Subway data real-time monitoring method and system based on multi-data collection

Abstract

The application discloses a subway data real-time monitoring method and system based on multi-data collection, and belongs to the technical field of subway passenger flow management. The system comprises a data collection module, a flow analysis module, a prediction management module and a visualization module. The data collection module is used for collecting station planning maps, historical logs and station and train information. The flow analysis module divides monitoring areas and blind areas in the station and collects mobile parameters of personnel in real time. Different monitoring areas lock the same personnel according to the mobile parameters to track, record the getting-on and getting-off conditions and the in-and-out station conditions. The prediction management module is used for predicting the getting-off points of personnel in each carriage of the train, and pre-controlling according to the difference between the number of getting-on personnel at each getting-on port and the number of accommodatable personnel in each carriage of the train to be arrived. The flow pressure of each station and the conveying conditions of each train are analyzed, and the congestion degree of the station is reduced by closing the gate entrance or adjusting the train speed. The visualization module is used for real-time display of the running conditions of the station and the train.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of subway passenger flow management, in particular to a subway data real-time monitoring method and system based on multi-data collection. BACKGROUND

[0002] As one of the main public transportation methods in cities, subways often carry a large number of passenger demands. Therefore, studying subway congestion phenomena is of great significance to alleviate urban traffic pressure.

[0003] At present, for the subway congestion problem, the technical means of passenger flow monitoring combined with train interval control are usually used to improve the efficiency of subway operation and reduce the possibility of congestion. However, this method has certain drawbacks, for example: 1. During the peak period, the remaining number of people that each car of the train can accommodate is different, and the number of people waiting at each boarding point of the station where the train is about to arrive is also different. If there is a large difference between the two, it will cause some cars to still be able to accommodate while some people cannot board because of the congestion at their boarding point, resulting in waste of transport resources. 2. For stations with too high passenger flow, the train running interval is usually shortened, but if the number of people getting off the train at these stations is greater than the number of people getting on, it will not only fail to alleviate the pressure of passenger flow at the station, but also increase the pressure of passenger flow at these stations, causing more serious congestion phenomena. 3. The existing technology only monitors the number of people in the station, and cannot grasp the different needs of the people in the station to achieve more efficient monitoring. Therefore, at present, a more intelligent and efficient subway passenger flow monitoring and management technology solution is needed to solve the above problems. SUMMARY

[0004] The purpose of the present application is to provide a subway data real-time monitoring method and system based on multi-data collection to solve the problems raised in the background art.

[0005] In order to solve the above technical problems, the present application provides the following technical solution: a subway data real-time monitoring method based on multi-data collection, which comprises the following steps:

[0006] S100, collect the station planning map and historical log, as well as the number of people in each station and the monitoring picture, and count the number of people in the station through the number of people passing through the gate.

[0007] S200, divide the monitoring area and blind area on the station planning map according to the coverage range of the monitoring picture and identify the target object, lock the same target object in different monitoring areas for tracking, and record the boarding and alighting conditions and the in-and-out station conditions.

[0008] S300, predict the getting-off point of each compartment of each train, and pre-control the difference between the number of passengers getting on at each entrance of each station and the number of passengers that each compartment of the train arriving at the station can accommodate; analyze the flow pressure of each station and the transportation of each train, and reduce the congestion degree of the station by closing the entrance of the gate or adjusting the speed of the train.

[0009] S400, real-time monitor the running status of each station and each train, and display it in the form of dynamic two-dimensional image on the large screen of the subway data center.

[0010] In S1, the station plan is a layout drawing of the inside of the subway station. The history log includes a personnel log, a station log, and a train log. The personnel log is a record of each person's ride, and each ride record includes an entry point and an exit point. The station log is a flow record of each station, and each flow record includes the number of people inside the station at different times, which is counted by counting the number of times the gate is entered and exited. The train log is a speed record of each train, and each speed record includes the speed of the train running on each section at different times. The section refers to the track between two stations. The monitoring picture is a real-time video inside the station.

[0011] In S200, the specific steps are as follows:

[0012] S201, obtain the coverage range of each camera monitoring picture in the station, and mark the location of the gate entrance and exit and the boarding and alighting port in the monitoring area. The area not covered in the station plan is regarded as a blind area, and each range-independent blind area is given a unique code.

[0013] There are overlapping areas between different monitoring areas, and there are only overlapping sections between monitoring areas and blind areas, which are the boundaries of monitoring areas and blind areas.

[0014] S202, a station flow set is established for each station. When a person in the monitoring area provides credential information to activate the gate entrance to enter the station, the credential information is put into the corresponding station flow set. A target detection algorithm is used to identify the person at the activated gate entrance in the corresponding monitoring area and collect feature information, and a target object is generated according to the feature information.

[0015] When the monitoring area includes the gate entrance or the boarding and alighting port, the area outside the station should be shielded to avoid misidentification. The identification and detection of the target object are only for the internal area of the station with the gate entrance or the boarding and alighting port as the boundary. When the person is in the external area of the station or enters the train, the tracking is not continued.

[0016] S203, each monitoring area is only responsible for tracking the target object in the corresponding monitoring picture coverage range, and real-time analysis and recording the speed and direction of each target object as a moving parameter. When the target object leaves the monitoring area, it is regarded as a pending object and the leaving time T is obtained leave , set the prediction duration TL pre , calculate the predicted position LOC pre after the prediction duration TL pre according to the last position and moving parameter of the pending object.

[0017] The prediction duration is set by the staff in advance, and the specific value is referred to the normal walking speed of a person to ensure that the predicted position is not in the corresponding monitoring picture coverage range of the original monitoring area. The calculation of the predicted position is first to calculate the moving distance by the speed in the last collected moving parameter and the prediction duration, and then to find the position corresponding to the moving distance as the predicted position along the direction in the last collected moving parameter from the last position of the pending object.

