Accurate spectrogram restoration algorithm for an echelle spectrometer based on adaptive parameters

Abstract

The echelle spectrometer is a high-resolution spectrometer that can realize transient direct readings of a full spectrum. To improve the accuracy of the spectrogram restoration model in calibration, multiple-integral time fusion, and an improved adaptive-threshold centroid algorithm are used to overcome noise and improve the accuracy of calculating the light spot position. A seven-parameter pyramid-traversal method is proposed to optimize the parameters of the spectrogram restoration model. The deviation of the spectrogram model is significantly reduced after the parameters are optimized, and the deviation curve fluctuation becomes mild, which greatly improves the model's accuracy after curve fitting.The test results show that the accuracy of the spot position determination algorithm proposed in this paper is 0.1 pixels. In addition to this, the accuracy of the spectral restoration model is controlled within 0.3 pixels in a short-wave stage and 0.7 pixels in a long-wave stage. Compared with the traditional algorithm, the accuracy of spectrogram restoration is more than two times, and the spectral calibration time is less than 45 min.