence time information and the environmental characteristic information includes:

taking the passenger flow residence time corresponding to the historical passenger flow residence time information as a passenger flow residence time label of the environment feature vector;

and training a passenger flow residence time prediction model based on the environment feature vector and the passenger flow residence time label.

In some embodiments of the present application, based on the above scheme, the evaluating a media delivery value of the location to be delivered with media in the predetermined time period based on the passenger flow volume predicted value and the passenger flow residence time predicted value includes:

and carrying out weighted operation on the passenger flow volume predicted value and the passenger flow residence time predicted value of the media to-be-released position in each time period of a preset date, and determining the media release value of the media to-be-released position in the preset date.

In some embodiments of the present application, based on the above scheme, the determining, by the passenger flow image, a first historical passenger flow volume of each time period of the historical location to be delivered with media includes:

segmenting the passenger flow image to obtain a head image corresponding to the passenger flow image;

and detecting the head image through a head detection model to obtain the passenger flow volume corresponding to the passenger flow image.

marking the head image to generate a label corresponding to the customer image;

carrying out gray processing, drying and smoothing processing and feature extraction on the passenger flow image to generate sample features;

and training the human head detection model through the sample characteristics and the labels corresponding to the passenger flow images.

and marking the head of the passenger flow image through pedestrian re-identification processing, and counting according to a marking result to obtain the passenger flow volume corresponding to the passenger flow image.

In a second aspect of the embodiments of the present application, there is provided a data processing apparatus, including:

the system comprises an information acquisition module, a service management module and a service management module, wherein the information acquisition module is used for acquiring historical passenger flow information of a position to be released of a medium and corresponding environment characteristic information, the historical passenger flow information comprises historical passenger flow information and historical passenger flow residence time information, and the position to be released of the medium is a preset area of a target station;

the passenger flow prediction module is used for predicting a passenger flow prediction value of a future predetermined time period through a passenger flow prediction model based on the historical passenger flow information and the environmental characteristic information;

the residence time prediction module is used for predicting a passenger flow residence time prediction value of the preset time period through a passenger flow residence time prediction model based on the historical passenger flow residence time information and the environmental characteristic information;

and the evaluation module is used for evaluating the media delivery value of the position to be delivered with the media in the preset time period based on the passenger flow predicted value and the passenger flow residence time predicted value.

In some embodiments of the present application, based on the above scheme, the information obtaining module is configured to:

and/or the presence of a gas in the gas,

In some embodiments of the present application, based on the above scheme, the information obtaining module is further configured to:

In some embodiments of the present application, based on the above solution, the passenger volume prediction module includes:

a first training unit, configured to train the passenger flow prediction model based on the historical passenger flow information and the environmental characteristic information, where the environmental characteristic information includes one or more of a date, weather, and a traffic congestion degree corresponding to the historical passenger flow information;

a first characteristic obtaining unit, configured to obtain environmental characteristic information corresponding to the predetermined time period;

and the passenger flow prediction unit is used for predicting the passenger flow prediction value of the preset time period through the trained passenger flow prediction model based on the environmental characteristic information corresponding to the preset time period.

In some embodiments of the present application, based on the above scheme, the first training unit is configured to:

In some embodiments of the present application, based on the above scheme, the information obtaining module is further configured to include:

In some embodiments of the present application, based on the above scheme, the residence time prediction module includes:

a second training unit, configured to train the passenger flow residence time prediction model based on the historical passenger flow residence time information and the environmental characteristic information, where the environmental characteristic information includes one or more of a date, weather, and a traffic congestion degree corresponding to the historical passenger flow volume information;

the second characteristic acquisition unit is used for acquiring environmental characteristic information corresponding to the preset time period;

and the residence time prediction unit is used for predicting the predicted value of the passenger flow residence time in the preset time period through the trained passenger flow residence time prediction model based on the environment characteristic information corresponding to the preset time period.

In some embodiments of the present application, based on the above scheme, the second training unit is configured to:

In some embodiments of the present application, based on the above, the evaluation module is configured to:

marking the head image to generate a label corresponding to the customer image;

In a third aspect of the embodiments of the present application, there is provided a data processing apparatus, including: a receiver, a processor, a memory, and a transmitter; the memory is used for storing computer programs and data, and the processor calls the computer programs stored in the memory to execute the data processing method provided by any embodiment of the first aspect.

A fourth aspect of the embodiments of the present application provides a computer-readable storage medium, where the computer-readable storage medium includes a computer program, and the computer program is configured to, when executed by a processor, perform the data processing method provided in any one of the embodiments of the first aspect.

