Abstract:Laser-induced breakdown spectroscopy (LIBS) faces challenges due to large fluctuations in characteristic spectral line intensity and unstable detection results in the quantitative analysis of low-concentration heavy metals in water, limiting its practical application. This study investigated the correlation between plasma image features and the characteristic spectral line intensities of uranium. A model to correct spectral line intensities was established by integrating multiple plasma image features, and the U II 367.007 nm and U II 409.013 nm spectral line intensities were corrected. The results show that after correction, the average relative standard deviation of the spectral line intensities for U II 367.007 nm and U II 409.013 nm decreased from 13.21% and 13.13% to 6.58% and 7.36%, respectively, resulting in enhanced stability of the characteristic spectral line intensities. Based on this, quantitative analysis of uranium elements in water was conducted. After calibration, the linear fitting R2 of the calibration curves for U II 367.007 nm and U II 409.013 nm improved from 0.989 and 0.976 to 0.995 and 0.993, respectively. The detection limits decreased from 17.66 μg/L and 16.95 μg/L to 11.03 μg/L and 7.55 μg/L, respectively, effectively lowering the detection limit for low-concentration uranium in water and improving the accuracy of quantitative analysis.