Research And Design Of Intelligent Monitoring System For Elevator Car Based On Computer Vision

According to the blueprint of the National New Urbanization Plan(2014-2020),large-scale popularization of high-rise buildings is an unavoidable trend.The elevator because of its high transmission efficiency,comfort,safe and reliable,gradually become people daily life,especially in high-rise building is vital part of transportation passage way.At the same time,the improvement of elevator safety performance becomes the key point to ensure the safety of people's life.So research how to realize the application of computer vision in combination with intelligent monitoring system of the embedded system to scientific management of elevator,improving the efficiency of the elevator using the safe and stable,has important theoretical and practical value.The system designed in this paper focuses on the switch status of the elevator car door and the number of passengers in the elevator car.The system combines embedded technology and computer vision two important technologies,which study how to realize collection and processing of image information in the elevator car and detected data presentation.In the end,users can use the browser to monitor monitoring system data in real time.Firstly,according to the specific application environment and analysis and comparison of video automatic identification schemes,a design scheme and software hardware platform of intelligent monitoring recognition system used in the elevator car based on computer vision are determined.At the same time,in this paper,it takes Exynos 4412 based on Cortex-A9 core and embedded Linux system as the control center of the intelligent monitoring system,selects SQLite3 as the local database and selects Boa server as embedded Web server.According to the needs of the system,it completes the construction of the software development platform that includes the establishment of the host development environment,the implementation of the embedded Linux platform and the transplant of OpenCV.Secondly,driving USB camera by ARM development board and capturing video images in the elevator car through the video capture program of V4L2.In order to reduce the complexity of the image and enhance the distinctiveness of the image features,it performs a series of image processing on the captured video frame image,which includes image grayscale,bilateral filtering,edge detection and edge refinement..Finally,this paper proposes a switch state detection algorithm based on Hough transform line detection and a statistic algorithm for people in elevator cars based on improved Hough transform circle detection.At the application layer through uses the OpenCV library function and the corresponding algorithm to detect the relevant status of the elevator car,then saving the test result in the local database.It will finally realize establish interaction between users and database with embedded Web serve.This article tests the functions of video image acquisition,elevator car door status detection,number of passengers in the car,and server management platform based on the computer vision of the elevator car intelligent monitoring system.The measured results show that the design of the monitoring system in this paper has the advantages of high stability,high accuracy,small equipment volume and convenient installation.This system is a preliminary application of the elevator monitoring and application.which embedded technology combined with computer vision.In this age,intelligent have been widely used,so that the system proposed in this paper has a certain practical value and application scene.