According to the data processing method, the device, the equipment and the storage medium provided by the embodiment of the application, on one hand, by combining historical passenger flow, passenger flow residence time characteristics and environmental characteristics such as date, weather, traffic congestion degree and other characteristic data of the position to be delivered with the media, a passenger flow predicted value and a passenger flow residence time predicted value of a future preset time period of the position to be delivered with the media are predicted through a machine learning model, and the passenger flow predicted value and the passenger flow residence time predicted value of the future preset time period of the position to be delivered with the media can be accurately predicted; on the other hand, the passenger flow predicted value and the passenger flow residence time predicted value of each time period in the future of the position to be delivered with the media can be accurately predicted, so that the media delivery value of the position to be delivered with the media at the rail transit station in each preset time period can be accurately evaluated based on the passenger flow predicted value and the passenger flow residence time predicted value, and the media delivery value can be maximized.

Drawings

In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present application, and for those skilled in the art, other drawings can be obtained according to these drawings without inventive exercise.

FIG. 1 is a schematic flow diagram of a data processing method provided in accordance with some embodiments of the present application;

FIG. 2 is a schematic flow chart diagram of a data processing method according to other embodiments of the present application;

FIG. 3 is a schematic flow diagram of passenger flow prediction model training provided in accordance with some embodiments of the present application;

FIG. 4 is a schematic flow diagram of a passenger flow residence time prediction model provided in accordance with some embodiments of the present application;

FIG. 5 is a schematic flow chart of a media value assessment for a media to-be-delivered location according to some embodiments of the present application;

FIG. 6 is a schematic illustration of obtaining environmental characteristic information provided in accordance with some embodiments of the present application;

FIG. 7 is a schematic flow chart of determining passenger flow provided in accordance with some embodiments of the present application;

FIG. 8 is a schematic block diagram of a data processing apparatus provided in accordance with some embodiments of the present application;

FIG. 9 is a schematic block diagram of a passenger flow prediction module provided in accordance with some embodiments of the present application;

FIG. 10 is a schematic block diagram of a residence time prediction module provided in accordance with some embodiments of the present application;

fig. 11 is a schematic block diagram of embodiments of a data processing apparatus provided in accordance with some embodiments of the present application.

Detailed Description

In order to make the objects, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present application.

At present, in most of technical schemes for evaluating the media delivery value, a predicted click rate corresponding to a media to be delivered is obtained through a click rate prediction model according to characteristic information of the media or advertisements to be delivered and user information, and the media delivery value is determined according to the predicted click rate. However, this solution is not suitable for evaluating the media delivery value in the rail transit field because the user information needs to be acquired.

Further, the delivery value of rail transit media such as subway media is greatly affected by the volume of passenger traffic and the residence time of passenger traffic, compared to conventional media. Based on the above, the basic idea of the embodiments of the present application is: the method comprises the steps of predicting a passenger flow predicted value and a passenger flow residence time predicted value of a media position to be released in a future preset time period by acquiring historical passenger flow and passenger flow residence time characteristics of the media position to be released in rail transit and combining with characteristic data such as date, weather and traffic congestion, and evaluating the media release value of the media position to be released in the future preset time period based on the passenger flow predicted value and the passenger flow residence time predicted value of the media position to be released. By combining the historical passenger flow, the passenger flow residence time characteristic, the date, the weather, the traffic congestion degree and other characteristic data of the position to be delivered with the media, the passenger flow predicted value and the passenger flow residence time predicted value of the position to be delivered with the media in the future predetermined time period can be more accurately predicted, and therefore the media delivery value of the position to be delivered with the media in the future predetermined time period can be more accurately evaluated based on the passenger flow predicted value and the passenger flow residence time predicted value.

Fig. 1 is a schematic flow diagram of a data processing method provided according to some embodiments of the present application. The data processing method may be applied to a server, where the server may be a physical server including an independent host, or a virtual server carried by a host cluster, or a cloud server, and the method includes steps S110 to S140, and the data processing method in the example embodiment is described in detail below with reference to the drawings.

In step S110, historical passenger flow information of the position to be delivered with the medium and corresponding environmental characteristic information are obtained, the historical passenger flow information includes historical passenger flow volume information and historical passenger flow residence time information, and the position to be delivered with the medium is a predetermined area of the target station.

In an exemplary embodiment, the location to be delivered with the media may be a predetermined area of a subway station, a high-speed railway station, or a railway station, which is not particularly limited in this application, and the location to be delivered with the media is described as a subway station. The environmental characteristic information is an environmental characteristic related to the passenger flow volume and the passenger flow residence time, and for example, the environmental characteristic information includes: and one or more of date, weather and traffic congestion degree corresponding to the historical passenger flow information. In an example embodiment, date and weather information corresponding to the historical passenger flow information is acquired through a web crawler technology, and traffic congestion degree information corresponding to the historical passenger flow information is acquired through an application program interface provided by a traffic department.

It should be noted that the environmental characteristic information may also be other suitable information, for example, the environmental characteristic may also be train operation time information, and important event or holiday information, and the like, which is also within the protection scope of the present application.

