Subway station hall data information analysis method and system based on artificial intelligence

Abstract

The application relates to the field of subway traffic, in particular to a subway station hall data information analysis method and system based on artificial intelligence, which comprises an image recognition module, a personnel dispersion module, a traffic flow prediction module, a boarding interval module and a gate shunting module; the image recognition module is used for acquiring personnel information in a carriage; the personnel dispersion module is used for calculating personnel dispersion; the traffic flow prediction module is used for predicting the traffic flow arriving at a station; the boarding interval module is used for calculating expected traffic flow and dividing boarding intervals; and the gate shunting module is used for adjusting the position of a gate. The application can identify the travel habits and demands of passengers, reasonably allocate traffic flow, alleviate the congestion phenomenon in a station hall, make passengers obtain better platform passing space, reduce the distance of passengers getting off, shorten the time cost of passengers getting on and off, and improve the overall passing efficiency and operation effect of a subway line.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of subway transportation, in particular to a subway station hall data information analysis method and system based on artificial intelligence. BACKGROUND

[0002] The subway station hall is a large space in the subway station, usually located at the entrance of the subway station, and is provided with an entrance and exit passage, a security check desk, an automatic ticket vending machine, a ticket gate and an escalator, etc., to facilitate passengers to take the train. The traditional subway gate is in parallel rows, and passengers can pass through the gate by using a subway card or buying a single ticket. The prior art (such as CN216531157U, CN209194402U, etc.) discloses various mobile gates, which make the arrangement of the gates more flexible.

[0003] During the peak of the subway, the train carriages are often very crowded, and passengers pass through the gate to take the escalator to the waiting hall. Due to the insufficient dispersion of the passenger flow, it will cause uneven distribution of passengers in each carriage of the train, leading to excessive crowding in some carriages, which not only affects the comfort of passengers, but also is not conducive to the stable operation of the train.

[0004] In addition, due to the different design of the entrance and exit positions of each subway station, passengers have to walk a long way to leave the subway station after getting off the train, which not only increases the burden of passengers, but also causes passengers to stay in the subway station, making the subway station hall more and more crowded. How to reasonably distribute the passenger flow has become a problem to be solved. SUMMARY

[0005] The purpose of the present application is to provide a subway station hall data information analysis method and system based on artificial intelligence to solve the problems raised in the background art.

[0006] In order to solve the above technical problems, the present application provides the following technical scheme: a subway station hall data information analysis system based on artificial intelligence, comprising: an image recognition module, a personnel dispersion module, a flow prediction module, a boarding interval module and a gate shunting module;

[0007] The image recognition module is used to obtain the monitoring video of passengers getting on and off the train in each carriage, and the number of people in each carriage is obtained by reading the monitoring video using AI video recognition technology.

[0008] The personnel dispersion module is used to calculate the personnel dispersion degree according to the position of each escalator, the position of the carriage and the number of passengers getting on the train, and to predict the number of passengers getting on and off the train in each carriage.

[0009] The flow prediction module is configured to predict the probabilities of passengers going to each arrival station after entering a station according to historical riding data and single-journey ticket purchase records of each passenger, calculate expected travel distances of passengers after getting off from each carriage according to locations of station hall exits and locations of escalators, and predict passenger flows of each station hall in a future period;

[0010] The riding interval module is configured to calculate expected passenger flows corresponding to locations of each escalator according to a passenger distribution of a next train, passenger flows and passenger dispersion degrees of each exit of a station hall, and divide riding intervals according to the predicted passenger flows and the probabilities of passengers arriving at each station.

[0011] The gate shunting module is configured to fix a gate above a movable device, divide riding intervals for each gate according to total expected travel distances of all passengers and the passenger distribution of the train, calculate adjustment distances of each gate according to the riding intervals, and adjust positions of the gates in a next period.

[0012] Further, the image recognition module comprises a video acquisition unit and an intelligent recognition unit.

[0013] The video acquisition unit is configured to acquire monitoring videos in each carriage of a train by using a monitoring system on a train door.

[0014] The intelligent recognition unit is configured to read contents in the monitoring videos and acquire boarding and alighting passenger information of each station.

[0015] Further, the passenger dispersion module comprises a riding positioning unit and a dispersion degree unit.

[0016] The riding positioning unit is configured to calculate probabilities of passengers boarding carriages at different distances according to boarding passenger numbers of each carriage, locations of escalators and numbers of passengers entering a station obtained at a gate.

[0017] The dispersion degree unit is configured to calculate a dispersion rate of passengers according to the probabilities of passengers boarding carriages at different distances, and predict the dispersion rate of passengers when a next train arrives according to a change trend of the dispersion rate.

[0018] Further, the flow prediction module comprises a ticket unit, an arrival station prediction unit and an entry and exit prediction unit.

[0019] The ticket unit is configured to acquire historical riding records of a passenger subway card and single-journey ticket purchase records.

