CN-121725527-B - Rail transit non-inductive ticket checking method and system based on fusion of space-time track and biological characteristics

CN121725527BCN 121725527 BCN121725527 BCN 121725527BCN-121725527-B

Abstract

The invention relates to the technical field of ticketing equipment, and discloses a track traffic non-inductive ticketing method and a track traffic non-inductive ticketing system based on fusion of space-time tracks and biological characteristics, wherein the track traffic non-inductive ticketing method comprises the steps of acquiring a continuous space coordinate sequence of a moving target in a ticketing area and a corresponding biological identification characteristic group in real time; the method comprises the steps of determining distribution weight according to the advancing deflection angle of a moving target relative to a sensor, calculating a verification confidence index, establishing a first buffer zone and a second buffer zone which comprise ticket checking decision lines and are distributed on two sides of the ticket checking decision lines and have different physical lengths on a passing path, monitoring displacement vectors of the moving target relative to a decision domain, and driving a ticket checking transaction state machine to migrate.

Inventors

ZHANG JUNLIN
WU GUOPING
XIANG HAO
LI DESHENG
YI CHAO
ZHANG QIANG
CHEN JIE
LIU HUA
LIU YANG

Assignees

四川省中车铁投轨道交通有限公司

Dates

Publication Date: 20260512
Application Date: 20260214

Claims (9)