Further, in an example embodiment, the historical passenger flow information includes historical passenger flow information, and the passenger flow information of each historical time period of the target subway station may be obtained in one or more of the following three ways:

the first method is as follows: obtaining passenger flow images of the media to-be-released position in each historical time period shot by a camera, and determining the historical passenger flow of the media to-be-released position in each historical time period through the passenger flow images;

the second method comprises the following steps: acquiring mobile terminal signal data of each time period of the history of the position to be released of the media, which is determined by a wireless local area network sniffing technology, and determining second history passenger flow of each time period of the history of the position to be released of the media according to the mobile terminal signal data;

the third method comprises the following steps: and passenger flow data of each time period of the history of the gate of the position to be released of the media is obtained, and third history passenger flow of each time period of the history of the position to be released of the media is determined according to the passenger flow data.

Furthermore, in an example embodiment, the historical passenger flow information includes historical passenger flow residence time information, and obtaining the historical passenger flow residence time information of the location where the medium is to be delivered includes: acquiring mobile terminal signal data of each time period of a historical position to be launched of a medium, which is determined by a wireless local area network sniffing technology; acquiring train running time information, corresponding train full load rate and passenger flow volume data of each time period of historical positions where media are to be delivered; and determining the passenger flow residence time of each time period of the history of the positions where the media are to be put based on the mobile terminal signal data, the train running time information, the train full load rate and the passenger flow volume data.

In step S120, a predicted value of the passenger flow volume for a predetermined period of time in the future is predicted by the passenger flow volume prediction model based on the historical passenger flow volume information and the corresponding environmental characteristic information.

In an example embodiment, a passenger flow prediction model is trained based on historical passenger flow information and the environmental characteristic information, wherein the environmental characteristic information comprises one or more of date, weather and traffic congestion degree corresponding to the historical passenger flow information; acquiring environmental characteristic information corresponding to a preset time period; and predicting the passenger flow prediction value of the preset time period through the trained passenger flow prediction model based on the environmental characteristic information corresponding to the preset time period. For example, an environment feature vector corresponding to a predetermined time period is generated, and the environment feature vector is input into the trained passenger flow prediction model to obtain a passenger flow prediction value of the predetermined time period.

It should be noted that the passenger flow prediction model may be a deep learning network model, a random forest model, or a logistic regression model, or may be other suitable machine learning models, such as a gradient boosting decision tree model or a bayesian model, which is not particularly limited in this application.

Further, in an example embodiment, training the passenger flow prediction model based on the historical passenger flow information and the environmental characteristic information includes: generating a corresponding environment characteristic vector based on environment characteristic information corresponding to the historical passenger flow volume information; taking the passenger flow corresponding to the historical passenger flow information as a passenger flow label of the environment characteristic vector; and training the passenger flow prediction model based on the environment feature vector and the passenger flow label.

In step S130, a predicted value of the passenger flow residence time in the predetermined time zone is predicted by the passenger flow residence time prediction model based on the historical passenger flow residence time information and the corresponding environmental characteristic information.

In an example embodiment, a passenger flow residence time prediction model is trained based on historical passenger flow residence time information and environmental characteristic information, wherein the environmental characteristic information comprises one or more of date, weather and traffic congestion degree corresponding to the historical passenger flow information; acquiring environmental characteristic information corresponding to a preset time period; and predicting the passenger flow residence time predicted value of the preset time period through the trained passenger flow residence time prediction model based on the environmental characteristic information corresponding to the preset time period. For example, an environment feature vector corresponding to a predetermined time period is generated, and the environment feature vector is input to the trained passenger flow residence time prediction model to predict a predicted value of the passenger flow residence time in the predetermined time period.

It should be noted that the passenger flow residence time prediction model may be a deep learning network model, a random forest model, or a logistic regression model, or may be other suitable machine learning models, such as a gradient boosting decision tree model or a bayesian model, which is not particularly limited in this application.

Further, in an example embodiment, training the passenger flow residence time prediction model based on historical passenger flow residence time information and environmental characteristic information includes: generating a corresponding environment feature vector based on the environment feature information; taking the passenger flow residence time corresponding to the historical passenger flow residence time information as a passenger flow residence time label of the environment feature vector; and training a passenger flow residence time prediction model based on the environment feature vector and the passenger flow residence time label.

In step S140, the media delivery value of the media to-be-delivered position in the predetermined time period is evaluated based on the passenger flow volume prediction value and the passenger flow residence time prediction value.

In an example embodiment, the effective value rate of media delivery in the predetermined time period is used as a weight, a weighted operation is performed on the predicted value of passenger flow and the predicted value of residence time of passenger flow in the predetermined time period, and the value of media delivery in the predetermined time period at the position to be delivered in the media is determined, wherein the effective value rate of media delivery is related to unit passenger flow and unit residence time of passenger flow. For example, the media delivery value rate per unit passenger flow volume and per unit residence time is obtained in advance from the media delivery value statistical data, and the media delivery value in the predetermined time period is evaluated using the media delivery value rate as a weight.

