Design And Realization Of Real-Time Online 3D Reconstruction System Of Temperature Field

At present,the 3D model has been widely used in plant visualization,medical,construction and equipment manufacturing and many other fields,But these three-dimensional models only have the geometric information of the object and the surface color information.In addition,the infrared temperature information is added to the visible light and mapped to the 3D model,which makes the application of 3D model visualization more valuable,Because the temperature information can reflect the information of the internal temperature field change of the object,it has a huge application prospect for plant equipment operation evaluation and fault diagnosis,medical analysis,building energy audit,urban planning,environmental protection,resource searching and fire rescue,While in the field of machine vision for the semantic recognition of objects is of great significance.The current 2D thermal imaging technology is the problem lies in the characteristics of the optical lens imaging plane caused by the limitations of large display range,making only the use of pictures or video equipment evaluation is not intuitive.At present,the application of two-dimensional infrared images is extensive,while the three-dimensional rendering method is still in the blank stage.In order to solve the above problems,we combine infrared image and three-dimensional model in equipment monitoring,and add infrared temperature information on the visible light and map to the three-dimensional model,which makes the application of infrared graphics in equipment detection more valuable.Compared with the traditional 2D thermal imaging technology,3D thermal imaging technology can produce three-dimensional object temperature field model,overcoming the problem that the display range is insufficient due to the characteristics of the optical lens planar imaging,which provides the basic model carrier for the numerical calculation of the temperature field and visual display.And 3D thermal imaging technology for the production operation optimization and safety production provides a new means of monitoring,which is a new complex industrial production process visualization technology.To this end,this thesis presents a fusion of infrared information can be three-dimensional model of the reconstruction process.The challenge of this method is to combine the infrared temperature image with the 3D model in a way that is not simply a replacement,but to add one dimension to the obtained model and to incorporate information from that dimension into the geometric model of the reconstruction process.The system is implemented as follows:Firstly,According to the characteristics of infrared temperature data,In the choice of sensors,the author conducted a large number of research,finally chose the RGB-D camera ASUS XTion PRO and the infrared camera Optris 1450 as the front-end input,and the author has done the time and space calibration,then use of three-dimensional printing technology to build a hardware platform.Secondly,According to the current international advanced real-time three-dimensional positioning and reconstruction methods.Combining the complementary information of the infrared temperature information and the surface color information in different environments,the pose of the camera is estimated by weighting the infrared image and the RGB image according to the different environment.The experimental results show that this method can obtain better pose estimation effect in the dark environment than using the surface color information alone.Finally,The two RGB images are simultaneously involved in the geometric model reconstruction of the object.In the reconstruction process,the visualization software is developed to provide users with the real-time reconstruction process of the switchable infrared model and the RGB model,in order to analyze and evaluate the reconstruction process.Finally,the output of the system is a compact 3D model,can be achieved in the same model surface display RGB data and infrared temperature data.