Research on seedling grading of Platycladus orientalis based on fuzzy C-means clustering algorithm improved by...

Platycladus orientalis seedlings possess significant ecological and economic value, but large individual quality variation often leads to poor post-transplant performance. Existing grading standards are oversimplified with ambiguous boundaries. In this study, eight-month-old P. orientalis seedlings were evaluated using in situ electrical impedance spectroscopy (EIS) for non-destructive root assessment, combined with three aboveground morphological parameters (stem diameter, plant height, branch number). These five indices served as reliable quality predictors. To overcome the limitations of conventional clustering methods-namely, the need to pre-specify the number of clusters and sensitivity to initial centres-we proposed a modified fuzzy C-means algorithm optimised by a genetic sparrow family search (GSFS, an improved version of the sparrow search algorithm, SSA), denoted as GSFS-FCM, which autonomously determines the optimal number of quality grades. The GSFS-FCM algorithm showed that six clusters minimised the loss function. From a practical production perspective, merging Clusters 4 and 5 into a single grade yields a five-grade classification that maintains biological utility and aligns with nursery practices. Compared with traditional methods, this five-grade system significantly enhanced post-transplant survival and growth performance, with intra-grade survival rates ranging from 85.0% to 97.8% (SE  1.2-2.1%) and clearly distinguishable inter-grade differences. This approach improves the scientific rigor of P. orientalis seedling grading and provides a novel framework for quality classification of other woody plant seedlings.