CN-121997194-A - Category information acquisition method and system for target place entrance and exit

CN121997194ACN 121997194 ACN121997194 ACN 121997194ACN-121997194-A

Abstract

The invention discloses a category information acquisition method and a category information acquisition system for an entrance and an exit of a target place, which particularly relate to the technical field of traffic monitoring and vehicle information fusion and are used for solving the problem of inaccurate vehicle identification caused by mismatching of multi-source time sequence data, wherein when a vehicle enters the place, the running track of the vehicle is firstly definitely tracked, and then different types of monitoring information are accurately corresponding to the same vehicle on the basis of the track, so that the problem of data dislocation caused by conditions such as acceleration, deceleration or lane change of the vehicle is effectively avoided; in addition, through unified coordination and comprehensive analysis on various monitoring data, the reliability of the data is intelligently judged, the situation that mismatching possibly exists can be actively screened out, rechecking can be timely carried out, the occurrence of vehicle identification errors and misjudgment is obviously reduced, the accuracy and reliability of information when vehicles enter and exit places are ensured, and the accuracy and the effectiveness of vehicle passing efficiency and on-site supervision are improved.

Inventors

JING YUANPENG
LIU ZHONGWEI
ZOU LIANGLONG

Assignees

山东元鸿智能科技有限公司

Dates

Publication Date: 20260508
Application Date: 20260129

Claims (10)

1. The category information acquisition method of the target place entrance is characterized by comprising the following steps: S1, continuously extracting a vehicle motion vector by an entrance laser micro-distance measuring array, and writing a motion vector main key into a temporary track index table; s2, after capturing a character frame, the license plate recognition module calls a motion vector main key, maps the main key to a license plate record and synchronously inserts a sequence time stamp; S3, searching a motion vector primary key when the remote sensing emission analyzer finishes analysis, predicting arrival delay write-up emission records according to the motion vector, and pushing the record to a fusion buffer area; S4, the dynamic weighing plate outputs the axle load value in real time and then refers to the same motion vector primary key, and performs time difference correction and sequence rearrangement with the emission record in the fusion buffer area to generate a complete multimode item; And S5, acquiring intersection density factors and scattered symmetry factors from the multimode entries by the fusion engine, deriving fusion confidence factors through machine learning, updating the vehicle type labels and issuing a gateway control end if the fusion confidence factors are higher than a preset threshold, otherwise, transferring the entries into a rechecking buffer zone and archiving the entries into a supervision database.
2. The method for acquiring category information of a destination site entrance as recited in claim 1, wherein the step S1 includes: The method comprises the steps of scanning vehicles in real time through a plurality of high-precision laser sensors, capturing position and speed information of the vehicles, generating a motion vector, generating a unique motion vector main key for each vehicle, ensuring the uniqueness of the motion vector main key based on initial time and initial position of the vehicles entering a detection area through a hash function, writing the motion vector main key and corresponding track data into a temporary track index table, wherein the temporary track index table comprises the motion vector main key, a timestamp, position coordinates and motion vector fields, and dynamically updating along with the movement of the vehicles.
3. The method for acquiring category information of a destination site entrance as recited in claim 2, wherein the step S2 includes: After capturing a character frame and identifying a license plate number, retrieving a vehicle track item closest to capturing time from a temporary track index table, acquiring a corresponding motion vector primary key, then pairing the motion vector primary key with the license plate number to generate an associated record, generating a sequence time stamp to record the identification completion time, and finally integrating the motion vector primary key, the license plate number and the sequence time stamp into a license plate record and storing the license plate record into a license plate database.
4. A method for acquiring category information of a destination site entrance and exit according to claim 3, wherein the step S3 includes the following steps: After the remote sensing emission analyzer finishes analysis, recording analysis completion time, retrieving a vehicle track item closest to the analysis completion time from a temporary track index table, extracting a motion vector primary key, predicting the position of the vehicle at the analysis completion time according to a motion vector sequence, and calculating arrival delay from an emission acquisition point to a predicted position.
5. The method for acquiring category information of a destination site entrance as recited in claim 4, wherein the step S3 further includes: And subtracting the arrival delay from the analysis completion time to obtain a corrected emission acquisition time stamp, integrating the motion vector primary key, the emission data and the emission acquisition time stamp into an emission record, and pushing the emission record to a fusion buffer area.
6. The method for acquiring category information of a destination site entrance as recited in claim 5, wherein step S4 includes: The method comprises the steps of measuring a vehicle axle load value by a dynamic weighing plate, recording measuring time, retrieving a motion track record closest to the measuring time from a temporary track index table, extracting a motion vector main key, retrieving an emission record from a fusion buffer area according to the motion vector main key, calculating the time difference between the axle load value measuring time and an emission acquisition time stamp, predicting the position of the vehicle at the measuring time by using a motion vector sequence, adjusting the emission record time stamp to realize time alignment, integrating the motion vector main key, a license plate number, corrected emission data, an axle load value sequence and the measuring time into multimode items, and arranging the multimode items according to the ascending order of the measuring time to generate an ordered multimode item list.
7. The method for acquiring category information of a destination site entrance as recited in claim 6, wherein step S5 includes: The method comprises the steps of extracting intersection density factors and scattered symmetry factors from multimode items, inputting the intersection density factors and the scattered symmetry factors into a pre-trained gradient lifting integrated classifier to output fusion confidence factors, and determining to update vehicle class labels or transfer multimode items into a rechecking buffer area by a fusion engine according to comparison of the fusion confidence factors and preset thresholds.
8. The method for acquiring category information of a destination site entrance as recited in claim 7, wherein the step S5 further includes: wherein the intersection density factor is obtained by calculating the total length of the intersection sections of the predicted track and the adjacent track divided by the target track length, and the scatter symmetry factor is obtained by calculating the discrete degree and the maximum-minimum deviation ratio of the multimode timestamp.
9. The method for acquiring category information of a destination site entrance as recited in claim 8, wherein the step S5 further includes: When the fusion confidence factor is higher than a preset threshold, the fusion engine determines the vehicle category according to license plate records, emission records and axle load value measurement data in the multimode item and issues a barrier gate control end, and when the fusion confidence factor is not higher than the preset threshold, the fusion engine transfers the multimode item to a review buffer zone and files the multimode item to a supervision database.
10. A category information acquisition system for a destination entry and exit, for implementing a category information acquisition method for a destination entry and exit according to any one of claims 1 to 9, comprising: continuously extracting a vehicle motion vector by the entrance laser micro-distance measuring array, and writing a motion vector main key into a temporary track index table; After capturing the character frame, the license plate recognition module invokes a motion vector primary key, maps the primary key to a license plate record and synchronously inserts a sequence time stamp; searching a motion vector primary key when the remote sensing emission analyzer finishes analysis, predicting arrival delay write-up emission records according to the motion vector, and pushing the record to a fusion buffer area; The dynamic weighing plate outputs the axle load value in real time and then refers to the same motion vector primary key, and performs time difference correction and sequence rearrangement with the emission record in the fusion buffer area to generate a complete multimode item; The fusion engine acquires the intersection density factor and the scattered symmetry factor from the multimode item, derives the fusion confidence factor through machine learning, updates the vehicle type label and issues the barrier gate control end if the fusion confidence factor is higher than a preset threshold, otherwise, the item is transferred into a recheck buffer zone and is archived to a supervision database.