[0018] S204, find the predicted position LOC pre in the station planning map and judge the area type, and if the type is a monitoring area, obtain the monitoring area code regard the pending object as a target object again, and continue to track according to the feature information provided by the original monitoring area . If the type is a blind area, obtain the blind area code mark all the monitoring areas adjacent to the blind area , set a sampling distance Q, and mark a sampling point every distance Q at the overlapping section of each marked monitoring area and the blind area ; calculate the required time of the pending object to each sampling point according to the predicted position and moving speed, and then add the leaving time T leave and the prediction duration TL pre to obtain the prediction time.

[0019] The sampling distance is set by the staff in advance, and the specific value is referred to the perimeter of the camera monitoring picture coverage range. The larger the value, the fewer the number of sampling points, and the less the consumed computing power; the smaller the value, the more the number of sampling points, and the more accurate the prediction result.

[0020] S205, establish a prediction set for the target object, and put the prediction time of each sampling point and the corresponding monitoring area code into the prediction set in time sequence. According to the order in the prediction set, lock the sampling point position in the prediction time, judge whether the pending object appears, and if it appears, regard the pending object as a target object again, and continue to track according to the feature information provided by the original monitoring area by the monitoring area corresponding to the sampling point where the pending object appears. If it does not appear, continue to lock and judge other monitoring areas according to the order in the prediction set until the pending object is found.

[0021] If the pending object is still not found after the latest predicted time in the prediction set, it indicates that the pending object is still in the blind area and any monitoring area in the prediction set has the possibility of appearing the pending object. All monitoring areas in the prediction set are controlled to lock all sampling point positions at the same time until the pending object is found.

[0022] S206, when the target object in the monitoring area provides the credential information to activate the gate exit to leave the station, the credential information is deleted in the corresponding station flow set. A vehicle flow set is established for each train, and when the target object in the monitoring area enters the train through the boarding gate, the corresponding car number is obtained, and the car number and the feature information of the target object are put into the corresponding vehicle flow set. When the person in the monitoring area enters the station through the alighting gate, the feature information is collected and the corresponding car number is obtained, and the similarity between each feature information in the corresponding vehicle flow set of the stopped train and the collected feature information is calculated, the target object with the highest similarity in the vehicle flow set is selected, and the credential information of the target object is transferred from the station flow set of the corresponding station of the boarding gate to the station flow set of the corresponding station of the alighting gate.

[0023] In S300, the specific steps are as follows:

[0024] S301, when the target object enters the train, it becomes a traveling object, and all ride records in the personnel log are retrieved according to the credential information of the traveling object, ride records with the same entry station as the current entry station are screened out, the number of occurrences of each exit station in these ride records is counted, and the exit station with the highest number of occurrences is selected as the predicted alighting point.

[0025] According to the route and direction of each train, all stations in front are analyzed, if the predicted alighting point does not belong to any of the stations in front, the first station in front that can reach the predicted alighting point through transfer is selected as the new predicted alighting point. If the traveling object does not have a ride record, the terminal station is automatically selected as the predicted alighting point according to the route and direction of the train.

[0026] S302, the train TR u The maximum number of people in each car is reduced by the number of traveling objects in the car to get the actual number of people that can be accommodated; according to the predicted alighting point of the traveling object, the predicted number of people XC n in each car is analyzed when the train arrives at the next station CZ i , and the actual number of people that can be accommodated is added to the predicted number of people XC i to get the predicted number of people that can be accommodated KRS i .

[0027] S303, the number of people in the station CZn The number of target objects at each boarding gate SC i , respectively, subtract the estimated number of people KRS n that can be accommodated in each carriage of the train that will arrive at the station CZ i , get the redundant number of each boarding gate, and warn the boarding gate whose redundant number is greater than zero, prompt the boarding gate that the number of waiting passengers is over limit, and the target objects will transfer to other nearby boarding gates according to the warning prompt.

[0028] The boarding gate refers to the boarding and alighting position corresponding to the train TR n that will arrive at the station CZ u . The number of target objects at each boarding gate is calculated and the difference between the estimated number of people KRS u that can be accommodated in each carriage is calculated to predict whether there is an overload situation in the carriage when the train arrives.

[0029] S304, obtain the flow record of each station, calculate the average value of the number of people in the flow record to obtain the reference number of each station, and divide the current total number of target objects in each station by the corresponding reference number in real time to obtain the flow index. Set the index threshold set {ZY1, ZY2, ZY3}, the elements in the set take increasing values. The station state is lightly crowded when the flow index is greater than ZY1 and less than or equal to ZY2, the station state is moderately crowded when the flow index is greater than ZY2 and less than ZY3, and the station state is severely crowded when the flow index is greater than or equal to ZY3.

[0030] S305, when the train TR u currently stays in the section ST P and arrives at the station CZ n that is lightly crowded, obtain the maximum speed SP max and the minimum speed SP min in all speed records of the train TR u in the section ST P , and the current speed SP now of the train TR u , substitute the adjustment speed SP adj calculated by the formula, and respond to the train driver end, and the driver adjusts the speed according to the adjustment speed. The adjustment speed calculation formula is as follows:

[0031]

[0032] In the formula, ZY tru is the flow index of the train TR u .

[0033] When the total number of passengers boarding at the station is greater than the total estimated number of people that the train can accommodate, the total estimated number of people that the train can accommodate is automatically selected as the total number of passengers boarding. The difference between the total number of passengers boarding and the total estimated number of passengers alighting is calculated to determine whether the number of people at the station increases or decreases after the train arrives. If it increases, the train speed is reduced to give the station a certain amount of time to relieve the passenger flow, and if it decreases, the train speed is increased to quickly relieve the passenger flow at the station.