[0020] The arrival station prediction unit is configured to calculate probabilities of passengers going to each arrival station in a next period.

[0021] The entry and exit prediction unit is configured to predict passenger flows of each station hall in a future period.

[0022] Further, the boarding section module comprises a train distribution unit, a station hall distribution unit and a section division unit;

[0023] The train distribution unit is configured to quantitatively determine the empty space in each carriage of the train;

[0024] The station hall distribution unit is configured to calculate the expected travel distance of passengers after getting off from each carriage according to the position of the station hall exit of the arrival station and the position of the escalator;

[0025] The section division unit is configured to calculate the expected passenger flow of each escalator and allocate the arrival station section to each gate according to the expected passenger flow.

[0026] Further, the gate shunting module comprises a mobile gate unit and a passenger allocation unit;

[0027] The mobile gate unit is configured to control the position of the gate so that the gate is aligned with different escalators, and the gates are connected by a retractable isolation belt;

[0028] The passenger allocation unit is configured to allocate the entry gate to the passenger according to the arrival station position of the passenger, and if the passenger holds a single-journey ticket, the passenger can only enter from the corresponding gate, and if the passenger enters by swiping the card, the passenger is recommended a waiting position in the bound electronic device after swiping the card.

[0029] The subway station hall data information analysis method based on artificial intelligence comprises the following steps:

[0030] Step S1. Obtain the monitoring video in each carriage of the train, use AI visual recognition technology to recognize the video, and obtain the personnel flow information in each carriage;

[0031] Step S2. Obtain the position of the escalator in the subway station hall, classify the train carriages according to the position of the escalator, verify the normality of the boarding rate in each type of carriage according to the personnel flow information in each type of carriage obtained in step S1, and calculate the personnel dispersion rate;

[0032] Step S3. According to the ticket sales record of single-journey tickets and the historical boarding record of subway card passengers in the last period, predict the passenger flow in the next period and calculate the probability of passengers going to each arrival station;

[0033] Step S4. According to the actual dispersion rate and the position of the entrance and exit of each arrival station, calculate the exit distance that passengers need to walk from each carriage, combine the probability of passengers going to each arrival station obtained in step S3 to calculate the expected distance of passengers exiting, and allocate the boarding position for passengers according to the expected distance and the empty space in each carriage of the train;

[0034] Step S5. According to the passenger distribution determined in step S4, the shunting position of each gate is calculated, the position of the gate is adjusted by the movable device, and after the passengers enter the station, the ticket checking gate and the waiting position are recommended for the passengers according to the probability of the passengers going to each arrival station.

[0035] Further, step S1 comprises:

[0036] Step S11. Obtain the monitoring video in each carriage of all trains when stopping in the station hall, and the monitoring video covers all passengers in the carriage;

[0037] Step S12. Train an artificial intelligence model using crowd statistical characteristics to obtain a trained AI visual recognition system, identify the monitoring video using the AI visual recognition system, and obtain personnel flow information in each carriage, including the number of passengers getting on, the number of passengers getting off, and the total number of passengers in the carriage.

[0038] This step can use artificial intelligence recognition equipment to identify the number of passengers getting on and off each carriage, monitor the passenger flow and density in the subway station hall, and the congestion of the channel entrance and exit in real time, so as to perform early warning and regulation.

[0039] Further, step S2 comprises:

[0040] Step S21. Obtain the position of the escalator in the subway station hall, mark the carriage opposite the lower end of the escalator as the reference carriage, and mark the number of carriages between the remaining carriages and the nearest reference carriage as U, U∈[0,N], where N represents the maximum value of the number of carriages between the train carriage and the nearest reference carriage, and the U value of the reference carriage is defined as 0;

[0041] Step S22. Obtain the number of passengers getting on in each carriage of the train obtained in step S1, count the average number of passengers getting on in each category of carriage, generate a data set F according to the mapping relationship between the average number of passengers getting on in the carriage and the U value, F={F0,F1,…,FN}, where F0 represents the average number of passengers getting on in the reference carriage, F1 represents the average number of passengers getting on in the carriage separated by one carriage from the reference carriage, and FN represents the average number of passengers getting on in the carriage separated by N carriages from the reference carriage. N} where F N represents the average number of passengers getting on in the train carriage with U value N;

[0042] Step S23. Verify the normality of the data set F according to the following formula, and calculate the personnel dispersion rate Q:

[0043]

[0044] where T represents the mean value of the number of passengers getting on all carriages, F i represents the mean value of the number of passengers getting on all carriages with U value i, represents the standard deviation of the data set F, Ku=3, and represents the overvalue peak degree constant;

[0045] Step S24. Obtain the personnel dispersion rate calculated when all trains stop at the station hall, and record the average value QV of the personnel dispersion rate as the actual dispersion rate of the station hall.

[0046] This step can identify the travel habits and needs of passengers, thereby optimizing subway line planning and station layout, and providing operational service quality.

