Research On Detection Method Of Safety Door In Construction Elevator

With the development of urbanization,the use of construction elevators has become more and more frequent.Elevator safety doors are the only entrances for construction workers and tools to enter the building through elevators.To avoid safety accidents such as construction workers or objectsfalling from buildings caused by safety doors that are not closed.This paper adopts the real-time monitoring of the state of the elevator's safety gate,and studies the state-of-the-art safety angle-recognition algorithm,which has important engineering practical value.Based on the analysis of the characteristics of the actual working conditions of the construction safety door,this article has determined the plan of monitoring the elevator safety door with machine vision,solved the elevator safety gate control problem economically and practically.Considering that the direction of movement of the security door is perpendicular to the optical axis of the camera,since the perspective transformation results in non-rigid deformation of the image,three methods are used to identify the corner of the security door.The first method is based on feature point matching image similarity estimation.This paper defines the similarity coefficient based on matching feature points.Considering the non-rigid deformation of the image,in order to improve the accuracy of the algorithm,this paper proposes the matching based on the vector field consensus algorithm.Purification.Through experiments,it is concluded that the vector field consistency algorithm can effectively eliminate the mis-match.Through segmental linearization,the feature point matching algorithm can measure the angle within the range,and the sensitivity of the range is higher than the range;the algorithm takes an average time of 500 ms.The second method is to use the feature change of the image frame of the security gate;through the image edge extraction,edge culling and connection,the vector coordinates of the image security door frame segment are obtained to obtain the rotation angle.The experiment concludes that when the safety door rotation angle is in the range,the algorithm can be used to detect the algorithm and the algorithm takes an average of 400 ms.The third method is based on Halcon's deformation template shape matching and similarity estimation.Through experiments,it is concluded that Halcon's deformed template matching algorithm can measure the angle within the range,but the sensitivity is low;the algorithm takes an average time of 0.63 ms.Considering the working conditions of the construction safety door comprehensively,this paper proposes the use of the frame feature algorithm,which can easily achieve a wide range of angle detection,and the algorithm time also meets the requirements of the control system.So the real-time monitoring of the state of construction elevator safety doors has been completed.