Further, if the total media delivery value of the scheduled date is to be acquired, performing weighted operation on the passenger flow volume predicted value and the passenger flow residence time predicted value of the media to-be-delivered position in each time period of the scheduled date, and determining the media delivery value of the media to-be-delivered position on the scheduled date.

According to the data processing method in the exemplary embodiment of fig. 1, on one hand, by combining the historical passenger flow volume, the passenger flow residence time characteristic, and the environmental characteristic data such as date, weather, traffic congestion degree, etc. of the position to be delivered with the media, the predicted value of the passenger flow volume and the predicted value of the passenger flow residence time in the future predetermined time period of the position to be delivered with the media are predicted by the machine learning model, so that the predicted value of the passenger flow volume and the predicted value of the passenger flow residence time in the future predetermined time period of the position to be delivered with the media can be accurately; on the other hand, the passenger flow predicted value and the passenger flow residence time predicted value of each time period in the future of the position to be delivered with the media can be accurately predicted, so that the media delivery value of the position to be delivered with the media at the rail transit station in each preset time period can be accurately evaluated based on the passenger flow predicted value and the passenger flow residence time predicted value, and the media delivery value can be maximized.

Fig. 2 is a schematic flow chart of a data processing method according to another embodiment of the present application. The data processing method may be applied to a server, where the server may be a physical server including an independent host, or a virtual server borne by a host cluster, or a cloud server, and the data processing method in the example embodiment is described in detail below with reference to the accompanying drawings.

In step S210, historical passenger flow information of each time segment of the historical predetermined time interval is obtained, and the historical passenger flow information includes historical passenger flow volume information and historical passenger flow residence time information.

In an example embodiment, traffic volume and traffic residence time data for various time periods of one or more days in the history are obtained. The passenger flow information of each historical time period of the target subway station can be obtained in one or more of the following three ways:

the first method is as follows: and obtaining passenger flow images of the media to-be-released position in each historical time period shot by the camera, and determining the historical passenger flow of the media to-be-released position in each historical time period through the passenger flow images. For example, time periods are divided from one day of a historical date, one day is divided into M time periods according to a certain time interval, wherein i represents the ith time period of the divided day, a passenger flow image of the current media to-be-delivered position acquired through a camera is transmitted into a human head detection model, the human head in the passenger flow image is identified through the human head detection model, and the human head in continuous image frames is subjected to target tracking through Re-id (pedestrian Re-Identification) processing, so that whether the passenger flow image is the same Person or not is judged, the number of the human head is counted, and the passenger flow A in the ith time period of the day is acquired_i。

In step S220, environment characteristic information corresponding to the historical passenger flow information, such as the environment characteristic information of week, month, year, weather, ground traffic congestion, important activities, and the like, is obtained.

In an example embodiment, the environmental characteristic information corresponding to the historical passenger flow information is acquired from an environmental characteristic database, and date data, traffic jam data and weather data of each historical time period are stored in the environmental characteristic information database. For example, referring to fig. 6, date information corresponding to the historical passenger flow information, e.g., week, month, year, is acquired from the date database 610; acquiring weather information corresponding to the historical passenger flow information from the server 620 through a web crawler technology; the ground traffic congestion degree information corresponding to the historical passenger flow information is acquired from the traffic congestion degree database through an Application Programming Interface (API), which may be an Interface of the traffic congestion degree database corresponding to a traffic map legend high-grade map. The acquired date information, weather information, and traffic congestion degree information are stored in the environmental characteristic database 600.

In step S230, a passenger flow prediction model is trained based on the historical passenger flow information and the corresponding environmental characteristic information. For example, based on the environmental characteristic information corresponding to the historical passenger flow volume information, a corresponding environmental characteristic vector is generated; taking the passenger flow corresponding to the historical passenger flow information as a passenger flow label of the environment characteristic vector; and training the passenger flow prediction model based on the environment feature vector and the passenger flow label.

In step S240, the environmental characteristic information of the future predetermined time period, such as the passenger flow, the week, the month, the year, the weather, the ground traffic congestion, the important activities, and the like, is obtained and input into a passenger flow prediction value model after training, so as to obtain the passenger flow prediction value of the future predetermined time period.

In step S250, a passenger flow volume residence time prediction model is trained based on the historical passenger flow residence time information and the environmental characteristic information. For example, based on the environmental feature information, generating a corresponding environmental feature vector; taking the passenger flow residence time corresponding to the historical passenger flow residence time information as a passenger flow residence time label of the environment feature vector; and training a passenger flow residence time prediction model based on the environment feature vector and the passenger flow residence time label.

In step S260, environmental characteristic information of a predetermined time period in the future, for example, characteristic information such as the residence time of passenger flow, week, month, year, weather, ground traffic congestion, and important activities, is obtained, and an environmental characteristic vector corresponding to the obtained environmental characteristic information is input into a passenger flow residence time prediction value model after training, so as to obtain a passenger flow residence time prediction value of the time period in the future.

