Methods of data processing in multi-wavelength thermometry

Three kinds of methods for processing the data of the multi-wavelength pyrometer are presented and are named curve auto-search method, curve auto-regression method and neural network method. The experimental results indicate that the calculated temperature and the spectral emissivity compared with the true target temperature and spectral emissivity have significant deviation using the curve auto-search and the curve auto-regression methods. However, the calculated temperature and the spectral emissivity with higher accuracy can be obtained using the neural network method.