Research On 3D Image Reconstruction For Complex Furnace Conditions In Blast Furnace

As an important smelting equipment in modern iron and steel industry,blast furnace provides liquid pig iron for a series of steel products.The distribution of charge level is one of the influencing factors of blast furnace operation.Reasonable charge level distribution can promote the reasonable development of gas flow,make sufficient chemical reaction in the blast furnace,reduce the waste of charge,and obtain high-yield and high-quality molten iron.Mastering the charge level distribution timely and accurately,combined with the optimal operation of distribution,can ensure the stability and efficiency of blast furnace production.Therefore,mastering and analyzing the charge level distribution plays an important role in actual production.In the complex environment inside the blast furnace,the monocular camera can only obtain the two-dimensional plane information when obtaining the charge level information of the blast furnace,and there are great defects in obtaining the charge level depth information.In this dissertation,two CCD industrial cameras are used to build a binocular vision system.The twodimensional charge level surface image is obtained based on the passive vision method.The threedimensional charge level surface is reconstructed through camera calibration,distortion correction,feature point extraction and matching,spatial point coordinate solution,triangulation technology and so on.Firstly,the charge level surface video information is collected through the binocular vision system,and the video information is transmitted to the computer through the video acquisition card to extract the charge level image.Secondly,an 8 * 10 chessboard calibration board is designed,15 chessboard images are taken,and the left and right cameras are calibrated by Zhang's calibration method to obtain the internal parameters and distortion parameters of the camera;Due to the radial distortion of the image caused by the camera lens,the distortion of the extracted charge level image is corrected.Then use the calibration tool stereo camera calibrator in MATLAB to calibrate the binocular,and obtain the external parameters between the two cameras,namely rotation matrix R and translation vector T.Then,the Scale Invariable Feature Transform(SIFT)algorithm with the characteristics of constant rotation scale,strong illumination adaptability and good robustness is used to extract and match the feature points of the charge level image,and then the Random Sample Consensus(RANSAC)algorithm is used to screen out the wrong feature point matching to obtain a more accurate matching.Finally,using the calibration results of monocular and binocular cameras and taking the left camera as the world coordinate system,the coordinates of the feature points of the charge level image in the three-dimensional space are calculated,that is,the three-dimensional coordinate points are recovered from the two-dimensional coordinate points.In order to obtain the intuitive spatial structure of the charge level distribution,the obtained threedimensional spatial coordinate points are regarded as a point set,and the Delaunay triangulation technology is used to restore the shape and depth of the charge level in the world coordinate system,so as to obtain the three-dimensional reconstruction results of the charge level.