In step S270, the passenger flow residence time predicted value and the passenger flow volume predicted value are introduced into the constructed subway media delivery value evaluation model to obtain a media delivery value metric value of the media to-be-delivered position in a future period of time, which is used to evaluate the media delivery value of the media to-be-delivered position in a future predetermined period of time. The media to-be-delivered position can be a subway station, a high-speed railway station, a train station and the like, and the application is not particularly limited to this.

Fig. 3 is a schematic flow chart illustrating training of a passenger flow prediction model according to some embodiments of the present application.

Referring to fig. 3, in step S310, image data of a position to be placed of a medium, which is captured by a camera, device signal data obtained by wireless local area network sniffing, and passenger flow data passing through a station gate are obtained.

In an example embodiment, as shown with reference to fig. 6, date information corresponding to historical passenger flow information, e.g., week, month, year, is acquired from the date database 610; acquiring weather information corresponding to the historical passenger flow information from the server 620 through a web crawler technology; the ground traffic congestion degree information corresponding to the historical passenger flow information is acquired from the traffic congestion degree database through an Application Programming Interface (API), which may be an Interface of the traffic congestion degree database corresponding to a traffic map legend high-grade map. The acquired date information, weather information, and traffic congestion degree information are stored in the environmental characteristic database 600.

In step S320, passenger flow volume information of each time segment of the history date is acquired according to the image data shot by the camera, the device signal data acquired by the wireless local area network sniffing, and the passenger flow data passing through the station gate.

In the exemplary embodiment, the passenger volume for each time period of the history date is acquired by the following three ways.

Time periods are divided into one day of the historical date, and one day is divided into M time periods according to a certain time interval, wherein i represents the ith time period of the divided day.

The first method is as follows: the method comprises the steps of transmitting a passenger flow image of a current media to-be-released position acquired through a camera into a human head detection model, identifying the human head in the passenger flow image through the human head detection model, carrying out target tracking on the human head in continuous image frames through Re-Identification (Re-Identification) processing, judging whether the image is the same Person or not, counting the number of the human head, and acquiring the passenger flow A in the ith time period in one day_i。

FIG. 7 is a schematic flow chart for determining passenger flow according to some embodiments of the present application.

Referring to fig. 7, in step S710, images of a passenger flow passing through a position to be delivered of media captured by a camera at different time periods are collected.

In step S715, the collected passenger flow image is divided into head images having fixed size areas according to a predetermined specification.

In step S720, the human head image is subjected to gray scale processing, denoising smoothing processing, and feature extraction.

In step S725, the sample library data is trained using the strong classification algorithm and the Haar algorithm.

In step S730, a human head detection model is obtained based on the training result in step S725.

In step S735, the serialized image data processed in step S720 is input into a human head detection model as input, the human head region of the image is segmented according to the same specification size as the sample in the sample library through the human head detection model to obtain a plurality of segmented sub-regions, and the central position of each segmented sub-region is the centroid position of the human head;

in step S740, the human head is marked by performing similarity comparison on the image sub-regions segmented in step S735 using re-id processing.

In step S745, the head data passing through the image area is counted based on the labeling result, and the passenger flow volume, which is the total number of people, is detected.

In step S750, history video stream data is acquired.

In step S755, the history video stream data is processed to obtain image frame data.

In step S760, performing gray scale processing, drying smoothing processing, and feature extraction on the image frame data to generate a sample feature library;

in step 765, the yolo algorithm is trained using the sample features in the sample feature library to generate a yolo model for head recognition. The method comprises the steps of utilizing a trained yolo model to conduct human head recognition on image frame data, comparing recognized human head images with human head images recognized by utilizing a strong classification algorithm and a Harr model, if one recognized human head image is the human head image and the other recognized human head image is not the human head image, putting the image into a later-stage manual check library for marking, and training the strong classification algorithm and the Harr again to obtain a new human head detection model and updating the yolo model.

According to the technical scheme in the example embodiment of fig. 7, the accuracy of passenger flow detection can be improved, the detection efficiency is high, and the method and the device can be applied to stations, shopping malls and other places, so that the media delivery value can be maximized.

The second method comprises the following steps: method for acquiring media to be launched by using Wireless Fidelity (WIFI) sniffing technologyPlacing positions, namely the number of WIFI devices in the current media to-be-placed area, and bringing the number of the WIFI devices into a common crowd proportion model of the electronic equipment so as to obtain the passenger flow B of the ith time period of a day calculated by a WIFI sniffing technology_iSpecifically, as shown in the calculation formula (1):

B_i＝wifi_num_i/(1-ε) (1)

wherein, B_iRepresenting the amount of passenger flow in the ith time period of the day, WIFI _ num, acquired by using the WIFI sniffing technology_iThe number of the electronic equipment in the ith time period of a day acquired by using a WIFI sniffing technology is represented, and epsilon represents the proportion of the number of the infrequent electronic equipment to the total number of people.