[0034] S306、When the train TR u The current section ST P Arriving at the station CZ n When the station is moderately congested or heavily congested, the flow index is first taken as ZY2, and then substituted into step S305 to calculate the adjusted speed and respond. Then the station state is analyzed, and if the station is heavily congested, all gate entrances are closed. If the station is moderately congested, the station CZ n The number of gate entrances S fz and the flow index ZY tru are substituted into the formula, and the number of closed gate entrances is calculated by taking the integer part downward, and the corresponding number of gate entrances is automatically closed; the calculation formula is as follows:

[0035]

[0036] In the formula, S close is the number of closed gate entrances.

[0037] In S400, the flow index change of each station and the running speed change of each train are monitored in real time, and the dynamic two-dimensional image is displayed in real time on the metro data center visual screen, and the passenger records, station flow records and train speed records are collected and put into the historical log.

[0038] The metro data real-time monitoring system based on multi-data collection includes a data acquisition module, a flow analysis module, a prediction management module and a visualization module.

[0039] The data acquisition module is used to acquire station planning maps, historical logs, and station and train information. The flow analysis module divides the monitoring area and the blind area in the station and acquires the movement parameters of the personnel in real time, and different monitoring areas lock the same personnel for tracking according to the movement parameters, and record the boarding and alighting conditions and the in-and-out station conditions. The prediction management module is used to predict the alighting points of the personnel in each train compartment, and to pre-control according to the difference between the number of passengers boarding at each boarding gate and the number of people that each train compartment of the train to be arrived can accommodate. The flow pressure of each station and the transportation of each train are analyzed, and the congestion degree of the station is reduced by closing the gate entrances or adjusting the train speed; the visualization module is used to display the running conditions of the station and the train in real time.

[0040] The data collection module comprises a station information collection unit, a train information collection unit and a historical log collection unit.

[0041] The station information collection unit is configured to collect a station plan and monitoring pictures and the number of personnel inside the station. The train information collection unit is configured to collect attributes and parameters of each train, the attributes including train numbers and car numbers, and the parameters referring to running speeds. The historical log collection unit is configured to collect personnel logs, station logs and train logs. The personnel logs refer to ride records of each personnel, and each ride record includes an entry point and an exit point. The station logs refer to flow records of each station, and each flow record includes the number of personnel inside the station at different times. The train logs refer to speed records of each train, and each speed record includes speeds on each section at different times, the section referring to a track between two stations.

[0042] The flow analysis module comprises a region division unit and a personnel tracking unit.

[0043] The region division unit is configured to divide monitoring regions and blind regions. In the station plan, a region where a monitoring picture of each camera covers is regarded as a monitoring region, and a region which is not covered in the station plan and is independent in range is regarded as a blind region.

[0044] The personnel tracking unit is configured to track target objects.

[0045] Firstly, a target detection algorithm is used to identify personnel entering the station through a gate entrance in a monitoring region and collect feature information, and a target object is generated according to the feature information. Each monitoring region only tracks target objects in the range covered by the respective monitoring picture, and the moving parameters of the target objects are analyzed and recorded. Secondly, when a target object leaves a monitoring region, a predicted position is analyzed and predicted according to the moving parameters. If the predicted position is in a monitoring region, the target object is tracked by the monitoring region where the predicted position is located. If the predicted position is in a blind region, sampling points are marked at overlapping sections of all adjacent monitoring regions of the blind region, the predicted time of the target object to each sampling point is calculated, and each monitoring region is notified to lock the sampling point position for target detection in the order of the predicted time, until the target object leaving the monitoring region is found and tracked. Finally, when a target object leaves the station through a gate exit, the tracking is cancelled. When a personnel enters a train through a boarding gate, the train number and car number are recorded. When a personnel enters the station through an alighting gate, feature information and the train number and car number are collected, similarity calculation is performed on the feature information and feature information corresponding to the same train number and the same car number recorded before, and a target object is generated by selecting feature information with the highest similarity.

[0046] The prediction management module comprises a congestion prediction unit and a debugging control unit.

[0047] The congestion prediction unit is used to predict the congestion situation of the station. First, the target object on the train is taken as a traveling object, the predicted alighting point of the traveling object is analyzed through the boarding record, the expected accommodated number of each compartment at the next station is counted according to the predicted alighting point of each traveling object, the difference between the number of target objects at each boarding gate of the next station and the expected accommodated number of the corresponding compartment is calculated respectively, and whether the boarding gate number exceeds the limit is judged. Then, the reference number of each station is calculated according to the number of personnel in the flow record, the total number of target objects in each station is divided by the corresponding reference number to obtain the flow index, and the congestion state of the station is divided according to the flow index. The congestion state includes light congestion, medium congestion and heavy congestion.

[0048] The debugging control unit is used to debug control the congestion situation. The boarding gate with the number of waiting passengers exceeding the limit is warned, and the target object is prompted to transfer to other nearby boarding gates. When the congestion state of the station is light, only the speed of the train about to arrive is adjusted to reduce the flow index. When the congestion state of the station is medium or heavy, the speed of the train about to arrive is adjusted and a certain number of gate entrances are closed to reduce the flow index.

[0049] The visualization module displays the flow index of each station and the traveling speed of each train in real time through the subway data center visualization large screen, and puts the boarding record of personnel, the flow record of the station and the speed record of the train into the historical log.

[0050] Compared with the prior art, the application has the following beneficial effects:

[0051] 1. Efficient personnel tracking: In the present application, different monitoring areas are used to identify and track the same personnel, the predicted position of the personnel is analyzed through the position and movement parameter of the personnel out of the picture, and the predicted time of the personnel to arrive at other monitoring areas is calculated according to the predicted position to lock and identify in turn. Compared with the mechanical target detection in the traditional technology, it is more efficient and saves computing resources.

[0052] 2. Accurate waiting planning: In the present application, the boarding of personnel is tracked and the alighting point is predicted according to the historical record, the accommodated number of each compartment of each train at the next station is calculated, and the waiting number of each boarding point at the next station is dynamically planned according to the difference between the boarding number of each boarding point and the accommodated number of the corresponding compartment. Compared with the traditional waiting mechanism, it is more accurate.