[0047] Further, step S3 comprises:

[0048] Step S31. Number all subway stations reachable from the station hall, and count the single-journey ticket sales data of the ticket vending machines within the time period t0 before the departure of the previous train and the time period t0 before the arrival of the next train, wherein t0 is a preset passenger travel time;

[0049] Record the number of single-journey ticket sales and the mapping between the arrival stations as a data set H, wherein H={H1,H2,…,H k ,…,H m}, and m represents the number of all subway stations reachable by train from the station hall, H k represents the number of single-journey tickets sold to the arrival station numbered k, wherein k is an integer and k∈[1,m];

[0050] Step S32. When a subway card is swiped to enter the station, obtain the historical travel records of the subway card, and calculate the probability P k of the swiping passenger going to the arrival station k, wherein P k =L k / FL, and L k represents the number of times the user has exited from the arrival station k in the historical travel records, and FL represents the total number of times the passenger has traveled historically.

[0051] This step divides the station hall into multiple travel intervals, which can alleviate congestion, avoid the concentration of crowds in a specific area, help reduce personnel gathering, reduce the discomfort caused by platform pushing and shoving, and improve station hall safety and passenger satisfaction.

[0052] Further, step S4 comprises:

[0053] Step S41. Obtain the positions of the escalators in the station hall of all arrival stations, and according to the method of steps S21-S22, record the carriages opposite the lower end of the escalators in the arrival stations as reference carriages, and record the number of carriages separated from the nearest reference carriage as G.

[0054] Step S42. Number the carriages, and record the numbering result as {W1,W2,…,We,…,Wx}, wherein x represents the number of carriages, We represents the carriage numbered e, and e is an integer and e∈[1,x];

[0055] Calculate the expected distance re of the single ticket passenger to get off at the arrival station k, re = QV·A(e, k), where A(e, k) is the distance of the car We to get off at the arrival station k, and QV is the actual dispersion rate of the station hall;

[0056] Step S43. When the arrival station of the single ticket passenger is k, the expected distance re of the passenger to get off is re = QV·A(e, k);

[0057] Calculate the expected distance Re of the subway card passenger to get off from the car numbered We:

[0058]

[0059] where P k represents the probability of the card swiping passenger to go to the arrival station k, A(e, k) represents the distance of the car We to get off at the arrival station k, and QV represents the actual dispersion rate of the station hall;

[0060] Step S44. For the single ticket passenger with the arrival station k, traverse all possible values of the car number e to obtain the minimum value of the expected distance re to get off, and the car number e corresponding to the minimum value is the recommended car;

[0061] For the subway card passenger, traverse all possible values of the car number e to obtain the minimum value of Re, and the car number e corresponding to the minimum value is the recommended car;

[0062] Step S45. Obtain the number of people in each car of the next train, subtract the number of people in the car from the load number of the car to obtain the number of empty people in the car, and assign the car to the passenger according to the time sequence of getting on the train. When the number of empty people in the recommended car is 0, go to step S44 to remove all full cars and recalculate the recommended car;

[0063] When all the cars are full, send a prompt to the dispatch center that the train is about to be full.

[0064] This step can allocate the arrival station and the boarding interval to each gate with the strategy of minimizing the total outbound distance, reducing the outbound distance of the passenger, shortening the time consumption of the passenger getting on and off the station, reducing the passenger congestion and waiting, improving the overall traffic efficiency and operation effect of the subway line.

[0065] Further, step S5 includes:

[0066] Step S51. Obtain the recommended car of all single ticket passengers with the arrival station k, set the escalator closest to the recommended car as the arrival station k, and add a mark;

[0067] Count the number of markers of each escalator, calculate the number of gates allocated to each escalator B, B = round (BS * EN / m), wherein round represents the rounding function, BS represents the total number of gates, EN represents the number of markers of the escalator, and m is the number of all subway stations that can be reached by the train from the station hall;

[0068] Step S52. Adjust the position of the gate using the movable device, so that the number of gates facing each escalator is the same as the number of gates allocated in step S51 B;

[0069] Step S53. After the subway card passenger enters the station, send the recommended car information to the mobile device registered by the subway card passenger, and guide the passenger to take the escalator closest to the recommended car to wait;

[0070] Step S54. After the current train departs at time t0, go to step S1 to allocate passengers for the next train.

[0071] This step can make passengers from different arrival stations disperse in the car position to disperse the train flow.

[0072] Compared with the prior art, the beneficial effects achieved by the present application are:

[0073] 1. The present application can use artificial intelligence recognition equipment to recognize the number of passengers getting on and off each car, monitor the passenger flow and density in the subway station hall, and the congestion of the passage entrance and exit in real time, so as to provide early warning and control, can identify the travel habits and needs of passengers, and optimize the subway line planning and station layout, and provide operation service quality.