The third method comprises the following steps: the gate is used for obtaining the passenger flow of the position to be released of the incoming and outgoing media and bringing the passenger flow into a passenger flow calculation model, the passenger flow calculation model is a statistical model, and the passenger flow C of the ith time period of a day is obtained_i。

In summary, the passenger flow volume calculation of the ith time slot of the day of the media to-be-delivered position is shown in the calculation formula (2):

PL_i＝A_i×α₁+B_i×α₂+C_i×α₃ (2)

wherein PL_iIndicating the traffic volume, alpha, of the ith time period of the day₁、α₂、α₃Respectively representing the confidence or weight of the passenger flow obtained by the three methods, A_iRepresenting the amount of traffic in the ith time period of the day, obtained by means of a camera, B_iRepresenting the amount of passenger traffic in the ith time period of the day, C, obtained using WIFI sniffing techniques_iIndicating the amount of traffic acquired using the gate for the ith time period of the day.

In S330, the passenger flow volume information and the environmental characteristics are newly input to the passenger flow volume prediction model, and the passenger flow volume prediction model is trained.

In an example embodiment, the passenger flow prediction model is a deep learning network model or a random forest model or a logistic regression model or a gradient boosting decision tree model. Taking a neural network model as a passenger flow prediction model as an example for explanation, external environment feature data such as weather and traffic congestion degrees and environment features of week, month and year data can be used as input vectors, historical passenger flow is used as a label, and predicted passenger flow is used as an output vector. Inputting the environment characteristic data and the label into the neural network model to train the neural network model, judging whether the neural network model is converged or not through a loss function in the training process, stopping training if the neural network model is converged, and continuing training through adjusting the parameters of the set model until the model is converged if the neural network model is not converged.

Fig. 4 is a schematic flow diagram of a passenger flow residence time prediction model provided according to some embodiments of the present application.

In step S410, device signal data, train operation time information, train full load rate, and corresponding passenger flow volume data obtained by wireless lan sniffing are acquired. The passenger flow volume data may be the passenger flow volume information acquired in step S320.

In step S420, the passenger flow residence time of each time period of the historical date per day is calculated based on the device signal data, the train operation time information, the train full rate, and the passenger flow volume data.

In an example embodiment, time periods are divided into one day and the other day of the historical date, and the one day is divided into M time periods according to a certain time interval, wherein i represents the ith time period of the divided day.

The residence time of each electronic device in a preset area of a media to-be-released position, namely a target station, is acquired through WIFI sniffing, and the first passenger flow residence time ARt in the ith time period acquired through the WIFI sniffing technology in one day_iThe calculation is shown in equation (3):

wherein, ARt_iRepresenting the passenger residence time of the ith time period, N, acquired by wifi sniffing in one day_iRepresenting the number of electronic devices acquired by utilizing wifi sniffing technology in the ith time period, wherein j representsUtilize wifi to sniff the j electronic equipment who acquires, t_jiRepresents the dwell time, beta, of the j-th electronic equipment ascertained in the ith time period of the day at the station_iA passenger retention time effective value factor representing the ith time period of the day.

Further, the second passenger flow residence time BRt of different time periods in a day can be obtained through the train running time chart, the full load rate and the passenger flow data of different time periods in the day_i。

Therefore, the calculation of the residence time of the passenger flow in the ith time zone in the day is shown in the calculation formula (4):

EPST_i＝ARt_i×λ₁+BRt_i×λ₂ (4)

wherein, EPST_iDenotes the traffic volume, lambda, of the ith time period of the day₁、λ₂Respectively representing the confidence level or weight, ARt, of the passenger flow residence time obtained by the two methods_iRepresenting the passenger flow residence time, BRT, of the ith time period of the day, obtained by using WIFI sniffing technology_iAnd the passenger flow residence time of the ith time period of the day is obtained by utilizing the train running time chart, the train full load rate and different passenger flow volumes in the day.

In step S430, the passenger flow residence time and the environmental characteristic information are input to the passenger flow residence time prediction model, and the passenger flow residence time prediction model is trained.

In an example embodiment, the passenger flow residence time prediction model is a deep learning network model or a random forest model or a logistic regression model or a gradient boosting decision tree model. Taking the neural network model as the passenger flow effective residence time prediction model as an example for explanation, external environment feature data such as weather and traffic congestion degree and week, month and year data can be used as input vectors, the corresponding passenger flow residence time is used as a label, and the passenger flow residence time is used as an output vector. Inputting the environment characteristics and the label into the neural network model to train the neural network model, judging whether the neural network model is converged or not through a loss function in the training process, stopping training if the neural network model is converged, and continuing training through adjusting the parameters of the set model if the neural network model is not converged until the model is converged.

Fig. 5 is a flow diagram of a media value assessment for a media to be placed location provided according to some embodiments of the present application.

Referring to fig. 5, in step S510, a media delivery value evaluation model of the target station is constructed according to the passenger flow volume information and the passenger flow residence time information of each historical time period.

