Due to its non-invasive nature, high sensitivity and selectivity, the laser-induced fluorescence (LIF) spectroscopy has become an essential tool for the qualitative and quantitative diagnosing of plasma parameters. To obtain the distribution of plasma parameters, deconvolution technique must be used to separate the pure Doppler effect from the raw LIF spectrum. However, the presence of inherent noise in the received signal can result in the amplification of high-frequency noise, thereby causing deviations from accurate solutions. To solve this problem, we propose an enhanced deconvolution algorithm based on Richardson-Lucy (R-L) method with the adaptive gain coefficient, which effectively mitigates the ringing artifacts and noise amplification existing in the conventional R-L algorithm, and discuss the influence of parameter selection of gain coefficient. We use the numerical reconstructed LIF spectra at different noise levels as well as the measured LIF spectra of Xe to test the effectiveness of the proposed algorithm. The results show that the enhanced algorithm can significantly suppress the ringing effect and noise, and at the same time recover the Doppler lineshape well. Moreover, this approach presents a novel perspective for spectrum deconvolution algorithms such as infrared spectrum, Raman spectrum X-ray spectrum, etc.
|