Guardrail crossing behavior recognition method, system and equipment based on skeleton key points and medium

CN121600592APending Publication Date: 2026-03-03CRSC COMM & INFORMATION GRP CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202511629493.X
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-11-07
Publication Date
2026-03-03

AI Technical Summary

Technical Problem

Existing fence-climbing detection technologies suffer from problems such as high false alarm rate, poor environmental robustness, inability to understand behavioral semantics, high computational resource consumption, and weak system stability, making it difficult to achieve high-precision, real-time fence-climbing behavior recognition under limited computing resources.

Method used

A hierarchical processing and multi-feature fusion method based on skeletal key points is adopted. Through human skeletal key point detection, temporal smoothing and confidence cleaning, physical space mapping, and hierarchical decision state machine judgment, the climbing behavior is identified and alarm information is output.

Benefits of technology

It achieves high-precision, low-false-alarm detection of fence crossings, possesses strong environmental robustness and low resource consumption, is easy to deploy at the edge, adapts to targets of different sizes and crossing speeds, and maintains stable detection performance.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN121600592A_ABST
    Figure CN121600592A_ABST
Patent Text Reader

Abstract

The invention relates to a skeleton key point-based guardrail crossing behavior identification method, system and device, and a medium, and the method comprises the steps: carrying out the detection of human skeleton key points of an input video stream, and obtaining sequence data containing key points of a preset type; performing time sequence smoothing and confidence-based data cleaning on the sequence data; based on pre-calibrated guardrail parameters, mapping image coordinates of key points in the sequence data into height information of a physical space; based on the height information, calculating vertical motion characteristics of the center of mass and the feet of the human body in the sequence data; inputting the vertical motion characteristics of the center of mass and the feet of the human body in the sequence data into a pre-established hierarchical judgment state machine for behavior logic judgment; and when the logical judgment result of the hierarchical judgment state machine is a crossing behavior, outputting alarm information. The method can be widely applied to the technical fields of intelligent algorithms, neural networks and skeleton key point detection.
Need to check novelty before this filing date? Find Prior Art