红外CCD是通过采集高于绝对零度物体发出红外线并转化为图像来进行在线温度场测量的, 得到的图像与物体的本身有一定的对应关系。由于受到很多干扰因素的影响, 得到的温度场信息并不完全准确。造成的误差有很多, 包括光学系统的误差、CCD器件本身带来的误差, 还有就是红外辐射的干扰。对于前两者通过图像处理方法可以消除一部分误差, 而红外辐射的干扰随机性比较大, 利用一般的图像处理不能够完全解决。采用预埋热电偶的方法实时监测几个固定点的温度, 然后利用这几个点的温度与CCD测的温度相比较, 从而实时来校准系统的发射率。并利用修正后的系统发射率来计算被测物体的温度场分布, 为了防止温度只是由于突然改变的系统的发射率发生突变。由此设计了温度场实时校准数学模型和修正系统, 消除了部分温度场在线测量误差和突变, 实现了红外CCD在线温度场的准确测量。实验结果表明, 利用温度场实时修正系统来消除红外CCD温度场实时测量误差是可行的。

The temperature-field information collected by infrared CCD is usually interfered by various measurement errors, including optical system error, CCD error and infrared radiation interference. Errors induced by optical system and CCD can usually be controlled by image processing techniques. However, these techniques can hardly deal with infrared radiation interferences due to their high randomness. To solve this problem, a series of thermo-couples were preset to monitor temperatures at fixed points, which were used to calibrate the system emissivity. Real-time measurement of temperature field was thus achieved based on the modified system emissivity. The mathematical model and system of calibration are established to prevent the sudden change of temperature with system emissivity. Error and mutation of the on-line temperature field measurement are eliminated. The test results show the feasibility of the calibration system of on-line temperature-field measurement.