Algorithm for detecting object and passenger behaviors on escalator floor plate based on machine vision

Abstract

The invention discloses an algorithm for detecting object and passenger behaviors on an escalator floor plate based on machine vision, which comprises the steps of 1) acquiring video images; 2) building a background model by using a CodeBook algorithm, and extracting an object and passenger foreground image; 3) performing head feature extraction on the foreground image by using an HOG descriptor; 4) performing head detection on the foreground image by using an AdaBoost cascade classifier; 5) tracking a passenger target by using a Kalman filter; 6) performing contour detection and ellipse fitting on the foreground image by using a Fitzgibbon algorithm; 7) removing the passenger target in the step 6), and tracking an object target by using the Kalman filter; and 8) and judging behaviors of the object and the passenger through analyzing the moving speed and the moving direction of center coordinates of the object and the passenger. According to the algorithm, detection and analysis can be performed on object and passenger behaviors on the escalator floor plate, an emergency plan can be started timely when an exception occurs, and potential hazards of an exceptional situation are enabled to be minimized.

Description

technical field [0001] The invention relates to the field of image pattern recognition and the field of elevator production technology, in particular to a machine vision-based algorithm for detecting floor board objects and passenger behavior of an escalator. Background technique [0002] With the rapid development of science and technology and the increasing awareness of people's safety, video surveillance has become an effective auxiliary means of public management and decision-making. In the modern society where escalators are becoming more and more popular, various safety issues follow. Nowadays, the video surveillance of escalators mostly stays in the stage of manual monitoring with image information as the main content. When an abnormality or accident occurs, the monitoring personnel often make the situation worse because they cannot respond in time or take corresponding measures. Today, with increasing labor costs, it is impossible to assign emergency personnel to eac...

AI Technical Summary

Problems solved by technology

At present, there is still relatively little research on machine vision monitoring technology for escalators in China. In view of its advantages of accuracy, intuition and low cost, it is necessary to speed up the research and development of corresponding technologies to reduce the incidence of escalator safety accidents

Images