Research On Temperature Measurement By Two-color Method Based On Image Processing

Temperature measurement is closely related to national defense,military,scientific research,industrial production and other activities.Different temperature measurement scenarios have different requirements for the accuracy and displayment of temperature measurement.Based on two-color pyrometry and digital image processing technology,the following works have been done in this paper:(1)A dual-path optical spectroscopy system was designed and built.The theory of temperature measurement by two-color pyrometry is based on the ratio of the intensity of monochrome light of two wavelengths to realize temperature measurement.Each channel of the color CCD collects a light intensity in a wide spectral range,and directly shooting the flame by CCD camera will introduce errors.In the dual-path optical spectroscopy system designed in this paper,two monochromatic filters with center wavelengths of 637 nm and 541 nm respectively were used to enable the CCD to receive two narrow-band monochromatic lights,which was more in line with the two-color temperature measurement theory.The combination of the reflective mirror and the beam splitter enabled the two-way filtered monochromatic light to be collected into one channel and received by a CCD camera,avoiding the post-matching of hightemperature targets in the two monochromatic images.(2)A monochrome gray value compensation model was designed.Since monochromatic light of 637 nm and 541 nm is interlaced in the R and G channels of the CCD,experiments were carried out in order to accurately extract the light intensity signals of the two wavelengths.A compensation model for the two channel signals is established,and the compensation coefficient is determined.The gray values of the R and G channels after compensation are accurately reflected respectively to the monochromatic light intensity of 637 nm and 541 nm,which further improves the accuracy of two-color temperature measurement.(3)The parameters such as white balance and exposure time in the temperature measurement system built in this paper were tested and selected.An iterative algorithm for calculating the flame temperature was proposed.The initial value of the temperature is determined by the gray value of the G channel,and then the iterative operation is performed by using the relationship model between the temperature measurement coefficient and the temperature to achieve accurate temperature measurement.The experimental comparison with the traditional method proved that the method improved the temperature measurement accuracy,and the relative error of the temperature measurement was 0.55%.The same flame was measured with different exposure times,and the flame images at various exposure times were collected,and the pixels with the gray value in a good range were image-fused to obtain a more accurate temperature distribution map of the flame.