[0074] 2. The present application divides the station hall into multiple boarding intervals, which can alleviate congestion and avoid crowds gathering in a specific area, which helps to reduce personnel gathering, reduce the discomfort caused by platform pushing and shoving, and improve station hall safety and passenger satisfaction, so that passengers can obtain better platform passing space and better waiting and boarding experience.

[0075] 3. The present application can move the ticket gate, use the characteristics of passengers waiting nearby, make passengers from different arrival stations disperse in the car position to disperse the train flow, and at the same time, allocate the arrival station and the boarding interval to each gate with the strategy of minimizing the total outbound distance, reduce the outbound distance of passengers, shorten the time cost of passengers entering and leaving the station, reduce the situation of passenger congestion and queuing, and improve the overall traffic efficiency and operation effect of the subway line. BRIEF DESCRIPTION OF DRAWINGS

[0076] The accompanying drawings are used to provide further understanding of the present application, and constitute a part of the specification, which together with the embodiments of the present application, serve to explain the present application, and do not constitute a limitation to the present application. In the drawings:

[0077] Figure 1 is a structural schematic diagram of a subway station hall data information analysis system based on artificial intelligence of the present application;

[0078] Figure 2 is a step schematic diagram of a subway station hall data information analysis method based on artificial intelligence of the present application. DETAILED DESCRIPTION

[0079] 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. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor fall within the scope of protection of the present application.

[0080] Please refer to Figure 1 The present application provides a technical solution: a subway station hall data information analysis system based on artificial intelligence, comprising: an image recognition module, a personnel dispersion module, a traffic prediction module, a boarding interval module and a gate shunting module.

[0081] The image recognition module is used to obtain monitoring videos of passengers getting on and off in each carriage of the train, read the monitoring videos by using AI video recognition technology, and obtain the number of people in each carriage;

[0082] The image recognition module comprises a video acquisition unit and an intelligent recognition unit.

[0083] The video acquisition unit is used to obtain monitoring videos in each carriage of the train by using a monitoring system on the train door.

[0084] The intelligent recognition unit is used to read the content in the monitoring videos and obtain the number of people getting on and off at each station.

[0085] The personnel dispersion module is used to calculate the personnel dispersion degree according to the position of each escalator, the position of the carriage and the number of people getting on, and predict the number of people getting on and off in each carriage of the train.

[0086] The personnel dispersion module comprises a boarding positioning unit and a dispersion degree unit.

[0087] The boarding positioning unit is used to calculate the probability of passengers getting on the carriage at different distances according to the number of people getting on each carriage, the position of the escalator and the number of people getting on at the gate.

[0088] The dispersion unit calculates the dispersion rate of the passengers according to the probability of the passengers getting on the carriages at different distances, and predicts the dispersion rate of the passengers when the next train arrives according to the change trend of the dispersion rate.

[0089] The flow prediction module is configured to predict the probability of the passengers going to each arrival station after entering a station according to the historical riding data and the single-journey ticket purchase records of each passenger, calculate the expected travel distance of the passengers after getting off each carriage according to the location of the station hall exit of the arrival station and the location of the escalator, and predict the passenger flow of each station hall in a future period.

[0090] The flow prediction module comprises a ticket unit, an arrival station prediction unit, and an entry and exit prediction unit.

[0091] The ticket unit is configured to obtain the historical riding records of the passenger's metro card and the single-journey ticket purchase records.

[0092] The arrival station prediction unit is configured to calculate the probability of each arrival station that the passengers entering a station go to in a next period.

[0093] The entry and exit prediction unit is configured to predict the passenger flow of each station hall in a future period.

[0094] The riding interval module is configured to calculate the expected passenger flow corresponding to each escalator location according to the personnel distribution of the next train, the passenger flow of each entrance and exit of the station hall, and the personnel dispersion degree, and divide the riding interval according to the predicted passenger flow and the arrival station probability of each passenger.

[0095] The riding interval module comprises a train distribution unit, a station hall distribution unit, and an interval division unit.

[0096] The train distribution unit is configured to quantitatively determine the empty space in each carriage of the train.

[0097] The station hall distribution unit is configured to calculate the expected travel distance of the passengers after getting off each carriage according to the location of the station hall exit of the arrival station and the location of the escalator.

[0098] The interval division unit is configured to calculate the expected passenger flow of each escalator, and allocate the arrival station interval to each gate according to the expected passenger flow.

[0099] The gate shunting module is configured to fix the gate above the movable device, divide the riding interval for each gate according to the total expected travel distance of all passengers and the personnel distribution of the train, calculate the adjustment distance of each gate according to the riding interval, and adjust the position of the gate in a next period.

[0100] The gate shunting module comprises a movable gate unit and a passenger allocation unit.

[0101] The mobile turnstile unit is used to control the position of the turnstile so that it is aligned with different escalators. The turnstiles are connected by a telescopic isolation belt.

[0102] The passenger allocation unit is used to allocate entry gates to passengers based on their arrival station location. If a passenger holds a single-journey ticket, they can only enter the station through the corresponding gate. If a passenger swipes their card to enter the station, the unit will recommend a waiting location for the passenger in the bound electronic device after the passenger swipes their card.

