Fault Detection Of Position Tracking Inside The Heating Furnace Based On Video Monitor

At present,the step-type heating furnace is indispensable to hot strip billet by rolling mill essential heat treatment equipment.For the first step of hot rolling production line,the step-type heating furnace is one of the most important equipment in the rolling production.Therefore,in the continuous fast automatic production,whether step-type heating furnace can work fine will directly affect the ability of the rolling production line's production.In the process of step-type type heating furnace production,steel's falling malfunction is one of the most difficult troubles to deal with.But when steels moving in the heating furnace,for some reason,the laser-aided feed determination or displacement sensing devices detect the steel for mistake,and this mistake can lead the error of tapping stroke's calculation,at last caused the steel's falling malfunction.Because the temperature in the heating furnace always up to one thousand degrees,and it is very difficult and time-consuming to be handled,and it influenced the production seriously.In this thesis,we used some knowledge of machine vision to process the video which was shot by the video monitor in the heating furnace in real-time.First of all,we processed every frame image which can effectively weaken the interference noise in the image by the means of image space filtering.And then we used the image histogram enhancement processing,to make the detail of the image become clear.Second,by comparing the traditional frame differential method,symmetric frame differential method and accumulation of frame differential method,and finally we selected the symmetrical frame difference method to get the targets.Due to the targets contained in the part of the walking beam movement area.Third,we through the area of morphological filter to remove the motion of the walking beam movement.Fourth,we processed the target area for thinning that can make the forefront of steel clear and avoid caused by straight line width is too wide and checked by mistake.Fifth,we used the Hough Transform to find out the most pixels in a straight line,and to determine the straight line for steel edge in a straight line.Sixth,we made up the camera imaging space transform defining coordinate system.And we established the relationship between the pixel coordinate system and world coordinate system.We calibrated the camera in 2-D plane.Compared with the traditional camera calibration,the in 2-D plane camera calibration has higher accuracy.The methods can be used to steels for position tracking in heating furnaceAt last we will measure the distance between steel edge and the line of laser-aided feed determination by the calibration of the camera.In that way,we can have a good position and tracking to the steel.Then we can get the fault diagnosis information,by the distance's difference between video and displacement sensing devices' measured value.When the difference is too large,the system will automatically alarm,and then the staff will finish this steel tracking by hand,to avoid steel's falling malfunction.