Research On Temperature Measurement Based On Fusion Of Billet Surface Image And Model Data

Billet surface temperature is one of the most important parameters in the steel industry. It can improve the billet quality, reduce the consumption of energy, and be used to implement closed-loop optimization control. This thesis detects the position between the secondary cooling zone export and the straightening machine, because of the adverse circumstances and oxide skin, image sensor can not obtain the real temperature information of the slab surface under oxide skin. Based on the former measurement system, this thesis, mainly study the issues about image non-uniformity, casting speed measurements, accuracy of fusion, in the method based on data fusion of image and model measurement.The main contents in this thesis are as followings:(1)For the nonuniformity of the slab surface image,calibration-based and scene-based (adaptive) nonuniformity correction (NUC) techniques are both studied in (his thesis.we discuss the origin of nonuniformity and compare the effect of different NUC methods.Then an adaptive non-uniformity correction(NUC) algorithm based on the billet moving statistical analysis is proposed. The method proposed can overcome the periodic calibration problem, and provide supports for the long-range stable operation of the temperature measurement system.(2)In the aspect of temperature field Calculation, the effect of casting speed is larger. To avoid these effects,digital particle image velocimetry(DPIV) technology is introduced to the casting speed detection.Take the surface oxide skin as tracing particles use cross-correlation technique to realize the casting speed real time measurement, Thus reduce the fluctuation of the model results caused by casting speed changes,reduce the system coupling and improve the system stability.(3) In the course of the temperature field calculation, Temperature model and image data are not well matched, To solve this problem, a fast correction method for Temperature curve based on the principle of CCD imaging is proposed.First, Calibrate the model curve Then matching with the information calculated by image data. Thereby reducing the errors of the matching process,reduce the probability of failure model matching and improve the accuracy of temperature measurement system.(4) This algorithm is verified by using the image processing software of Visual C++,OpenCV and Matlab in this thesis.The simulation and experiment results show that this algorithm is very effective.