利用小波变换技术对辐射信号进行预处理和去噪,实现了辐射信号的高精度检测;进而,设计了一种组合神经网络发射率模型,依据测温仪测得的亮温可以高精度地辨识出被测目标的发射率模型,这种建模方法不需给定发射率固定的样本数据,给出具体理论推导过程.仿真结果表明,此方法是获知连续光谱发射率及真温的一种有效的方法.

By using the technique of wavelet transform for pretreatment and flitting the weak radiation signal has been accurately detected. Moreover, a new method for modeling the emissivity and wavelength of a multiwavelength pyrometer has been proposed and developed for obtaining accurately spectral emissivity and wavelength. In this method, it was unnecessary to give sample data of fixed emissivity model. Then, the detailed aglorithem was described and the simulation was done for checking the validity of algorithm.