[0053] 3. Intelligent flow regulation: The application judges the influence of the difference between the boarding number and the alighting number of each train on the congestion degree of the station, and delays the increase of congestion or increases the congestion relief speed by adjusting the speed of the train about to arrive. The flow regulation is realized by closing the gate entrance, and the station flow regulation method is more intelligent than the traditional technology.

[0054] Compared with the prior art, the subway passenger flow control efficiency can be improved by the personnel tracking, accurate waiting planning and intelligent flow regulation. BRIEF DESCRIPTION OF DRAWINGS

[0055] The accompanying drawings are included to provide a further understanding of the application and are incorporated in and constitute a part of this specification, illustrate embodiments of the application and are used to explain the application, but do not constitute a limitation of the application. In the drawings:

[0056] Figure 1 is a flow diagram of the subway data real-time monitoring method based on multi-data collection of the application;

[0057] Figure 2 is a structural diagram of the subway data real-time monitoring system based on multi-data collection of the application. DETAILED DESCRIPTION

[0058] The technical solutions in the embodiments of the application will be clearly and completely described below with reference to the drawings in the embodiments of the application. Obviously, the described embodiments are only part of the embodiments of the application, rather than all the embodiments of the application. Based on the embodiments in the application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the application.

[0059] Please refer to Figure 1 The application provides a subway data real-time monitoring method based on multi-data collection, which comprises the following steps:

[0060] S100, collecting a station planning map and historical logs, and the number of personnel in each station and monitoring pictures, and counting the number of personnel in the station by the number of people passing through the gate.

[0061] S200, dividing monitoring areas and blind areas on the station planning map according to the coverage range of the monitoring pictures and identifying target objects, locking the same target object in different monitoring areas for tracking, and recording the boarding and alighting conditions and the in-and-out station conditions.

[0062] S300, predicting the alighting points of personnel in each carriage of each train, pre-controlling according to the difference between the number of boarding personnel at each boarding gate of each station and the number of people that can be accommodated in each carriage of the train arriving at the station, analyzing the flow pressure of each station and the conveying conditions of each train, and reducing the congestion degree of the station by closing the gate entrance or adjusting the train speed.

[0063] S400, real-time monitoring the running conditions of each station and each train, and displaying the conditions on the visual large screen of the subway data center in the form of dynamic two-dimensional images.

[0064] In S1, the station plan is a layout drawing of the inside of a subway station. The history log includes a personnel log, a station log, and a train log. The personnel log is a record of each person's ride, and each ride record includes an entry point and an exit point. The station log is a record of the flow of each station, and each flow record includes the number of people inside the station at different times, which is counted by counting the number of times the gate is entered and exited. The train log is a record of the speed of each train, and each speed record includes the speed of the train traveling on each section at different times. A section refers to the track between two stations. The monitoring picture is a real-time video of the inside of the station.

[0065] In S200, the specific steps are as follows:

[0066] S201, obtain the coverage range of each camera monitoring picture inside the station, and mark the area covered by the coverage range as a monitoring area on the station plan. Each monitoring area of a camera is given a unique code. Mark the locations of the gate entrances and exits and the boarding and alighting locations in the monitoring area. The area not covered in the station plan is marked as a blind area, and each independent blind area is given a unique code.

[0067] There are overlapping areas between different monitoring areas, and there are only overlapping sections between monitoring areas and blind areas. An overlapping section is the boundary line between a monitoring area and a blind area.

[0068] S202, a station flow set is established for each station. When a person in a monitoring area provides credential information to activate the gate entrance to enter the station, the credential information is placed in the corresponding station flow set. A target detection algorithm is used to identify the person at the activated gate entrance in the corresponding monitoring area and collect feature information, and a target object is generated based on the feature information.

[0069] When a monitoring area includes a gate entrance or exit or a boarding or alighting location, the area outside the station should be shielded to avoid misidentification. The identification and detection of a target object are only for the area inside the station that is divided by the gate entrance or exit or the boarding or alighting location. When a person is in the area outside the station or enters a train, the tracking is not continued.

[0070] S203, each monitoring area is only responsible for tracking target objects within the coverage range of the corresponding monitoring picture. The speed and direction of each target object are analyzed and recorded in real time as movement parameters. When a target object leaves a monitoring area, it becomes a pending object and the time T of leaving is obtained. leave Set a prediction duration TL pre The predicted location LOC pre after the prediction duration TL pre is calculated based on the last location of the pending object and the movement parameters.

[0071] The prediction duration is set by the staff in advance, and the specific value is referred to the normal walking speed of the person, to ensure that the predicted position is not in the coverage range of the original monitoring area corresponding monitoring picture. The calculation of the predicted position is first to calculate the moving distance by the speed in the last collected moving parameter and the prediction duration, and then to find the position corresponding to the moving distance as the predicted position along the direction in the last collected moving parameter from the last position of the object to be determined.

[0072] S204, find the predicted position LOC in the station planning map pre and judge the type of the area, if the type is a monitoring area, get the monitoring area code regard the object to be determined as the target object again, and track the object to be determined by the monitoring area according to the feature information provided by the original monitoring area. If the type is a blind area, get the blind area code mark all the monitoring areas adjacent to the blind area , set the sampling distance Q, and mark a sampling point every distance Q at the overlapping section of each marked monitoring area and blind area ; calculate the required duration of the object to be determined to each sampling point according to the predicted position and moving speed, and then add the escape time T leave and the prediction duration TL pre to get the prediction time.

[0073] The sampling distance is set by the staff in advance, and the specific value is referred to the perimeter of the camera monitoring picture coverage range. The larger the value, the fewer the number of sampling points, and the less the computing power consumed; the smaller the value, the more the number of sampling points, and the more accurate the prediction result.