1. The track traffic non-inductive ticket checking method based on the fusion of the space-time track and the biological characteristics is characterized by comprising the following steps: step S1, acquiring a continuous space coordinate sequence of a moving object in a ticket checking area in real time, and extracting a biological identification feature group aligned with the continuous space coordinate sequence on a sampling time sequence; s2, calculating similarity probability values of each sub-item feature in the biological recognition feature group and a preset account feature library, and calculating a verification confidence index representing identity attribution certainty according to distribution weights determined by the advancing deflection angles of the moving targets relative to the sensors, wherein the advancing deflection angles are selected At the position of To the point of Range of range Obtaining face features, gait structural features and clothing texture feature sample sequences corresponding to the sampling points for discrete sampling points of step length, and calculating information entropy of the dimension features under corresponding deflection angles Information entropy Characterizing feature matching uncertainty according to information entropy The ratio of reciprocal to the sum of the reciprocal of the feature dimensions determines the weight coefficient Weight coefficient The calculation formula of (2) is as follows: wherein The total dimension of the biometric feature, Is the first Information entropy of dimensional characteristics under the current deflection angle; step S3, a first buffer area and a second buffer area which comprise a ticket checking decision line and are distributed on two sides of the ticket checking decision line are established on a passing path, and the physical length of the first buffer area along the advancing direction of the ticket checking decision line is set to be larger than the physical length of the second buffer area on the exiting side; Step S4, monitoring relative displacement vectors of the moving target relative to the first buffer zone, the second buffer zone and the ticket checking decision line, and driving the ticket checking transaction state machine to execute state migration, wherein when the moving target passes through the first buffer zone, the centroid position of the moving target crosses the ticket checking decision line, and the verification confidence index is not lower than a preset deduction confidence threshold value, the ticket checking transaction state machine jumps to a charging compliance state and generates a deduction instruction message; In the step S4, the ticket checking transaction state machine executes the following operations under the transaction suspension state, namely, the step S501 is used for maintaining closed-loop tracking of a moving object and starting an auxiliary acquisition window to acquire lateral area characteristics or back area characteristics of the moving object, the step S502 is used for carrying out increment correction on a biological identification characteristic set by utilizing the acquired lateral area characteristics or back area characteristics and recalculating a verification confidence index, and the step S503 is used for tracing the moment that the moving object crosses a ticket checking decision line and supplementing a deduction instruction message when the corrected verification confidence index reaches a deduction confidence threshold.
2. The track traffic non-inductive ticket checking method based on the fusion of space-time trajectories and biological characteristics according to claim 1, wherein in step S2, the calculation rule of the verification confidence index is as follows: , wherein, Is a verification confidence index; Is the total number of independent feature dimensions contained in the set of biometric features, and Is a positive integer; to assign corresponding ones of the weights A weight coefficient of the dimensional feature; Is the corresponding first in the similarity probability value Probability components of the dimensional features.
3. The track traffic non-inductive ticket checking method based on the fusion of space-time track and biological characteristics according to claim 1, wherein in step S3, the random swing displacement generated by the moving object at the edge of the ticket checking decision line is filtered by using the displacement delay interval formed by the physical length of the first buffer zone, so as to prevent the state jump from vibrating.
4. The track traffic non-inductive ticket checking method based on space-time track and biological feature fusion according to claim 1 is characterized by further comprising the following substeps of monitoring a cloud envelope volume value of a moving target point corresponding to a continuous space coordinate sequence, transmitting a modulated infrared detection sequence to a ticket checking area when the cloud envelope volume value of the moving target point exceeds a preset single object volume threshold value, collecting phase shift data formed by reflection of a moving target surface, and executing logic decoupling on a plurality of overlapped moving targets from the continuous space coordinate sequence according to space gradient distribution of the phase shift data.
5. The track traffic non-sensing ticket checking method based on space-time track and biological feature fusion according to claim 1 is characterized in that the acquisition of the continuous space coordinate sequence comprises the steps of acquiring a depth image stream of a ticket checking area by using a depth perception camera, setting the sampling frequency to be not lower than 30Hz, extracting moving target centroid coordinates in the depth image stream, connecting the moving target centroid coordinates of each sampling frame to generate the continuous space coordinate sequence, calculating displacement vectors between adjacent sampling frames in the continuous space coordinate sequence, and determining the physical invasion depth of a moving target relative to a first buffer area or a second buffer area.
6. The track traffic non-inductive ticket checking method based on space-time track and biological feature fusion of claim 1 is characterized by further comprising the steps of calculating path curvature continuity parameters of a continuous space coordinate sequence in step S701, comparing the path curvature continuity parameters before and after the interruption and evolution trend of verification confidence indexes when a moving target is subjected to track interruption and reappears in a sensor coverage blind area in step S702, and executing associated transparent transmission of ticket checking marks if the deviation is in a preset range.
7. The method for track traffic non-inductive ticketing based on temporal-spatial trajectory and biometric fusion of claim 1, further comprising identifying an accompanying attribute in the continuous spatial coordinate sequence, the accompanying attribute comprising a geometric deviation value of the carrier relative to a centroid of the moving object, and automatically calibrating an effective centroid position in the continuous spatial coordinate sequence according to the geometric deviation value to compensate for trajectory displacement deviation caused by luggage occupancy.
8. The track traffic non-inductive ticket checking method based on space-time track and biological feature fusion according to claim 1, wherein the biological recognition feature group comprises face features, gait structural features and clothing texture features, the distribution weight is adjusted in a real-time linear compensation mode according to the ambient illumination intensity data fed back by the photosensitive element, the method further comprises the following steps of step S1001, recording a migration path, a verification confidence index sequence and a generation time of a fare deduction instruction message of a ticket checking transaction state machine, and step S1002, packaging the recorded migration path, verification confidence index sequence and generation time into a ticket checking storage card data block and storing the ticket checking storage card data block into a local memory as ticket auditing basis.
9. The track traffic non-inductive ticket checking system based on the fusion of the space-time track and the biological characteristics is used for realizing the track traffic non-inductive ticket checking method based on the fusion of the space-time track and the biological characteristics as set forth in claim 1, and is characterized by comprising a data acquisition module, a characteristic processing module, a space decision module and a ticket checking transaction module: the data acquisition module is used for acquiring a continuous space coordinate sequence of a moving object in the ticket checking area in real time and synchronously extracting a biological recognition feature group aligned with the continuous space coordinate sequence on a sampling time sequence; The feature processing module is used for calculating a verification confidence index according to the similarity probability value of each sub-feature in the biological recognition feature group and a preset account feature library and combining the distribution weight determined by the advancing deflection angle of the moving target relative to the sensor Wherein, the advancing deflection angle is selected At the position of To the point of Range of range Obtaining face features, gait structural features and clothing texture feature sample sequences corresponding to the sampling points for discrete sampling points of step length, and calculating information entropy of the dimension features under corresponding deflection angles Information entropy Characterizing feature matching uncertainty according to information entropy The ratio of reciprocal to the sum of the reciprocal of the feature dimensions determines the weight coefficient Weight coefficient The calculation formula of (2) is as follows: wherein The total dimension of the biometric feature, Is the first The feature processing module is also used for calculating the path curvature continuity parameter of the continuous space coordinate sequence and identifying the geometric deviation value of the carrier relative to the centroid of the moving object so as to calibrate the effective centroid position; The space decision module is used for establishing a first buffer area and a second buffer area which comprise a ticket checking decision line and are distributed on two sides of the ticket checking decision line on a passing path, setting the physical length of the first buffer area on the incoming side of the ticket checking decision line along the advancing direction to be larger than the physical length of the second buffer area on the outgoing side so as to establish a decision domain with physical displacement delay characteristics; The ticket checking transaction module is used for monitoring the relative displacement vector of the moving target relative to the decision domain, driving the ticket checking transaction state machine to execute state migration, crossing the ticket checking decision line when the moving target passes through the first buffer zone and the centroid position of the moving target, and checking the confidence index When the charging confidence value is not lower than the preset charging confidence threshold value, skipping to a charging compliance state and generating a charging instruction message, and checking a confidence index When the deduction confidence threshold is lower than the deduction confidence threshold, jumping to a transaction suspension state, and locking a continuous space coordinate sequence until the complete biological feature data is acquired; the ticket checking transaction state machine executes the following operations under the transaction suspension state, namely maintaining closed-loop tracking of a moving target, starting an auxiliary acquisition window to acquire lateral region features or back region features of the moving target, performing incremental correction on a biological recognition feature group by utilizing the acquired lateral region features or back region features, recalculating a verification confidence index, and tracing the moment that the moving target crosses a ticket checking decision line when the corrected verification confidence index reaches a deduction confidence threshold value, and supplementing a deduction instruction message.