[0103] like Figure 2 As shown, the method for analyzing subway station concourse data based on artificial intelligence includes the following steps:

[0104] Step S1. Acquire surveillance video footage from each carriage of the train, and use AI visual recognition technology to identify the video footage to obtain information on the flow of people in each carriage;

[0105] Step S1 includes:

[0106] Step S11. Obtain surveillance video footage from each carriage when all trains stop at the station hall, wherein the surveillance video footage covers all passengers in the carriage;

[0107] Step S12. Train an artificial intelligence model using crowd statistical features to obtain a trained AI visual recognition system. Use the AI ​​visual recognition system to identify the surveillance video to obtain the personnel flow information in each carriage. The personnel flow information includes: the number of people boarding, the number of people alighting, and the total number of people in the carriage.

[0108] Step S2. Obtain the location of the escalators in the subway station hall, classify the train cars according to the location of the escalators, and perform normality verification on the passenger flow information in each type of car based on the passenger flow information obtained in Step S1, and calculate the passenger dispersion rate.

[0109] Step S2 includes:

[0110] Step S21. Obtain the location of the escalator in the subway station hall. The car directly opposite the bottom of the escalator is denoted as the reference car. The number of cars between the remaining cars and the nearest reference car is denoted as U, where U∈[0,N], and N represents the maximum number of cars between the train car and the nearest reference car. The U value of the reference car is defined as 0.

[0111] Step S22. Obtain the number of passengers boarding in each carriage of the train obtained in Step S1, calculate the average number of passengers boarding in each category of carriage, and generate a dataset F based on the mapping relationship between the average number of passengers boarding in each carriage and the U value, wherein F = {F0, F1, ..., F2}. N}, where F N This represents the average number of passengers boarding a train carriage with a U value of N.

[0112] Step S23. Perform normality verification on the data set F according to the following formula to calculate the personnel dispersion rate Q:

[0113]

[0114] wherein T represents the mean value of the number of passengers on all carriages, F i represents the mean value of the number of passengers on all carriages with U value i, represents the standard deviation of the data set F, Ku=3 represents the excess kurtosis constant;

[0115] Step S24. Obtain the personnel dispersion rate calculated when all trains stop on the day, and record the average value QV of the personnel dispersion rate as the actual dispersion rate of the station hall.

[0116] Step S3. According to the ticketing records of single-journey tickets and the historical ride records of subway card passengers in the last period, the passenger flow in the next period is predicted, and the probability of passengers going to each arrival station is calculated.

[0117] Step S3 includes:

[0118] Step S31. All subway stations reachable from the station hall are numbered, and the single-journey ticket sales data of the ticket vending machines within t0 time length before the departure of the last train to t0 time length before the arrival of the next train are counted, wherein t0 is the preset passenger ride time.

[0119] The mapping of the number of single-journey ticket sales and the arrival stations is recorded as a data set H, wherein H={H1, H2, …, H k ,…,H m}, wherein m represents the number of all subway stations reachable by train from the station hall, H k represents the number of single-journey tickets sold to the arrival station numbered k, wherein k is an integer and k∈[1, m];

[0120] Step S32. When the subway card is swiped to enter the station, the historical ride records of the subway card are obtained, and the probability P k of the swiping passenger going to the arrival station k is calculated, wherein P k =L k / FL, wherein L k represents the number of times of exiting from the arrival station k in the historical ride records of the user, and FL represents the total number of times of historical rides of the passenger.

[0121] Step S4. According to the actual dispersion rate and the location of the entrances and exits of each arrival station, the exit distance of passengers from each carriage is calculated, the expected distance of passengers exiting is calculated by combining the probability of passengers going to each arrival station obtained in step S3, and the passengers are allocated to the boarding positions according to the expected distance and the empty space in each carriage of the train.

[0122] Step S4 comprises:

[0123] Step S41. Obtain the position of all arrival station station hall escalators, according to the method of steps S21-S22, the arrival station escalator lower end opposite the car is recorded as the reference car, and the remaining car is separated from the nearest reference car by the number of cars G;

[0124] Step S42. Number the car, and the numbering result is recorded as {W1, W2, …, We, …, Wx}, where x represents the number of cars, We represents the car numbered e, e is an integer and e∈[1, x];

[0125] Calculate the outbound distance A(e, k) of the car We in the arrival station k, A(e, k) = |Ue-G(e, k)|·HL, where Ue represents the number of cars separated from the nearest reference car in the current station hall car We, G(e, k) represents the number of cars separated from the nearest reference car in the station hall of the arrival station k, and HL is the length of a section of car;

[0126] Step S43. When the arrival station number of the single journey ticket passenger is k, the expected distance re of the passenger out of the station is re=QV·A(e, k);