In an exemplary embodiment, a media placement value assessment model is established by the following equations (5) and (6), as shown in the calculation equation (5) and the calculation equation (6):

MDV_i＝PL_i×EPST_i×γ_i (5)

wherein, MDV_iRepresenting the value of media delivery, PL, for the ith time period of the day_iIndicating the volume of traffic in the ith time period of the day, EPST_iRepresenting the residence time, gamma, of the passenger flow in the ith time period of the day_iThe effective value rate of the media delivery in the ith time slot in one day is represented, M represents the total time slot of one day, beta represents the media delivery value conversion factor, and MDV represents the media delivery value of one day.

In step S520, the subway media delivery value metric value for a future period of time is obtained by substituting the predicted passenger flow volume and the residence time into the media delivery value calculation model, and the media delivery value for a future period of time is evaluated.

In an exemplary embodiment, the value of media placement at a media location to be placed is evaluated based on the predicted passenger flow and residence time, as shown in the following calculation formula (7) and calculation formula (8):

MDV_{i_predict}＝PL_{i_predict}×EPST_{i_predict}×γ_i (7)

wherein PL_{i_predict}Indicating the predicted traffic volume for the ith time of day, EPST_{i_predict}Effective passenger residence time prediction value, MDV, representing predicted ith time of day_{i_predict}Represents the predicted value, gamma, of media delivery in the ith time period of the day_iRepresenting the effective value rate of media delivery in the ith time period in a day, M representing the total time period divided into M time periods in a day, beta representing the conversion factor of the media delivery value, MDV_predictRepresenting the predicted media delivery value for the day.

Fig. 8 is a schematic block diagram of a data processing apparatus provided in accordance with some embodiments of the present application.

Referring to fig. 8, the data processing apparatus 800 includes: the information acquisition module 810 is configured to acquire historical passenger flow information of a position to be released of a medium and corresponding environment characteristic information, where the historical passenger flow information includes historical passenger flow volume information and historical passenger flow residence time information, and the position to be released of the medium is a predetermined area of a target station; a passenger flow prediction module 820, configured to predict a passenger flow prediction value of a future predetermined time period through a passenger flow prediction model based on the historical passenger flow information and the environmental characteristic information; a residence time prediction module 830, configured to predict a predicted value of the residence time of the passenger flow in the predetermined time period through a passenger flow residence time prediction model based on the historical information of the residence time of the passenger flow and the environmental characteristic information; an evaluation module 840, configured to evaluate a media delivery value of the location to be delivered with media in the predetermined time period based on the passenger flow volume prediction value and the passenger flow residence time prediction value.

In some embodiments of the present application, based on the above scheme, the information obtaining module 810 is configured to:

and/or the presence of a gas in the gas,

In some embodiments of the present application, based on the above scheme, the information obtaining module 810 is further configured to:

In some embodiments of the present application, based on the above solution, referring to fig. 9, the passenger volume prediction module 820 includes:

a first training unit 910, configured to train the passenger flow prediction model based on the historical passenger flow information and the environmental characteristic information, where the environmental characteristic information includes one or more of a date, weather, and a traffic congestion degree corresponding to the historical passenger flow information;

a first characteristic obtaining unit 920, configured to obtain environmental characteristic information corresponding to the predetermined time period;

a passenger flow prediction unit 930, configured to predict the passenger flow prediction value of the predetermined time period through the trained passenger flow prediction model based on the environmental characteristic information corresponding to the predetermined time period.

In some embodiments of the present application, based on the above scheme, the first training unit 910 is configured to:

In some embodiments of the present application, based on the above scheme, the information obtaining module 810 is further configured to include:

In some embodiments of the present application, based on the above scheme, referring to fig. 10, the residence time prediction module 830 includes:

a second training unit 1010, configured to train the passenger flow residence time prediction model based on the historical passenger flow residence time information and the environmental characteristic information, where the environmental characteristic information includes one or more of a date, weather, and a traffic congestion degree corresponding to the historical passenger flow volume information;

a second characteristic obtaining unit 1020, configured to obtain environmental characteristic information corresponding to the predetermined time period;

a residence time prediction unit 1030, configured to predict a predicted value of the passenger flow residence time in the predetermined time period through the trained passenger flow residence time prediction model based on the environment feature information corresponding to the predetermined time period.

In some embodiments of the present application, based on the above scheme, the second training unit 1010 is configured to:

In some embodiments of the present application, based on the above scheme, the evaluation module 840 is configured to:

marking the head image to generate a label corresponding to the customer image;

The data processing device provided by the embodiment of the application can realize each process in the foregoing method embodiments, and achieve the same functions and effects, which are not repeated here.

Fig. 11 shows a schematic structural diagram of a first embodiment of a data processing apparatus according to some embodiments of the present application, and as shown in fig. 11, a data processing apparatus 1100 according to this embodiment may include: a memory 1110 and a processor 1120.

Optionally, the data processing apparatus may further comprise a bus. Wherein, the bus is used for realizing the connection between each element.