Description

Rail transit non-inductive ticket checking method and system based on fusion of space-time track and biological characteristics Technical Field The invention belongs to the technical field of ticketing equipment, and particularly relates to a track traffic non-inductive ticket checking method and system based on fusion of space-time tracks and biological characteristics. Background In the current automatic fare collection system of rail transit, the open type noninductive fare collection improves the passing efficiency of a station by canceling a physical gate, the noninductive fare collection generally adopts a depth perception unit and a vision acquisition component to acquire the motion track and biological characteristics of a passing object, and utilizes the identified passenger identity information to establish a binding relationship with a specific passing event, so that the mainstream technical mode provides a basic path for constructing a non-contact urban rail transit fare collection system; the problem of logical fault between identity attribution and physical track is difficult to solve by simply relying on perceived hardware performance stacking or general recognition algorithm iteration, for example, chinese patent application with an authorized bulletin number of CN119785387B discloses a space-ground collaborative target recognition method based on self-evolution visual prompt learning, recognition accuracy in a cross-platform environment is improved through view decoupling feature extraction and feature refinement, but in a track traffic charging scene, mapping ambiguity caused by identity evidence discreteness cannot be eliminated by simply improving recognition stability, physical displacement delay characteristic deep mining of passing objects is lacking, working conditions such as random swing near a decision perimeter, centroid deviation caused by luggage occupation, track adhesion under high-density passenger flow and the like are faced, a recognition result and charging action trigger are in a weak correlation state, charging host object inversion is generated, so that false transaction records are frequently generated, and the requirements of charging settlement law certainty and evidence integrity cannot be met. However, the rail transit station is taken as a typical high-density passenger flow application scene, has extremely high requirements on legal certainty of charging transactions, and in practical engineering application, physical displacement is a continuous evolution process and is influenced by factors such as ambient light shadow fluctuation, object posture randomness or physical shielding among human bodies, and the like; aiming at the above challenges, simply increasing the sampling frequency or introducing a complex recognition algorithm and other linear improvement paths can cause a great increase in data processing load, and the mapping ambiguity between discrete identity evidence and continuous physical tracks is difficult to eliminate, and the trigger mode depending on the instantaneous discrimination result leads to easy generation of logical charging host-guest inversion when facing the situations such as short-time loitering of passengers, physical adhesion of objects, track cross collision and the like, so that the shaking ticket checking system is used as a service foundation of a charging logic arbiter. Therefore, how to realize asynchronous association and logic latching of discrete identity features in continuous space-time coordinate vectors and construct a deterministic arbitration mechanism aiming at complex traffic intentions becomes a technical problem to be solved by the invention. Disclosure of Invention The invention provides a track traffic non-inductive ticket checking method based on fusion of space-time track and biological characteristics, which comprises the following steps: step S1, acquiring a continuous space coordinate sequence of a moving object in a ticket checking area in real time, and extracting a biological identification feature group aligned with the continuous space coordinate sequence on a sampling time sequence; S2, calculating a similarity probability value of each sub-item feature in the biological identification feature group and a preset account feature library, and calculating a verification confidence index representing identity attribution certainty according to an allocation weight determined by a traveling deflection angle of a moving target relative to a sensor; step S3, a first buffer area and a second buffer area which comprise a ticket checking decision line and are distributed on two sides of the ticket checking decision line are established on a passing path, and the physical length of the first buffer area along the advancing direction of the ticket checking decision line is set to be larger than the physical length of the second buffer area on the exiting side; And S4, monitoring relative displacement vectors o