Simplified spectral fluctuation correction method of LIBS based on plasma image assistance

Accurate and precise quantitative analysis is the key for the application of laser-induced breakdown spectroscopy (LIBS). Although LIBS offers significant advantages for multi-element detection, its measurement accuracy and stability are often compromised by spectral fluctuations. To address this problem, image-assisted correction methods have developed as a powerful data-processing strategy. In this work, we propose a novel spectral fluctuation correction method named the spectral bias-error and plasma parameters stepwise correction method based on image assistance (SBEPPSC-IA), which integrates a two-stage process: plasma parameter correction and then followed by spectral bias-error correction based on image information. SBEPPSC-IA takes full advantage of both individual and population information from spectra and plasma images across the entire training set. This dual-stage approach effectively mitigates calibration curve nonlinearity and quantifies spectral bias-error, thereby enhancing LIBS analytical accuracy and measurement stability comprehensively. Experimental results show that SBEPPSC-IA significantly improves calibration for multiple elements, elevating the R² values for Mn, Ni, Cr, and Si to 0.9929, 0.9803, 0.9950, and 0.9626, respectively. All evaluation indexes for the training set were reduced. Notably, the prediction set exhibited average reductions of 39.51 % in root-mean-square-error (RMSE), 27.26 % in average relative error (ARE), and 77.90 % in average standard deviation (ASTD), with a particularly pronounced improvement in signal stability. An ablation study confirmed that initial nonlinearity mitigation based on image assistance is essential for the subsequent bias-error correction to achieve better quantitative analysis results. In conclusion, SBEPPSC-IA demonstrates considerable potential for improving the quantitative performance of LIBS technology.