Position Monitor Of Billet At Reheating Furnace Discharge Point Based On Image Processing

Reheating furnace is important equipment in the production line of steel rolling.Compared with the previous pusher-type furnace,the widely applied walking beam furnace recently has a considerable amount of merits.It can process raw billet of various kinds productively and flexibly,avoiding scratching on the surface of the slab at the same time.Nevertheless,in the daily operation of industrial furnace,it is a commonly seen situation that operation problem in processing the billet cannot be noticed timely which leads to a failure in steel tapping smoothly.The billet may even fall off from the walking beam to the bottom of the furnace if the problem deteriorates.Walking beam furnace,however,is flow-type production equipment which has a consecutive ability of production at a fast speed.Undoubtedly,problems in operation are sure to greatly influence the whole production process.This thesis majors in applying a real time monitoring of the position of the billet based on image processing methods.Due to the complicated structure inside the furnace,the quality of images gathered from the furnace is relatively low,which is also the major issue discussed in this thesis.Therefore,the choice of image processing method is of vital importance in this process.This thesis firstly conduct comparative experiment based on classical edge detection operator.Afterwards,further researches are made based on Canny edge detector which boasts with good continuity and noise immunity.On the basis of traditional Canny edge detector and self-adaptive Canny edge detector,a method that can calculate high and low threshold values automatically by the program is proposed to extract the front edge of the billet.With the help of polymerization continuation algorithm,unnecessary lines are removed.It can acquire a complete edge of the billet which can extend to the lower boundary of the image.In order to conduct real-time monitor after getting the front edge of the billet,the thesis applies the idea of first measuring the distance then modeling,and chooses the last step of the walking beam,i.e.the step to furnace door,to be the object of experiment.It employs the method of partial least square regression(PLS)modeling to find out the connection between the real-time position of the billet and the distance from front edge of the billet to bricklaying line on the delivery side outside furnace door after calculating(shown in nonlinear regression equation).Finally,via conducting experiments,it intends to testify the result of this method can meet the accuracy requirement.Therefore,the purpose of acquiring the exact distance from the billet to outside the furnace door by reading the position of the billet in the image can be achieved.