[0127] Calculate the expected distance Re of the subway card passenger out of the station from the car numbered We:

[0128]

[0129] Where P k represents the probability of the card swiping passenger going to the arrival station k, A(e, k) represents the outbound distance of the car We in the arrival station k, and QV represents the actual dispersion rate of the station hall;

[0130] Step S44. For single journey ticket passengers arriving at station k, traverse all possible values of car number e to obtain the minimum value of the expected outbound distance re, and the car number e corresponding to the minimum value is the recommended car;

[0131] For subway card passengers, traverse all possible values of car number e to obtain the minimum value of Re, and the car number e corresponding to the minimum value is the recommended car;

[0132] Step S45. Obtain the number of people in each car of the next train, obtain the number of empty cars by subtracting the number of people in the car from the load number of the car, and assign the car to the passenger according to the time sequence of entering the station. When the number of empty cars in the recommended car is 0, go to step S44, and after removing all full cars, recalculate the recommended car;

[0133] When all the carriages are full, the dispatch center is prompted that the train is about to be full.

[0134] Step S5. According to the passenger distribution determined in step S4, the shunting position of each gate is calculated, the position of the gate is adjusted by the movable device, and after the passengers enter the station, the ticket gate and the waiting position are recommended for the passengers according to the probability of the passengers going to each arrival station.

[0135] Step S5 includes:

[0136] Step S51. Obtain the recommended carriages of all single-journey ticket passengers of the arrival station k, set the escalator closest to the recommended carriages as the entrance escalator of the arrival station k, and add a mark;

[0137] Iterate through all possible values of k, count the number of marks of each escalator, and calculate the number of gates B allocated to each escalator, wherein B = round(BS·EN / m), wherein round represents the rounding function, BS represents the total number of gates, EN represents the number of marks of the escalator, and m is the number of all subway stations that can be reached by the train from the station hall;

[0138] Step S52. The position of the gate is adjusted by the movable device so that the number of gates directly opposite each escalator is the same as the number of gates B allocated in step S51;

[0139] Step S53. After the subway card passengers enter the station, the information of the recommended carriages is sent to the mobile device registered by the subway card passengers, and the passengers are guided to descend from the escalator closest to the recommended carriages;

[0140] Step S54. After the current train departs at time t0, go to step S1 to allocate passengers for the next train.

[0141] Embodiment:

[0142] There are 20 ticket gates and 5 descending elevators in the subway station A, and the train has 15 carriages. The descending elevators are directly opposite the 2nd, 5th, 8th, 11th and 13th carriages of the train. Through the image recognition system, the total number of passengers on the 2nd, 5th, 8th, 11th and 13th carriages is 75, the total number of passengers on the 1st, 3rd, 4th, 6th, 7th, 9th, 10th, 12th and 14th carriages is 81, and the number of passengers on the 15th carriage is 5. The dispersion rate of the platform is calculated to be 1.38. According to the dispersion rate of all previous trains on the same day, the average dispersion rate of the day is calculated to be 1.4.

[0143] In the preset time, the number of the single-journey tickets is 120, and the number of the subway cards entering the station is 80, so the number of times that each train compartment is selected as the recommended compartment is respectively 15, 15, 20, 40, 10, 20, 5, 15, 13, 7, 10, 5, 5, 15 and 5, and the passengers carried by the No. 1 and No. 2 down elevators are the most, so the gate controller controls the passengers to gather at the No. 1 and No. 2 elevators.

[0144] It should be noted that the relational terms herein such as first and second and the like are used solely to distinguish one from another entity or action, without necessarily requiring or implying any actual 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.

[0145] Finally, it should be noted that the above-described embodiments are merely preferred embodiments of the present application, but are not used to limit the present application, and although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions recorded in the foregoing embodiments, or equivalently replace some technical features. Any modification, equivalent replacement, improvement, etc. 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 analyzing subway station concourse data based on artificial intelligence, characterized in that: The method includes the following steps: Step S1. Acquire surveillance video footage from each carriage of the train, and use AI visual recognition technology to identify the video footage to obtain information on the flow of people in each carriage; Step S2. Obtain the location of the escalators in the subway station hall, classify the train cars according to the location of the escalators, and combine the passenger flow information in each type of car obtained in Step S1 to perform normality verification on the passenger ratio in each type of car and calculate the passenger dispersion rate. Step S3. Based on the ticket sales records of single-journey tickets in the previous period and the historical travel records of subway card passengers, predict the passenger flow in the next period and calculate the probability of passengers going to each arrival station. Step S4. Based on the entrance and exit locations of each arrival station, calculate the distance passengers need to walk from each carriage to board the train. Combining the probability of passengers going to each arrival station obtained in Step S3 and the actual dispersion rate obtained in Step S2, calculate the expected distance for passengers to exit the station. Based on the expected distance and the available space in each carriage of the train, assign boarding positions to passengers. Step S5. Based on the passenger allocation determined in step S4, calculate the diversion position of each gate, adjust the position of the gate using a movable device, and recommend waiting locations for passengers after they enter the station based on the probability that they will go to each arrival station.

