A multi-feature fusion green pass abnormal vehicle auditing method and system

By employing a multi-feature fusion-based auditing method for abnormal green channel vehicles, and utilizing a multi-modal feature fusion network and edge computing architecture, the system addresses the issues of single inspection dimensions and data silos in green channel vehicle auditing. This enables accurate identification and real-time response to complex toll evasion behaviors, thereby improving the system's credibility and efficiency.

CN122196815APending Publication Date: 2026-06-12HENAN COMMUNICATIONS INVESTMENT GROUP CO LTD ANYANG BRANCH
View PDF 1 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
HENAN COMMUNICATIONS INVESTMENT GROUP CO LTD ANYANG BRANCH
Filing Date
2026-03-09
Publication Date
2026-06-12

AI Technical Summary

⚠Technical Problem

Existing green channel vehicle auditing technologies suffer from limitations such as single inspection dimensions, severe data silos, and low efficiency of manual auditing. This results in limited ability to identify complex toll evasion behaviors, high false positive and false negative rates, and a lack of real-time processing and edge collaboration mechanisms, affecting the system's practicality and reliability.

⚗Method used

A multi-feature fusion method for auditing abnormal green channel vehicles is adopted. By acquiring multi-source feature data, a passage feature vector, a structured cargo feature map, a behavior trajectory code, and a spatiotemporal constraint factor are generated. A multi-modal feature fusion network is used for deep semantic interaction. Combined with an edge computing and central cloud platform collaborative architecture, real-time intelligent analysis and interpretable auditing are achieved.

🎯Benefits of technology

It significantly improves the ability to accurately identify complex toll evasion behaviors, reduces the risk of misjudgment and omission, meets the high concurrency and low latency requirements of highway scenarios, and provides auditors with clear risk assessment criteria, thereby enhancing the system's practicality and decision credibility.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122196815A_ABST
    Figure CN122196815A_ABST
Patent Text Reader

Abstract

The application discloses a multi-feature fusion green pass abnormal vehicle auditing method and system, and belongs to the technical field of computers. The method comprises the following steps: acquiring vehicle pass, cargo images, historical behaviors and space-time context features, and respectively generating pass feature vectors, structured cargo feature maps, behavior trajectory codes and space-time constraint factors; inputting the four types of features into a multi-modal feature fusion network, and outputting an abnormal risk score after feature alignment, cross-modal attention interaction and gate fusion; and generating and issuing an auditing alarm instruction based on the comparison between the score and a dynamic risk threshold. The application also provides an edge-cloud collaborative architecture, and integrates scene adaptive adjustment and an interpretable basis generation mechanism based on gradient contribution in the network. The application realizes deep semantic fusion and dynamic interaction of multi-source heterogeneous features, significantly improves the recognition accuracy of complex fee evasion behaviors and the real-time response of auditing, and has good interpretability and scene adaptive ability.
Need to check novelty before this filing date? Find Prior Art