CN-121982790-A - ETC truck lane entrance method

CN121982790ACN 121982790 ACN121982790 ACN 121982790ACN-121982790-A

Abstract

The invention provides an ETC truck lane entrance method, which relates to the technical field of intelligent transportation and comprises the following steps of S1, synchronously collecting multi-source sensing data of a vehicle when the vehicle enters an identification area, S2, carrying out fusion processing and real-time verification on the multi-source sensing data, S3, generating a transaction decision instruction based on consistency comparison and matching verification results, S4, executing corresponding verification operation according to the type of an abnormal processing flow until the transaction is completed or the transaction is converted into manual processing, and acquiring multi-dimensional information such as three-dimensional physical contours, accurate wheelbase, cargo form, real-time axle weight, total weight, motion state and the like of the vehicle by a laser radar, a vision system and a dynamic weighing platform through time-space synchronous triggering and data fusion, so that a solid foundation is laid for the follow-up accurate verification.

Inventors

WANG PENGZHUANG
TIAN SHENGXIONG
SHAO ZHENGDA
ZHANG XIANG
HE WENTAO
SUN CE
GUO XIAOYU
Jia Maoxia
HUANG QIANG
LI XINGYUN
LI QUANFA
WANG GANG
ZHANG CHAOXIANG
Jiao Ziling
YANG XUDONG
WANG LINGXIAO
HE YAOZHONG
ZHU JIAN
Ren Feilin
CHEN MINRUI
XI WEI
LV YUN
LIU XU
LIU SHENGQING
MEI LEXIANG
HUANG YUN
ZHENG JINGJING
GAO SHAN
DENG JUAN
LI YONG

Assignees

交通运输部路网监测与应急处置中心

Dates

Publication Date: 20260505
Application Date: 20260209

Claims (10)

1. An ETC truck lane entry method, characterized by comprising the steps of: S1, synchronously acquiring multisource perception data of a vehicle when the vehicle enters an identification zone; the multi-source perception data at least comprises vehicle three-dimensional contour and characteristic image data acquired based on laser radar and vision fusion, vehicle weighing data acquired based on a dynamic weighing platform and vehicle-mounted unit issuing information acquired based on an OBU interaction module; s2, carrying out fusion processing and real-time verification on the multisource sensing data; the identified vehicle type axle number and license plate number information are subjected to consistency comparison with the vehicle-mounted unit issuing information, and the vehicle weighing data and a preset vehicle physical structure model are subjected to matching verification; s3, generating a transaction decision instruction based on a consistency comparison and matching verification result; if the data are consistent and the verification is passed, executing a normal ETC transaction flow and writing in entry information; If the data is inconsistent or the verification fails, triggering an abnormal processing flow; And S4, executing corresponding checking operation according to the type of the abnormal processing flow until the transaction is completed or the processing is changed into manual processing.
2. The method according to claim 1, wherein in step S1, the simultaneous acquisition of multi-source sensory data is triggered by: Detecting the arrival of a vehicle through a ground induction coil, and triggering a laser radar, a vision system and a dynamic weighing platform to start synchronously; if the ground induction coil fails, the trigger mode is automatically switched to a trigger mode based on radar motion detection, and the detection speed threshold is not lower than 5km/h.
3. The method according to claim 1, wherein the fusing and real-time checking of the multisource perceptual data in step S2 specifically comprises: s21, based on laser radar point cloud data and visual image data, identifying the model code, the axle number and the license plate number of the vehicle through an AI model identification model; S22, strictly comparing the identified number of axles with the number of axles in the issuing information of the vehicle-mounted unit, and performing double-factor matching on the identified license plate number and the license plate number in the issuing information of the vehicle-mounted unit, wherein the double-factor matching comprises comparison of OCR (optical character recognition) results and RFID (radio frequency identification) reading results; s23, comparing the total weight and axle weight data acquired by the dynamic weighing platform with an expected weight range estimated based on the axle number of the vehicle model, and analyzing whether the weight distribution proportion accords with a preset vehicle physical structure model.
4. The method of claim 3, wherein in step S22, the two-factor matching requires that the license plate number similarity be not less than 95% to be considered consistent, and wherein in step S23, the comparison is performed using different gross weight error thresholds based on whether the vehicle is overrun.
5. The method according to claim 1, wherein in step S3, if the data is consistent and verified, a 0-ary transaction is performed and entry information including the encrypted tamper-proof tag is written to the on-board unit.
6. The method according to claim 1, wherein the exception handling procedure in step S4 includes: if the release information is inconsistent with the identification result, starting license plate secondary identification and combining manual rechecking; if the weighing data are abnormal, guiding the vehicle to an artificial lane for rechecking; If fake license plates or axle number disguising behaviors are suspected, invoking an anti-cheating strategy library for analysis, and determining whether to trigger sound and light alarm and vehicle locking according to analysis results.
7. The method of claim 6, wherein invoking the anti-cheating policy library for analysis comprises: analyzing the vibration spectrum characteristics of the vehicle when passing through the weighing platform, and comparing the vibration spectrum characteristics with the frequency domain distribution of a normal vehicle; and analyzing the vehicle characteristics in the history passing records, and comparing the vehicle characteristics with the current identification characteristics in a consistency way.
8. The method according to claim 1, further comprising step S5, after completion of the transaction or exception handling: generating a verification report and a data consistency score of the transaction; and carrying out iterative optimization on the AI vehicle type recognition model and the anti-cheating strategy library based on the verification report and the abnormal sample.
9. The method of claim 1, wherein the method automatically switches to CPC card-based transaction mode and attempts to reestablish OBU communications when processing an OBU communication failure.
10. The method of claim 1, wherein the total processing delay of the acquisition, fusion processing, verification and transaction decision of the multi-source sensory data is no more than 200 milliseconds.

