| In the process of cement clinker firing,the kiln lining needs to bear the influence of high temperature and variable temperature,as well as the effect of erosion and chemical erosion of materials and air flow.The working environment is very bad,and the substrate insulation and protective materials often fall off and fail.Therefore,the workload of kiln lining maintenance and refractory brick laying is very large.The space of kiln body is very narrow.It emits many kinds of harmful gases inside.The ventilation condition is poor and the dust with high concentration is diffused.In the process of paving,workers need to enter the interior of the kiln,complete the erection of prevention devices,eliminate the insulation materials with substrate failure,clean the substrate bottom,pave new refractory bricks,check and repair the defective parts and other complex and fine work,with poor working environment,high labor intensity and high risk.This thesis analyzes the types of defects that affect the quality of refractory brick paving,and evaluates the quality of the paving through machine vision detection method.Without workers participating in the completion of this part of the work tasks,workers can be avoided from being hurt,human factors can be reduced from false detection,production efficiency can be improved,and enterprise income can be increased.The main research is as follows:(1)This thesis expounds the steps of determining the parameters of refractory brick in rotary kiln,selects the type of brick according to the national standard through the diameter and size of kiln body,and calculates the distribution of brick laying.The components of machine vision include hardware and software.According to the structure and principle of the hardware,different types of hardware are selected for combination,and the computer model and software are determined to form the software of quality inspection system.(2)According to the paving procedure,the steps of image acquisition are designed.Use Zhang’s calibration method to calibrate the camera,use MATLAB software to get the camera’s internal parameters,and eliminate the effects of lens distortion.Based on the image denoising of harmonic filter,the experiment shows that the method can remove the image noise of firebrick.By using smooth linear filter as image enhancement method,the edge features of the detected object in the image are obvious.(3)Based on the dynamic threshold detection segmentation method,the target features can be segmented,but the results are over segmented or under segmented,which cannot be used as defect detection targets.Therefore,the morphological segmentation technology is used again to perform morphological processing on the target with poor segmentation effect.Through the evaluation of image segmentation and segmentation results,it is proved that the segmentation method in this thesis is feasible.By using the method of shape feature extraction,the defect detection object of firebrick is obtained.(4)This thesis introduces the common defects of refractory brick paving,and classifies the defect characteristics.Based on the checking index of each test object,the algorithm of defect identification of firebrick is determined.According to the research of Bias classification method,based on the defect type,the feature of PBS state classification is extracted,and the simple Bias classifier model is established.In order to break through the influence of the special scene inside the rotary kiln on the collected image,this paper takes the detection of the quality of the refractory brick as the goal,aiming at the shortcomings of the existing traditional detection methods and the future trend development,carries out the research of the detection of the quality of the refractory brick based on the machine vision,and uses the effective image enhancement means and a variety of image segmentation methods,which is the response of the visual detection It provides the theoretical basis and practical basis. |
PDF Full Text Request