2. The method for analyzing subway station hall data information based on artificial intelligence according to claim 1, characterized in that: Step S1 includes: Step S11. Obtain surveillance video footage from each carriage when all trains stop at the station hall, wherein the surveillance video footage covers all passengers in the carriage; Step S12. Train an artificial intelligence model using crowd statistical features to obtain a trained AI visual recognition system. Use the AI ​​visual recognition system to identify the surveillance video to obtain the personnel flow information in each carriage. The personnel flow information includes: the number of people boarding, the number of people alighting, and the total number of people in the carriage. Step S2 includes: Step S21. Obtain the location of the escalator in the subway station hall. The car directly opposite the bottom of the escalator is denoted as the reference car. The number of cars between the remaining cars and the nearest reference car is denoted as U, where U∈[0,N], and N represents the maximum number of cars between the train car and the nearest reference car. The U value of the reference car is defined as 0. Step S22. Obtain the number of passengers boarding in each carriage of the train obtained in Step S1, calculate the average number of passengers boarding in each category of carriage, and generate a dataset F based on the mapping relationship between the average number of passengers boarding in each carriage and the U value, wherein F = {F0, F1, ..., F2}. N }, where F N This represents the average number of passengers boarding a train carriage with a U value of N. Step S23. Perform normality verification on dataset F using the following formula, and calculate the dispersion rate Q: Where T represents the average number of passengers boarding in all carriages, and F i This represents the average number of passengers boarding all carriages with a value of U = i. The standard deviation of dataset F is represented by Ku = 3, which represents the excess kurtosis constant. Step S24. Obtain the passenger dispersion rate calculated when all trains stop on the day, and record the average passenger dispersion rate QV as the actual dispersion rate of the station hall.

3. The method for analyzing subway station hall data information based on artificial intelligence according to claim 2, characterized in that: Step S3 includes: Step S31. Number all subway stations accessible from the station hall and collect data on single-journey tickets sold by the ticket vending machines within t0 hours before the departure of the previous train and t0 hours before the arrival of the next train, where t0 is the preset passenger travel time. The mapping between the number of single-journey tickets sold and the arrival stations is denoted as dataset H, where H = {H1, H2, ..., Hk, ..., Hm}, where m represents the number of all subway stations that can be reached by train from the station hall, and Hk represents the number of single-journey tickets sold with arrival station number k, where k is an integer and k ∈ [1, m]. Step S32. When the subway card is swiped to enter the station, the historical travel record of the subway card is obtained, and the probability Pk of the swiping passenger going to the arrival station k is calculated, where Pk = Lk / FL, where Lk represents the number of times the user exits from the arrival station k in the historical travel record, and FL represents the total number of times the passenger has taken the subway in the historical record.

4. The method for analyzing subway station hall data information based on artificial intelligence according to claim 3, characterized in that: Step S4 includes: Step S41. Obtain the location of all escalators in the station hall of the arrival station. Following the method of steps S21-S22, the car directly opposite the bottom of the escalator at the arrival station is recorded as the reference car, and the number of cars between the remaining cars and the nearest reference car is recorded as G. Step S42. Number the carriages, and record the numbering results as {W1,W2,…,We,…,Wx}, where x represents the number of carriages, We represents the carriage numbered e, and e is an integer and e∈[1,x]; Calculate the exit distance A(e,k) of car We in the arrival station k, where A(e,k) = |Ue-G(e,k)|·HL, where Ue represents the number of cars between car We in the current station hall and the nearest reference car, G(e,k) represents the number of cars between car We in the station hall of the arrival station k and the nearest reference car, and HL is the length of a car segment; Step S43. When the arrival station number of a one-way ticket passenger is k, the expected distance for the passenger to exit the station is re = QV·A(e,k); Calculate the expected distance Re for a metro card passenger exiting from car number We: Where P k Let A(e,k) represent the probability that a passenger using a card will travel to arrival station k, let A(e,k) represent the distance that carriage We will travel from the station k, and let QV represent the actual dispersion rate in the station hall. Step S44. For a one-way ticket passenger whose destination is k, iterate through all possible values ​​of carriage number e to obtain the minimum value of the expected distance re from the exit station. The carriage number e corresponding to this minimum value is the recommended carriage. For metro card passengers, iterate through all possible values ​​of car number e to find the minimum value of Re. The car number e corresponding to the minimum value is the recommended car. Step S45. Obtain the number of people in each carriage of the next train. Subtract the number of people in the carriage from the carriage's capacity to get the number of empty carriages. Assign carriages to passengers according to their arrival time. When the number of empty carriages in the recommended carriages is 0, go to step S44, remove all full carriages, and recalculate the recommended carriages. When all carriages are full, a notification is sent to the dispatch center that the train is about to be full.