The memory 1110 is used for storing computer programs and data, and the processor 1120 calls the computer programs stored in the memory 1110 to execute the technical solution of the data processing method provided by any one of the foregoing method embodiments.

Wherein the memory 1110 is electrically connected to the processor 1120 directly or indirectly, so as to realize data transmission or interaction. For example, the components may be electrically connected to each other via one or more communication buses or signal lines, such as a bus. The memory 1110 stores computer-executable instructions for implementing a data access control method, including at least one software functional module that can be stored in the memory 1110 in the form of software or firmware, and the processor 1120 executes various functional applications and data processing by running the computer programs and modules stored in the memory 1110.

The Memory 1110 may be, but is not limited to, a Random Access Memory (RAM), a Read Only Memory (ROM), a Programmable Read-Only Memory (PROM), an Erasable Read-Only Memory (EPROM), an electrically Erasable Read-Only Memory (EEPROM), and the like. The memory 1110 is used for storing programs, and the processor 1120 executes the programs after receiving the execution instructions. Further, the software programs and modules within the memory 1110 may also include an operating system, which may include various software components and/or drivers for managing system tasks (e.g., memory management, storage device control, power management, etc.), and may communicate with various hardware or software components to provide an operating environment for other software components.

Processor 1120 may be an integrated circuit chip having signal processing capabilities. The Processor 1120 may be a general-purpose Processor, and includes a Central Processing Unit (CPU), a Network Processor (NP), and the like. The various methods, steps, and logic blocks disclosed in the embodiments of the present application may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. It will be appreciated that the configuration of FIG. 11 is merely illustrative and may include more or fewer components than shown in FIG. 11 or have a different configuration than shown in FIG. 11. The components shown in fig. 11 may be implemented in hardware and/or software.

The embodiments of the present application further provide a computer-readable storage medium, on which a computer program is stored, where the computer program, when executed by a processor, can implement the data processing method provided in any of the above method embodiments.

The computer-readable storage medium in this embodiment may be any available medium that can be accessed by a computer or a data storage device such as a server, a data center, etc. that is integrated with one or more available media, and the available media may be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media (e.g., SSDs), etc.

Those of ordinary skill in the art will understand that: all or a portion of the steps of implementing the above-described method embodiments may be performed by hardware associated with program instructions. The program may be stored in a computer-readable storage medium. When executed, the program performs steps comprising the method embodiments described above; and the aforementioned storage medium includes: various media that can store program codes, such as ROM, RAM, magnetic or optical disks.

Finally, it should be noted that: the above embodiments are only used for illustrating the technical solutions of the present application, and not for limiting the same; although the present application has been described in detail with reference to the foregoing embodiments, it should be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; and the modifications or the substitutions do not make the essence of the corresponding technical solutions depart from the scope of the technical solutions of the embodiments of the present application.

Claims

1. A data processing method, comprising:

2. The method of claim 1, wherein obtaining historical passenger flow information for locations where media is to be delivered comprises:

and/or the presence of a gas in the gas,

3. The method of claim 2, further comprising:

4. The method according to any one of claims 1 to 3, wherein the predicting, by a passenger volume prediction model, a passenger volume prediction value for a predetermined period of time in the future based on the historical passenger volume information and the environmental characteristic information comprises:

5. The method of claim 4, wherein training the passenger prediction model based on the historical passenger information and the environmental characteristic information comprises:

6. The method of claim 1, wherein obtaining historical passenger flow residence time information for the location to which the media is to be delivered comprises:

7. The method of claim 6, wherein determining the passenger flow residence time of the media to-be-delivered location history for each time period based on the mobile terminal signal data, the train running time information, the train full load rate, and the passenger flow volume data comprises:

8. The method according to claim 1, 6 or 7, wherein the predicting a passenger flow residence time prediction value for the predetermined time period by a passenger flow residence time prediction model based on the historical passenger flow residence time information and the environmental characteristic information comprises:

9. The method of claim 8, wherein training the passenger flow residence time prediction model based on the historical passenger flow residence time information and the environmental characteristic information comprises:

10. The method of claim 1, wherein evaluating a value of media delivery to the location for media delivery over the predetermined time period based on the passenger flow volume prediction value and the passenger flow residence time prediction value comprises:

11. The method of claim 2, wherein determining a first historical passenger flow for each time period of the historical location for media placement via the passenger flow image comprises:

12. The method of claim 11, further comprising:

marking the head image to generate a label corresponding to the customer image;

13. The method of claim 11, further comprising:

14. A data processing apparatus, comprising:

15. A data processing apparatus, characterized by comprising: a processor and a memory; the memory is used for storing computer programs and data, and the processor calls the computer programs stored in the memory to execute the data processing method of any one of claims 1 to 13.

16. A computer-readable storage medium, characterized in that the computer-readable storage medium comprises a computer program which, when being executed by a processor, is adapted to carry out the data processing method of any one of claims 1 to 13.