[0074] S205, establish a prediction set for the target object, and put the prediction time of each sampling point and the corresponding monitoring area code into the prediction set in time sequence. According to the order in the prediction set, lock the sampling point position in the corresponding monitoring area within the prediction time, judge whether the object to be determined appears, if it appears, regard the object to be determined as the target object again, and track the object to be determined by the monitoring area corresponding to the sampling point where the object to be determined appears according to the feature information provided by the original monitoring area. If it does not appear, continue to lock and judge other monitoring areas according to the order in the prediction set until the object to be determined is found.

[0075] If the object to be determined is still not found after the latest prediction time in the prediction set, it means that the object to be determined is still in the blind area and any monitoring area in the prediction set may appear the object to be determined at any time. Lock all the sampling point positions of all the monitoring areas in the prediction set at the same time until the object to be determined is found.

[0076] S206. When a target object within the monitoring area activates the gate exit and leaves the station using credential information, the credential information is deleted from the corresponding station flow set. A vehicle flow set is established for each train. When a target object within the monitoring area enters a train through the boarding gate, the corresponding carriage number is obtained, and the carriage number and the target object's feature information are added to the corresponding vehicle flow set. When personnel within the monitoring area enter the station through the alighting gate, feature information is collected, and the corresponding carriage number is obtained. The similarity of each feature information under that carriage number in the vehicle flow set corresponding to the stopped train is calculated with the collected feature information. The feature information with the highest similarity in the vehicle flow set is selected to generate a target object, and the target object's credential information is transferred from the station flow set corresponding to the boarding gate to the station flow set corresponding to the alighting gate.

[0077] In S300, the specific steps are as follows:

[0078] S301. Once the target object enters the train, it is considered a travel object. Based on the travel object's credentials, all travel records are retrieved from the personnel log. Travel records with the same entry station as the current entry station are selected. The number of times each exit station appears in these travel records is counted, and the exit station with the highest number of appearances is selected as the expected disembarkation point.

[0079] Based on the analysis of all stations ahead of each train's route and direction, if the expected drop-off point is not among any of the stations ahead, the first station among all stations ahead that can be reached by transfer is automatically selected as the new expected drop-off point. If the passenger has no travel history, the final destination station is automatically selected as the expected drop-off point based on the train's route and direction.

[0080] S302, Train TR u The actual capacity of each carriage is obtained by subtracting the number of passengers traveling in that carriage from its maximum capacity; the train's journey to the next station CZ is analyzed based on the passengers' expected disembarkation points. n The estimated number of passengers disembarking in each carriage XC i Add the actual capacity to the estimated number of passengers disembarking to XC i The estimated capacity of each carriage (KRS) i .

[0081] S303, Statistical Station CZ n The number of target objects at each boarding gate (SC) i Subtract the CZ station you are about to arrive at from each station. n The estimated number of passengers per carriage of the train (KRS) iThe system obtains the redundant number of passengers at each boarding gate, issues an early warning for boarding gates with a redundant number greater than zero, indicating that the number of passengers waiting at that boarding gate has exceeded the limit, and the target passenger is transferred to another nearby boarding gate according to the early warning.

[0082] The boarding gate refers to the station CZ where you are about to arrive. n Train TR u The corresponding boarding and alighting positions are fixed, with each boarding gate corresponding to one carriage. The difference between the number of target objects at each boarding gate and the estimated capacity of the corresponding carriage is calculated to predict whether the carriage will be overloaded when the train arrives.

[0083] S304. Obtain the traffic flow records for each station. Calculate the average number of people at each station based on the number of people recorded in the traffic flow records to obtain the baseline number of people. Divide the current total number of target objects at each station by the corresponding baseline number of people in real time to obtain the traffic flow index. Set an index threshold set {ZY1, ZY2, ZY3}, where the values ​​of elements in this set increase sequentially. Stations with a traffic flow index greater than ZY1 and less than or equal to ZY2 are classified as lightly congested; stations with a traffic flow index greater than ZY2 and less than ZY3 are classified as moderately congested; and stations with a traffic flow index greater than or equal to ZY3 are classified as heavily congested.

[0084] S305, when train TR u The current section of road ST P Arrival Station CZ n To obtain train TR during periods of light congestion u On section ST P The highest speed SP among all speed records max and minimum speed SP min and train TR u Current driving speed SP now Substituting into the formula, the adjustment speed SP is calculated. adj The adjusted speed response is sent to the train driver's end, and the driver adjusts the speed accordingly. The formula for calculating the adjusted speed is as follows:

[0085]

[0086] In the formula, ZY tru For train TR u Traffic index.

[0087] When the total number of passengers boarding at the station exceeds the train's total estimated capacity, the train's total estimated capacity is automatically selected as the total number of passengers boarding. The difference between the total number of passengers boarding and the total estimated number of passengers disembarking is calculated to determine whether the number of passengers at the station increases or decreases after the train arrives. If the number increases, the train speed is reduced to allow the station some time to alleviate the passenger flow; if the number decreases, the train speed is increased to alleviate the passenger flow at the station as quickly as possible.

[0088] S306、When the train TR u The current section ST P The upper arrival station CZ n If the station is moderately congested or heavily congested, the flow index is first taken as ZY2, and then substituted into the step S305 to calculate the adjusted speed and respond. Then, the station state is analyzed, and if the station is heavily congested, all gate entrances are closed. If the station is moderately congested, the station CZ n The number of gate entrances S fz The flow index ZY tru is substituted into the formula, and the number of closed gate entrances is calculated by taking the integer part downward, and the corresponding number of gate entrances is automatically closed; the calculation formula is as follows:

[0089]

[0090] In the formula, S close is the number of closed gate entrances.

[0091] In S400, the flow index change of each station and the running speed change of each train are monitored in real time, and are displayed in real time on the metro data center visual screen in the form of a dynamic two-dimensional image, and the personnel ride records, station flow records and train speed records are collected and put into the history log.