5. The method for analyzing subway station hall data information based on artificial intelligence according to claim 4, characterized in that: Step S5 includes: Step S51. Obtain the recommended carriages for all single-journey ticket passengers whose arrival station is k, set the escalator closest to the recommended carriage as the entrance escalator for arrival station k, and add a mark; Iterate through all possible values ​​of k, count the number of markers for each escalator, and calculate the number of turnstiles B assigned to each escalator. B = round(BS·EN / m), where round represents the rounding function, BS represents the total number of turnstiles, EN represents the number of markers for the escalator, and m is the number of subway stations that can be reached by train from the station hall. Step S52. Use the movable device to adjust the position of the turnstiles so that the number of turnstiles facing each escalator is the same as the number of turnstiles B allocated in step S51; Step S53. After a metro card passenger enters the station, the recommended carriage information is sent to the mobile device registered by the metro card passenger, and the passenger is guided to descend the escalator closest to the recommended carriage to wait for the train; Step S54. After the current train departs from the station for a period of time t0, proceed to step S1 to assign passengers to the next train.

6. A subway station hall data information analysis system based on artificial intelligence, characterized in that: The system includes the following modules: image recognition module, personnel dispersion module, traffic flow prediction module, boarding section module, and turnstile diversion module; The image recognition module is used to acquire monitoring videos of passengers getting on and off the train in each carriage, and uses AI video recognition technology to read the monitoring videos to obtain the number of people in each carriage. The personnel dispersion module is used to calculate the personnel dispersion degree based on the location of each escalator, the location of the carriage, and the number of passengers boarding, and to predict the number of passengers boarding and alighting in each carriage of the train. The traffic prediction module is used to predict the probability of a passenger going to each arrival station after entering the station based on each passenger's historical travel data and single-journey ticket purchase record, and to calculate the expected travel distance of passengers after getting off the train from each carriage based on the location of the station hall exit and the location of the escalator, and to predict the passenger flow in each station hall in the future period. The boarding section module is used to calculate the expected passenger flow corresponding to each escalator position based on the passenger distribution of the next train, the passenger flow and passenger dispersion at each entrance and exit of the station hall, and to divide the boarding section according to the predicted passenger flow and the arrival probability of each passenger. The gate diversion module is used to fix the gate above the movable device. Based on the total expected travel distance of all passengers and the distribution of passengers on the train, it divides the travel area for each gate, calculates the adjustment distance of each gate based on the travel area, and adjusts the position of the gate in the next cycle.

7. The subway station hall data information analysis system based on artificial intelligence according to claim 6, characterized in that: The image recognition module includes: a video acquisition unit and an intelligent recognition unit; The video acquisition unit is used to acquire monitoring videos of each carriage of the train using the monitoring system on the train doors; The intelligent recognition unit is used to read the content in the surveillance video and obtain the number of passengers getting on and off at each station; The personnel dispersion module includes: a vehicle positioning unit and a dispersion unit; The passenger positioning unit is used to calculate the probability of passengers boarding in carriages at different distances based on the number of passengers boarding in each carriage, the location of the escalator, and the number of passengers entering the station obtained at the turnstile. The dispersion unit calculates the passenger dispersion rate based on the probability of boarding in carriages at different distances, and predicts the passenger dispersion rate when the next train arrives based on the trend of the dispersion rate.

8. The subway station hall data information analysis system based on artificial intelligence according to claim 7, characterized in that: The traffic prediction module includes: a ticketing unit, an arrival station prediction unit, and an entry / exit prediction unit; The ticketing unit is used to obtain passengers' historical travel records from their metro cards and the sales records of single-journey tickets; The arrival station prediction unit is used to calculate the probability of each arrival station that the arriving passenger will go to in the next period; The entry / exit prediction unit is used to predict the passenger flow in each station hall within a future period.

9. The subway station hall data information analysis system based on artificial intelligence according to claim 8, characterized in that: The passenger section module includes: a train distribution unit, a station hall distribution unit, and a section division unit; The train distribution unit is used to quantitatively determine the available space in each carriage of the train. The station hall distribution unit is used to calculate the expected travel distance of passengers after disembarking from each carriage, based on the location of the station hall exit and the location of the escalator. The interval division unit is used to calculate the expected passenger flow for each escalator and allocate arrival station intervals to each gate according to the expected passenger flow.

10. The subway station hall data information analysis system based on artificial intelligence according to claim 9, characterized in that: The turnstile diversion module includes: a mobile turnstile unit and a passenger allocation unit; The mobile turnstile unit is used to control the position of the turnstile so that it is aligned with different escalators. The turnstiles are connected by a telescopic isolation belt. The passenger allocation unit is used to allocate entry gates to passengers based on their arrival station location. If a passenger holds a single-journey ticket, they can only enter the station through the corresponding gate. If a passenger swipes their card to enter the station, the unit will recommend a waiting location for the passenger in the bound electronic device after the passenger swipes their card.

Images