Description

Category information acquisition method and system for target place entrance and exit Technical Field The invention relates to the technical field of traffic monitoring and vehicle information fusion, in particular to a category information acquisition method and system for a target place entrance and exit. Background The license plate recognition camera, the remote sensing emission monitoring beam and the embedded dynamic weighing plate are arranged at the entrance of the large logistics park, so that the transportation vehicle can synchronously finish emission standard judgment and axle load verification in a uniform-speed passing state. And the license plate frame flow, the emission spectrum line and the axle load characteristics generated in each detection link are transmitted to a fusion platform in real time through edge nodes, and the platform integrates multiple paths of data by utilizing a sliding time window algorithm to generate a vehicle type label and drive a barrier gate and a supervision record. In a dense traffic flow environment, the actions of acceleration, deceleration and lane changing continuously change the speed and track of the vehicle, so that the time stamp output by each detection link is dithered. The camera is affected by illumination fluctuation to generate exposure delay, the emission beam can return a result after spectrum analysis is completed, the weighing plate outputs weight immediately, and various output time sequences are difficult to keep synchronous, so that the record of the same vehicle is repeatedly misplaced. When the platform writes the data table according to the arrival sequence, the emission record is not returned yet, the weight information of the following vehicles arrives already, the aggregation logic is easy to mix different vehicle data under crowded flow, and finally, the tags of the superbus light vehicles and the compliant heavy vehicles are interchanged, the barrier gate is triggered by mistake, and the law enforcement chain is distorted. The traditional association mode only depends on license plate character consistency, ignores space-time drift in high-density streams, lacks real-time track prediction compensation, and becomes a core obstacle for the landing of a multi-source detection fusion scheme. In order to solve the above problems, a technical solution is now provided. Disclosure of Invention In order to overcome the defects of the prior art, the embodiment of the invention provides a category information acquisition method and a category information acquisition system for a target place entrance and exit, which are characterized in that when a vehicle enters a place, the running track of the vehicle is firstly definitely tracked, and then different types of monitoring information are accurately corresponding to the same vehicle on the basis of the track, so that the problem of data dislocation caused by acceleration, deceleration or lane changing of the vehicle is effectively avoided. In order to achieve the above purpose, the present invention provides the following technical solutions: A kind of information acquisition method of the entrance and exit of the goal place, comprising the steps of: S1, continuously extracting a vehicle motion vector by an entrance laser micro-distance measuring array, and writing a motion vector main key into a temporary track index table; s2, after capturing a character frame, the license plate recognition module calls a motion vector main key, maps the main key to a license plate record and synchronously inserts a sequence time stamp; S3, searching a motion vector primary key when the remote sensing emission analyzer finishes analysis, predicting arrival delay write-up emission records according to the motion vector, and pushing the record to a fusion buffer area; S4, the dynamic weighing plate outputs the axle load value in real time and then refers to the same motion vector primary key, and performs time difference correction and sequence rearrangement with the emission record in the fusion buffer area to generate a complete multimode item; And S5, acquiring intersection density factors and scattered symmetry factors from the multimode entries by the fusion engine, deriving fusion confidence factors through machine learning, updating the vehicle type labels and issuing a gateway control end if the fusion confidence factors are higher than a preset threshold, otherwise, transferring the entries into a rechecking buffer zone and archiving the entries into a supervision database. In a preferred embodiment, step S1 comprises the following: The method comprises the steps of scanning vehicles in real time through a plurality of high-precision laser sensors, capturing position and speed information of the vehicles, generating a motion vector, generating a unique motion vector main key for each vehicle, ensuring the uniqueness of the motion vector main key based on initial time and initial position of the vehicle