[0092] Please refer to Figure 2 The present application provides a subway data real-time monitoring system based on multi-data collection, which comprises a data acquisition module, a flow analysis module, a prediction management module and a visualization module.

[0093] The data acquisition module is used to acquire station planning maps, history logs and station and train information. The flow analysis module divides the monitoring area and blind area in the station and acquires the movement parameters of personnel in real time, and different monitoring areas lock the same personnel for tracking according to the movement parameters, and record the boarding and alighting conditions and the in-and-out station conditions. The prediction management module is used to predict the alighting points of personnel in each carriage of the train, and pre-control according to the difference between the number of boarding personnel at each boarding gate and the number of accommodated personnel in each carriage of the train to be arrived. The flow pressure of each station and the transportation conditions of each train are analyzed, and the congestion degree of the station is reduced by closing the gate entrances or adjusting the train speed; the visualization module is used to display the running conditions of the station and the train in real time.

[0094] The data acquisition module comprises a station information acquisition unit, a train information acquisition unit and a history log acquisition unit.

[0095] The station information acquisition unit is configured to acquire a station layout and a monitoring picture and a number of personnel in the station. The train information acquisition unit is configured to acquire attributes and parameters of each train, the attributes including a train number and a carriage number, and the parameters being a running speed. The historical log acquisition unit is configured to acquire personnel logs, station logs and train logs. The personnel logs are records of each personnel, each record including an entry point and an exit point. The station logs are records of a flow of each station, each record including a number of personnel in the station at different times. The train logs are records of a speed of each train, each record including a speed of the train on each section at different times, the section being a track between two stations.

[0096] The flow analysis module includes a region division unit and a personnel tracking unit.

[0097] The region division unit is configured to divide a monitoring region and a blind region. In the station layout, a region covered by a monitoring picture of each camera is regarded as the monitoring region, and a region not covered by the station layout and independent in range is regarded as the blind region.

[0098] The personnel tracking unit is configured to track a target object.

[0099] Firstly, a target detection algorithm is used to identify personnel entering the station through a gate entrance in the monitoring region and to acquire feature information, and a target object is generated according to the feature information, each monitoring region only tracks the target object in the range covered by the monitoring picture, and moving parameters of the target object are analyzed and recorded. Secondly, when the target object leaves the monitoring region, a predicted position is analyzed and predicted according to the moving parameters, if the predicted position is in the monitoring region, the target object is tracked by the monitoring region where the predicted position is located; if the predicted position is in the blind region, a sampling point is marked at an overlapping section of all adjacent monitoring regions of the blind region, a predicted time of the target object to each sampling point is calculated, and each monitoring region is notified to lock the sampling point position for target detection in order according to the predicted time, until the target object leaving the monitoring region is found and tracked. Finally, when the target object leaves the station through a gate exit, the tracking is cancelled; when the target object enters a train through a boarding gate, a train number and a carriage number are recorded; when the personnel enter the station through an alighting gate, feature information and the train number and the carriage number are acquired, similarity calculation is performed on the feature information and feature information corresponding to the same train number and the same carriage number recorded before, and a target object is generated by selecting feature information with the highest similarity.

[0100] The prediction management module includes a congestion prediction unit and a debugging control unit.

[0101] The congestion prediction unit is used to predict station congestion. First, the target passengers on the train are considered as the travel targets. The predicted alighting points of these targets are analyzed through passenger records. Based on the predicted alighting points of each travel target, the estimated capacity of each carriage at the next station is calculated. The difference between the number of target passengers at each boarding gate at the next station and the estimated capacity of the corresponding carriage is calculated to determine if the number of people waiting at the boarding gate exceeds the limit. Next, the baseline number of passengers at each station is calculated based on the number of people in the flow records. The current total number of target passengers at each station is divided by the corresponding baseline number to obtain the flow index. Based on the flow index, the station is classified into congestion states: mild congestion, moderate congestion, and severe congestion.

[0102] The commissioning control unit is used to adjust and control congestion conditions. It issues warnings for boarding gates with excessive passenger numbers, prompting passengers to move to other nearby boarding gates. When the station is lightly congested, the flow rate is reduced simply by adjusting the speed of arriving trains. When the station is moderately or heavily congested, the flow rate is reduced by adjusting the speed of arriving trains and closing a certain number of turnstile entrances.

[0103] The visualization module displays the flow index of each station and the speed of each train in real time through the large visualization screen of the subway data center, and puts the passenger travel records, station flow records and train speed records into the historical log.

[0104] Example 1:

[0105] Assuming train TR js The current section of road ST js Arrival Station CZ js For mild congestion, station CZ js The traffic flow index is 1.3, and the index threshold set is {1.2, 1.4, 1.6}; through analysis of train TR js On section ST js After reviewing all speed records, we found that the maximum speed was 75 km / s and the minimum speed was 45 km / s. (Train TR) js Current speed is 60 km / s, predicted when it arrives at station CZ js The total number of passengers boarding is greater than the total number of passengers disembarking. Substituting these values ​​into the formula, we can calculate the train's TR. js Adjustment speed:

[0106] Train TR js Adjustment speed:

[0107] It is to be noted that, in the present text, relational terms such as first and second and the like can be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises", "comprising", or any other variations thereof, are intended to cover a non-exclusive inclusion such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but can include other elements not expressly listed or inherent to such process, method, article, or apparatus.

