基于质心提取的线性拟合法，提出了一种自动化快速处理中阶梯光栅光谱仪谱图的方法，解决了中阶梯光栅光谱仪无法直接通过交叉色散形成的二维光谱图标定入射光波长的问题。根据中阶梯光栅光谱仪光谱图背景光的分布特点，采用背景扣除法处理光谱图，提高了后续光谱图光斑提取算法的运算效率。利用基于质心的曲线拟合中心定位算法，将光斑位置的提取精度提高到亚像素量级。建立了中阶梯光栅光谱仪理论谱图与实际二维谱图的自动匹配模型，并根据光栅的线性色散特性，采用插值法计算谱图还原模型中未列出的坐标位置处的波长。实验结果表明: 该算法对整个谱图的处理时间不超过3 s，波长提取精度达10-2 nm，不仅实现了中阶梯光栅光谱仪波长的自动化提取，还提高了波长提取精度。

An automatic algorithm to rapidly process the spectrogram of the echelle grating spectrometer was proposed based on the linear fitting method of centroid extraction, solving the problem that the echelle grating spectrometer can not calibrate wavelength of incident rays through the two-dimensional spectrogram which was formed by cross dispersion. The background correction method is adopted to process the spectrogram based on the distribution characteristics of background light of the spectrogram of echelle grating spectrometer, thereby the calculation efficiency of the optical spot extraction algorithm of the follow-up spectrogram of echelle grating spectrometer was improved. The curve fitting center positioning algorithm based on centroid was employed to improve the extraction precision of the optical spot position to the sub-pixel magnitude. The auto-match model of the theoretical spectrogram of echelle grating spectrometer and the actual two-dimensional spectrogram were established and the method of interpolation was adopted to calculate the wavelength at the coordinate position, which is not listed in the recovery model of the spectrogram based on the linear dispersion characteristics of optical grating. The result shows that: the processing time of the whole spectrogram by this algorithm does not exceed 3 s and the wavelength extraction precision is up to 10-2 nm, not only achieving the automatic wavelength extraction of the echelle grating spectrometer but also improving the wavelength extraction precision.