Seed vigor directly affects emergence rate, early emergence capacity, and seedling uniformity in ginseng. This study integrated pre-sowing laser treatment with hyperspectral prediction for rapid evaluation of ginseng seed vigor. Cracked-stage Mountain Cultivated Ginseng Seeds and American Ginseng Seeds were treated with four semiconductor laser conditions, including red light, blue light, red–blue 3:1, and red–blue 6:1, for 5, 15, and 30 min. Emergence rate, emergence energy, and emergence index were evaluated, and 350–2500 nm hyperspectral reflectance data were collected. Different preprocessing, feature selection, and machine learning methods were compared to construct an optimized support vector regression model for emergence rate prediction. The results showed that laser effects were seed-type-specific and parameter-dependent. Mountain Cultivated Ginseng Seeds responded strongly, with 5 min blue light treatment producing the best performance, achieving a 100.0% emergence rate, 40.4 percentage points higher than the control, with significantly higher emergence energy and emergence index. In American Ginseng Seeds, no treatment differed significantly from the control, although the 30 min red light treatment showed the highest emergence rate trend. Savitzky–Golay smoothing combined with the second derivative was the optimal preprocessing method. The Caterpillar Fungus Optimizer-based support vector regression (CFO-SVR) joint optimization model achieved the best prediction performance, with Rp2 of 0.9517, RMSEP of 2.4766, and MAEP of 1.0730. These results provide a feasible technical reference for pre-sowing treatment of ginseng seeds and rapid, non-destructive evaluation of emergence vigor.
Xue et al. (Thu,) studied this question.