[0108] Finally, it should be noted that the above-mentioned only constitutes the preferred embodiments of the present application and is not intended to limit the present application. Although the present application has been described in detail with reference to the foregoing embodiments, it will be apparent to those skilled in the art that modifications, equivalent replacements, improvements and the like of the technical solutions described in the foregoing embodiments can still be made. Any modifications, equivalent replacements, improvements and the like made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for real-time monitoring of subway data based on multiple data collection methods, characterized in that: The monitoring method includes the following steps: S100: Collect station planning maps and historical logs, as well as the number of people and monitoring footage at each station, and count the number of people entering and exiting the station through turnstiles; S200: Divide the station planning map into monitoring areas and blind spots according to the coverage of the monitoring screen and identify target objects. Different monitoring areas lock the same target object for tracking and record the boarding and alighting and entering and leaving the station. S300: Predict the disembarkation points of passengers in each carriage of each train, and make advance adjustments based on the difference between the number of passengers boarding at each boarding gate of each station and the capacity of each carriage of the train that is about to arrive at the station; analyze the flow pressure of each station and the transport status of each train, and reduce the congestion of the station by closing the gate entrance or adjusting the speed of the train. S400 monitors the operation status of each station and each train in real time and displays it on the large visualization screen of the subway data center in the form of dynamic two-dimensional images; In S300, the specific steps are as follows: S301. Once the target object enters the train, it is treated as the travel object. Based on the travel object's credentials, all travel records are retrieved from the personnel log. Travel records with the same entry station as the current entry station are filtered out. The number of times each exit station appears in these travel records is counted, and the exit station with the highest number of appearances is selected as the expected disembarkation point. S302, the train The actual capacity of each carriage is obtained by subtracting the number of passengers traveling in that carriage from its maximum capacity; the train's journey to the next station is then analyzed based on the passengers' expected disembarkation points. The estimated number of passengers disembarking in each carriage at that time Add the actual capacity to the estimated number of passengers getting off the bus. Get the estimated capacity of each carriage ; S303, Statistics Station Number of target objects at each boarding gate Subtract the station you are about to arrive at from each station. The estimated number of passengers per carriage of the train The system obtains the redundant number of passengers at each boarding gate, issues an early warning for boarding gates with a redundant number greater than zero, indicating that the number of passengers waiting at that boarding gate has exceeded the limit, and the target person is transferred to another nearby boarding gate according to the early warning. S304. Obtain the traffic flow records for each station, calculate the average number of people in the traffic flow records to obtain the baseline number of people at each station, and divide the current total number of target objects at each station by the corresponding baseline number of people in real time to obtain the traffic flow index; set the index threshold set. The elements in this set take values ​​in ascending order; the flow index is greater than... and less than or equal to The station is in a state of mild congestion, with a flow index greater than [missing information]. and less than The station is in a state of moderate congestion, with a flow index greater than or equal to [missing information]. The station is in a state of severe congestion; S305, when the train Current location Up to the station To obtain train tickets during periods of light congestion On the road section The highest speed among all speed records and minimum speed and trains Current driving speed Substituting into the formula, the adjustment speed is calculated. The adjusted speed response is sent to the train driver's end, and the driver adjusts the speed accordingly. The formula for calculating the adjusted speed is as follows: In the formula, For train Traffic index; S306, when the train Current location Up to the station When the congestion level is moderate or heavy, first determine the flow index value. Substitute the input to step S305 to calculate and adjust the speed and respond; then analyze the station status. If the station is severely congested, close all turnstile entrances; if the station is moderately congested, obtain the station's status. Number of turnstile entrances With traffic index Substitute all the components into the formula, round down, and calculate the number of gate entrances to be closed. The corresponding number of gate entrances will then be automatically closed. The calculation formula is as follows: In the formula, The number of gate entrances to be closed.

2. The real-time monitoring method for subway data based on multiple data collection as described in claim 1, characterized in that: In S1, the station planning map refers to the layout drawing of the subway station; the historical log includes the personnel log, station log, and train log; the personnel log refers to the travel record of each person, and each travel record includes the entry and exit points of the station; the station log refers to the flow record of each station, and each flow record includes the number of people inside the station at different times, which is counted by counting the number of times the turnstiles enter and exit; the train log refers to the speed record of each train, and each speed record includes the speed of the train traveling on each section at different times, and the section refers to the track between two stations; the monitoring screen refers to the real-time video inside the station.

3. The real-time monitoring method for subway data based on multiple data collection as described in claim 2, characterized in that: In S200, the specific steps are as follows: S201. Obtain the coverage area of ​​each camera's monitoring image within the station, and designate the area where the coverage area is located as the monitoring area in the station planning map. Generate a unique code for the monitoring area of ​​each camera. Mark the locations of turnstile entrances and exits and boarding / alighting points within the monitoring area, and designate areas not covered in the station planning map as blind spots, generating a unique code for each independent blind spot. S202. Each station establishes a station traffic set. When a person in the monitoring area provides credentials to activate the gate entrance and enter the station, the credentials are placed into the corresponding station traffic set. The target detection algorithm is used to identify the person at the activated gate entrance in the corresponding monitoring area and collect feature information. The target object is generated based on the feature information. S203. Each monitoring area is only responsible for tracking target objects within the coverage area of ​​the corresponding monitoring screen, and analyzing and recording the speed and direction of each target object in real time as movement parameters. When a target object leaves the monitoring area, it is designated as a pending object, and the time of departure is recorded. Set the prediction duration The prediction duration is calculated based on the last location and movement parameters of the object to be determined. Predicted position after ; S204. Locate the predicted location on the station planning map. It also determines the type of the area; if the type is a monitored area, it obtains the monitored area code. The pending object is reclassified as the target object, and the monitoring area... Continue tracking based on the feature information provided by the original monitoring area; if the type is a blind spot, obtain the blind spot code. Mark all areas with blind spots Set sampling distance for adjacent monitoring areas In each marked monitoring area and blind spot At intervals of overlapping road sections Mark a sampling point; calculate the time required for the object to reach each sampling point based on the predicted position and moving speed, and then add the detachment time. and predicted duration Obtain the predicted time; S205. Establish a prediction set for the target object, and put the prediction time and corresponding monitoring area code of each sampling point into the prediction set in chronological order; notify the corresponding monitoring area to lock the sampling point position in the prediction time in the order of the prediction set, and determine whether a pending object appears. If it appears, the pending object is re-designated as the target object, and the monitoring area corresponding to the sampling point where the pending object appears continues to track it based on the feature information provided by the original monitoring area. If it does not appear, continue to notify other monitoring areas to lock and judge according to the order in the prediction set until the pending object is found; S206. When a target object in the monitoring area provides credential information to activate the gate exit and leave the station, the credential information is deleted from the corresponding station flow set; a vehicle flow set is established for each train. When a target object in the monitoring area enters the train through the boarding gate, the corresponding carriage number is obtained, and the carriage number and the target object's feature information are put into the corresponding vehicle flow set. When personnel in the monitoring area enter the station through the exit, feature information is collected and the corresponding carriage number is obtained. The similarity of each feature information under the carriage number in the traffic flow set corresponding to the stopped train is calculated with the collected feature information. The feature information with the highest similarity in the traffic flow set is selected to generate the target object. The credential information of the target object is transferred from the station traffic flow set corresponding to the boarding gate to the station traffic flow set corresponding to the alighting gate.

