Objective  Spectral drift poses a unique challenge when observing moving targets using acousto-optic tunable filter (AOTF) spectrometers. Therefore, there is a need for an online spectral calibration method based on acousto-optic interaction. Utilizing the imaging position of the target spectrum and driving frequency, a reverse ray tracing model was constructed to achieve real-time calibration of the spectral data, ensuring stability and accuracy for subsequent detection, recognition, and tracking of the target. The developed mid-infrared AOTF spectral detection system with parallel entering light was employed for experimental verification. The results demonstrate that the correction accuracy of spectral drift is better than 4.45% for simulated moving targets with different fields of view. This improvement is beneficial for enhancing the application capabilities of spectral detection for moving targets.

  Methods  To address the issue of drift in mid-infrared AOTF spectral data under parallel light incidence conditions, a spectral calibration method based on the model of acousto-optic interaction is proposed. Initially, the principles of AOTF are briefly introduced, highlighting the use of a parallel light incidence structure to mitigate axial chromatic aberration in the mid-infrared band. The specific spectral calibration methods are then outlined. The reverse ray tracing method is employed, enabling the direct calculation of the spectrum from real-time image coordinates of the target and driving frequency (Fig.2). This involves computations such as refractive index calculation in three-dimensional space, coordinate system transformation, momentum matching, and more. Under ideal conditions, simulations of frequency drift in the image plane are conducted. The proposed spectral calibration method is experimentally validated using a self-developed prototype in the laboratory, and the performance parameters are presented in Tab.1. Importantly, as the proposed method is based on the acousto-optic mechanism model, no hardware modifications, such as changes to the optical structure, are required. This online spectral calculation method meets the application requirements for detecting the spectrum of moving targets.

  Results and Discussions   The validation experiment for dynamic spectral correction involves using a combination of a high-temperature blackbody and infrared filters as a narrowband light source. To simulate the moving target, sampling points are set (Fig.6). Initially, by selecting the region of interest (ROI), the frequency response at different positions of the target can be obtained. Experimental results indicate that the frequency response of the same target varies with different fields of view (Fig.7), leading to drift in the calculated target spectrum from the tuning curve. The method proposed in this article is then utilized to calibrate the target spectral response, resulting in a significant suppression of spectral drift before and after calibration. The spectral drift of the full field of view can be controlled within 4.45%. However, there are still some errors after spectral calibration. Firstly, the spectral full width at half maximum (FWHM) of AOTF varies with the field of view (FOV), which was not considered in the model. Secondly, there is a fitting error in the installation and adjustment of the system. Thirdly, random sampling errors occurred during the experimental process.

  Conclusions  The spectral data of aerial moving targets obtained by the AOTF spectral detection system may drift with FOV, affecting the extraction of spectral features and subsequently being unable to ensure stable tracking of the target. The spectral correction method based on the principle of acousto-optic interaction can perform real-time correction of the spectra of moving targets. Laboratory validation experiments have shown that the calibration method can effectively suppress spectral drift. After calibration, the accuracy of the spectral data cube of target can be ensured. The work of this article has certain significance for AOTF spectral detection from static targets to moving targets.