Elevator abnormal behavior detection method, device, equipment and storage medium

By extracting the skeleton and edge features of elevator car video frames using a dual recognition model and combining them with multi-dimensional information to detect abnormal behavior in the car, the problem of low accuracy in existing technologies is solved, and higher detection accuracy and generalization ability are achieved.

CN117133050BActive Publication Date: 2026-06-09SHENZHEN INOVANCE TECH CO LTD
View PDF 1 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
SHENZHEN INOVANCE TECH CO LTD
Filing Date
2023-08-28
Publication Date
2026-06-09

AI Technical Summary

Technical Problem

In existing technologies, visual technology-based methods for detecting abnormal behavior in car cabins suffer from low accuracy, particularly in distinguishing between door-pulling behavior and head-holding behavior, leading to frequent false detections.

Method used

A dual recognition model is used to perform target detection and feature extraction on elevator car video frame images, obtaining human skeleton features and edge features respectively. Through matching and comparison analysis, combined with multi-dimensional information, abnormal behavior is identified and predicted, and error correction is performed to improve accuracy.

Benefits of technology

It achieves more accurate detection of abnormal behavior in the car, improves detection accuracy and generalization ability, and reduces false detection rate.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117133050B_ABST
    Figure CN117133050B_ABST
Patent Text Reader

Abstract

The application discloses a kind of car abnormal behavior detection method, device, equipment and storage medium, related to elevator safety technical field, method includes: obtaining the video frame image of elevator car;Video frame image is detected and skeleton is recognized using first identification model, and first behavior feature and human skeleton feature are obtained;Video frame image is detected and semantic segmentation using second identification model, and second behavior feature and human edge feature are obtained;Matching processing is carried out based on first behavior feature and second behavior feature, and abnormal behavior recognition result is obtained;Contrast analysis is carried out based on human skeleton feature and human edge feature, and abnormal behavior prediction result is obtained;According to abnormal behavior recognition result and abnormal behavior prediction result, final detection result is obtained.The application solves the problem of low accuracy of abnormal behavior detection in related art, and achieves the effect of improving the accuracy and generalization ability of abnormal behavior detection.
Need to check novelty before this filing date? Find Prior Art