宽动态RAW图像辐射测温试验研究

Experimental Research on Temperature Measurement by Wide Dynamic RAW Image Processing

获取高精度的原始图像数据是开展辐射测温研究的重要基础,而这受制于测量元件的动态响应范围。利用RAW图像的线性感光特性从图像中提取红、绿基色谱色值,并与相机曝光时间作比值后得到比色测温所需的多光谱检测信号,该方法能够消除相机快门、白平衡设置对辐射测温的影响,提高标定精度,拓宽动态响应范围。通过在一台水泥回转窑上开展试验研究,对不同工况下的火焰温度和黑度分布进行了在线测量,并利用红外高温计对辐射图像测温精度进行了验证,测温误差小于3%。该研究将显著提高辐射图像处理精度,有助于改善现有的基于彩色CCD的火焰温度场可视化技术。

Dynamic response range of the measuring element becomes to restrict the development of radiation thermometry technology. In our research, digital CCD camera and Gige network transmission technology are used to record RAW image, in which the ratio of spectral color and camera exposure time is proportional to the incident monochromatic radiation intensity, and the proportional coefficient can be determined by blackbody furnace calibration, thus the ratio of spectral color and exposure time in red and green wavelength are extracted from RAW image to measure temperature by colorimetric method. Compared with the existing radiation image thermometry technology, this method can eliminate the influence of camera shutter and white balance settings, improve the calibration accuracy and broaden the scope of dynamic response. Experiment was carried out on a cement kiln, in which the flame image temperature and emissivity in different case were measured continuously, and then the measurement error of the radiation image thermometry method mentioned above was verified by infrared pyrometer, which was less than 3%. This research can abstract highly accurate monochromatic radiation intensity distribution from flame images, which will improve the existing flame temperature visualization method based on color CCD camera.