Description

ETC truck lane entrance method Technical Field The invention relates to the technical field of intelligent traffic, in particular to an ETC truck lane entrance method Background ETC systems play a key role in improving road traffic efficiency, however, in current truck lane entry transaction scenarios, the systems still have significant limitations in accurately and comprehensively sensing the actual state of the vehicle, which is mainly due to inherent drawbacks of the data acquisition and processing modes thereof. First, existing systems typically rely on a single or limited sensing module at the data acquisition level. For example, only a laser radar or a visual camera is adopted to obtain a two-dimensional or simple three-dimensional outline of a vehicle, the complex body structure, the cargo shape and the precise wheelbase distribution of a heavy truck are difficult to reconstruct by the data, the dynamic weighing platform can provide the total weight of the vehicle, but the precise distribution of the axle load, the axle group relation and the vibration characteristics of the vehicle in a motion state lack of depth perception; In the aspect of data processing and decision, the prior proposal lacks of deep fusion and systematic real-time verification of multi-source heterogeneous data, most systems only carry out simple data comparison (such as license plate number check) or threshold judgment (such as overrun detection), but do not construct a comprehensive verification chain integrating laser point cloud, visual images, dynamic weighing curves and OBU release information; The defects directly cause the problems of operation safety and efficiency, namely that the charging accuracy is difficult to guarantee, toll loss or dispute can be caused by the fact that the axle number of the vehicle is wrongly identified, space is provided for various cheating behaviors (such as towing pounds, jumping pounds and fake preferential vehicle types), the fairness of charging is improved, and when data contradiction occurs, the system is excessively dependent on manual intervention for checking, so that the lane passing efficiency is reduced (particularly when the traffic of the vehicle is large), and the management cost is increased. Disclosure of Invention The invention aims to solve the defects in the prior art, and provides an ETC truck lane entrance method which can deeply fuse multisource sensing data, realize intelligent real-time verification and automatically execute reliable transaction decisions so as to fundamentally improve the accuracy of entrance identification, the anti-cheating capability and the automation level and ensure fair charging and efficient lane smoothness. In order to achieve the above purpose, the present invention adopts the following technical scheme: An ETC truck lane entry method comprising the steps of: S1, synchronously acquiring multisource perception data of a vehicle when the vehicle enters an identification zone; the multi-source perception data at least comprises vehicle three-dimensional contour and characteristic image data acquired based on laser radar and vision fusion, vehicle weighing data acquired based on a dynamic weighing platform and vehicle-mounted unit issuing information acquired based on an OBU interaction module; s2, carrying out fusion processing and real-time verification on the multisource sensing data; the identified vehicle type axle number and license plate number information are subjected to consistency comparison with the vehicle-mounted unit issuing information, and the vehicle weighing data and a preset vehicle physical structure model are subjected to matching verification; s3, generating a transaction decision instruction based on a consistency comparison and matching verification result; if the data are consistent and the verification is passed, executing a normal ETC transaction flow and writing in entry information; If the data is inconsistent or the verification fails, triggering an abnormal processing flow; And S4, executing corresponding checking operation according to the type of the abnormal processing flow until the transaction is completed or the processing is changed into manual processing. Further, in the step S1, the synchronous acquisition of the multi-source sensing data is triggered by the following manner: Detecting the arrival of a vehicle through a ground induction coil, and triggering a laser radar, a vision system and a dynamic weighing platform to start synchronously; if the ground induction coil fails, the trigger mode is automatically switched to a trigger mode based on radar motion detection, and the detection speed threshold is not lower than 5km/h. Further, the step S2 of performing fusion processing and real-time verification on the multisource sensing data specifically includes: s21, based on laser radar point cloud data and visual image data, identifying the model code, the axle number and the license plate number of the