Research On Temperature Field Measurement System Based On Digital Image Processing Technology

High temperature measurement technology is closely related to national defense,military,scientific experiments and industrial and agricultural production,which is essential key technology in the modern scientific development.However,most of the traditional high temperature measurement methods use contacts measurement method.There are low temperature measurement efficiency,can not get the temperature field distribution and other issues.The high temperature field measurement technology based on digital image processing is a new non-contact temperature measurement method.It combines the digital image processing technology and the radiation temperature measurement technology to measure the high temperature field quickly and accurately.Compared to the traditional temperature measurement method,temperature measurement efficiency and temperature measurement accuracy have been greatly improved.Therefore,based on digital image processing temperature field measurement technology has become the focus of the industry.In this thesis,the basic principles and methods of radiation temperature measurement,the working principle of color CCD and the principle of color CCD temperature measurement are studied in detail.The color CCD(Charge Coupled Device)image sensor and digital image processing technology are used to design a high temperature field measurement system based on digital image processing.Firstly,the hardware of the measuring system is designed,and then the software system is designed.The functions of image acquisition,image preprocessing(image filtering and image segmentation),temperature field calculation and pseudo color display are completed.At last,the calibration experiment of the temperature measurement system is carried out,and the experiment of temperature measurement is carried out.The experimental data are analyzed by the least squares method and BP neural network respectively,and the accurate temperature measurement model is established and the range of temperature measurement and the accuracy of temperature measurement are determined.The experimental results show that BP neural network algorithm with smaller error,better error stability.