Research On 3D Temperature Field Reconstruction Of Material Surface Based On Infrared-Visible Light Image Mapping

In the industrial production process,the acquisition of the 3D temperature field of the material surface is of great significance for monitoring the degree of material reaction and monitoring the operation status of the equipment.In this thesis,the 3D temperature field of the material surface is reconstructed by combining infrared thermography temperature measurement method and visible light 3D reconstruction method with the simulated blast furnace material surface as the research object.The main research results and innovation points of the thesis are as follows.(1)Aiming at the problem that the thermal image texture is fuzzy,there is much noise and it is difficult to calibrate the thermal imager,a geometric calibration method of infrared thermal imager based on twostage enhancement of thermal image is proposed.A new calibration plate based on different emissivity materials is designed to improve the contrast of the calibration image and strengthen the texture edge before thermal imaging;A thermal image enhancement method based on texture denoising and multi-scale structure enhancement is proposed to reduce image noise and enhance texture details after thermal imaging,effectively improve the effect of calibrated image,and complete the calibration of infrared thermal imager.(2)Aiming at the problem that it is difficult to accurately obtain the three-dimensional shape of the material surface,a three-dimensional shape reconstruction method based on binocular stereo vision is proposed.The three-dimensional scene of the material surface is reconstructed with a dual visible camera.Based on the stereo matching algorithm integrating census transform and gradient transform,the correct matching of the material surface image with weak texture and discontinuous parallax is realized,and the material surface parallax map is obtained from the matching points,and then the three-dimensional morphology and spatial coordinates of the material surface are calculated and restored.(3)Aiming at the problem that it is difficult to accurately match the temperature point of the material surface thermal image with the threedimensional scene,a two-stage mapping method of temperature space point cloud is proposed.In the first stage,the rough mapping relationship between temperature pixels and spatial point cloud is established by using the method of re projection of thermal image to three-dimensional space;In the second stage,according to the infrared visible image feature matching point pair,the visible image feature points are re projected to the three-dimensional space,the fine mapping spatial point cloud of a small number of infrared feature points is established,the closed polyhedron is constructed with the rough mapping point cloud of feature points,and the centroid deviation between them is solved to correct a large number of rough mapping results,so as to realize the overall accurate mapping between temperature points and spatial point cloud.Experiments show that this method realizes the accurate correlation mapping between material surface temperature and spatial point cloud,and can effectively reconstruct the 3D temperature field of material surface.