4. The real-time monitoring method for subway data based on multiple data collection as described in claim 3, characterized in that: In the S400 system, the changes in the flow index of each station and the speed of each train are monitored in real time. These changes are displayed in real time on the large visualization screen of the metro data center in the form of dynamic two-dimensional images. Passenger travel records, station flow records, and train speed records are collected and put into the historical log.

5. A real-time subway data monitoring system based on multiple data collection, wherein the system is applied to the real-time subway data monitoring method based on multiple data collection as described in any one of claims 1-4, characterized in that: The system includes a data acquisition module, a traffic analysis module, a forecasting management module, and a visualization module; The data acquisition module collects station planning maps, historical logs, and station and train information; the traffic analysis module divides the station into monitoring zones and blind zones and collects personnel movement parameters in real time. Different monitoring zones track the same person based on movement parameters, recording boarding and alighting and station entry and exit; the predictive management module predicts the alighting points of passengers in each train carriage and makes advance adjustments based on the difference between the number of passengers boarding at each boarding gate and the capacity of each carriage of the arriving train; it analyzes the traffic pressure of each station and the transport status of each train, reducing station congestion by closing gate entrances or adjusting train speed; the visualization module displays the real-time operation status of stations and trains.

6. The real-time monitoring system for subway data based on multiple data collection as described in claim 5, characterized in that: The data acquisition module includes a station information acquisition unit, a train information acquisition unit, and a historical log acquisition unit; The station information collection unit is used to collect station planning maps, as well as monitoring images and personnel numbers inside the station; the train information collection unit is used to collect the attributes and parameters of each train, including train number and carriage number, and the parameters refer to the travel speed; the historical log collection unit is used to collect personnel logs, station logs, and train logs.

7. The real-time monitoring system for subway data based on multiple data collection as described in claim 6, characterized in that: The traffic analysis module includes a region segmentation unit and a personnel tracking unit; The area division unit is used to divide the monitoring area and the blind spot; in the station planning map, the area covered by each camera's monitoring image is designated as the monitoring area, and the area not covered and with an independent range in the station planning map is designated as the blind spot. The personnel tracking unit is used to track target objects; First, a target detection algorithm is used to identify personnel entering the station through the turnstile entrance in the monitored area and collect their feature information. Target objects are generated based on this feature information. Each monitoring area only tracks target objects within its own monitoring screen coverage area, analyzing and recording the movement parameters of these target objects. Second, when a target object leaves the monitored area, its predicted location is analyzed based on the movement parameters. If the predicted location is within the monitored area, the monitoring area where the predicted location is located continues tracking. If the predicted location is in a blind spot, sampling points are marked on all overlapping sections of monitoring areas adjacent to the blind spot. The predicted time from the target object to each sampling point is calculated, and each monitoring area is notified sequentially according to the predicted time to lock the sampling point location for target detection until a target object leaving the monitored area is found and tracking continues. Finally, tracking is canceled when a target object leaves the station through the turnstile exit. When a target object enters a train through the boarding gate, the train number and carriage number are recorded. When a person enters the station through the alighting gate, feature information, train number, and carriage number are collected. The feature information is compared with previously recorded feature information corresponding to the same train number and carriage number, and the feature information with the highest similarity is selected to generate the target object.

8. The real-time monitoring system for subway data based on multiple data collection as described in claim 7, characterized in that: The predictive management module includes a congestion prediction unit and a debugging and control unit; The congestion prediction unit is used to predict station congestion. First, the target objects on the train are taken as the travel objects. The predicted alighting points of the travel objects are analyzed through the passenger records. Based on the predicted alighting points of each travel object, the estimated number of passengers in each carriage at the next station is calculated. The difference between the number of target objects at each boarding gate of the next station and the estimated number of passengers in the corresponding carriage is calculated to determine whether the number of people waiting at the boarding gate exceeds the limit. Then, the baseline number of passengers at each station is calculated based on the number of people in the flow records. The current total number of target objects at each station is divided by the corresponding baseline number to obtain the flow index. The congestion status of the station is divided according to the flow index. The congestion status includes mild congestion, moderate congestion and severe congestion. The commissioning control unit is used to adjust and control congestion; it provides early warnings for boarding gates with excessive waiting numbers, prompting passengers to move to other nearby boarding gates; when the station is slightly congested, it reduces the flow index by adjusting the speed of the approaching train; when the station is moderately or heavily congested, it reduces the flow index by adjusting the speed of the approaching train and closing a certain number of turnstile entrances.

9. The real-time monitoring system for subway data based on multiple data collection as described in claim 8, characterized in that: The visualization module displays the flow index of each station and the speed of each train in real time through the large visualization screen of the subway data center, and puts the passenger travel records, station flow records and train speed records into the historical log.

Images