The Elevator Traction Machine And Wire Rope Fault Detection Method Based On Image Processing

Elevator is an important vehicle in modern society. Because of lack of the maintenance facility personnel and the backward detection method, there are security risks in the elevator.To protect passengers' personal safety and improve the elevator running safety, reliability, it is important to discover the elevator problems timely.Digital image processing technology can accurately, quickly obtain the fault information in images. A detection method based on the image processing is proposed in this paper, using the method to detect the more important part of the traction machine and the steel wire in elevator system which has a good application prospect and practical.The main works are as follows:(1) Refer to the fault detection technology at home and abroad. Analysis of the structure of the elevator machine, the wire rope and the common fault. Determine the fault detection method based on image processing. Study on the pretreatment method of the gray image of the elevator machine and the wire rope. The image of the elevator traction machine is obtained by the infrared technology. The image of the elevator wire rope is obtained by the CCD technology. To analyze the characteristic of the image gray algorithm of the average method, the weighted average method and the maximum value of method, the weighted average method is used to make the image of the elevator machine and the wire rope to a gray in our paper.(3) Study on the image of the elevator machine and the wire rope denoising method. The image of the elevator machine and the wire rope is denoising after graying. Analysis of the median filter, the wavelet transform, the Contourlet transform denoising algorithms, An improved Contourlet transform denoising algorithm is proposed. Firstly, the image of the Contourlet coefficients is obtained by the Contourlet transform. Secondly, the Contourlet coefficients are contracted by the neighborhood shrinkage method. The threshold and the shrinkage factor is improved to avoid the disadvantages that the threshold can not change with the change of the decomposition scale, and the shrinkage factor is too large. The Contourlet coefficients are processed by the improved threshold and the shrinkage factor. Lastly, the image is denoising by the inverse Contourlet transform. The simulation experiments show that the improved algorithm can better protect the image details of the elevator machine and the wire rope, avoiding the Gibbs phenomenon.(4) Study on the image of the elevator machine and the wire rope segmentation method. The image of the elevator machine and the wire rope is segmentation after denoising. An improved Geodesic Active Contour model is propose that combine with the Geodesic Active Contour model and Local Binary Fitting model. Firstly, the energy functional of the LBF model is introduced into the energy functional of the local variance information and the global variance information. To overcome the energy functional fall into the local minimum in the complex images of gray scale, the image is over segmentation. Secondly, the energy functional is normalized as the edge stopping function of the GAC model which avoiding the edge stoppingfunction of the original GAC model can not be separated from the weak boundary and no boundary region in the image. Lastly, the new edge stopping function is treated by the gradient descent method. The curve is driven to the image boundary of the elevator machine and the wire rope. The image is edge segmentation. The simulation results show that the Segmentation accuracy is obviously improved by the improved algorithm.(5) Study on the detection method of the fault defect degree of the elevator machine and the wire rope. The feature regions is extracted in the image of the elevator traction machine and the wire rope after segmentation by the Bayes cutout algorithm. The feature regions are disposed by the binarization, calculation the degree of the fault defect which the degree of the fault defect is quantified. The research results to improve the traction machine and wire rope detection,prediction of traction machine and wire rope fault, it has an important application prospect.