The application discloses a factory intelligent breeding method and
system for high-quality
aralia elegans seedlings, which realizes the whole life cycle intelligent breeding of
aralia elegans
seedling detection, evaluation, sorting, cultivation and re-detection through
aralia elegans
seedling cultivation, 360-degree multi-spectrum high-
throughput phenotype collection, Swin-Unet model semantic segmentation based on a four-element composite
loss function, leaf area,
chlorophyll index,
anthocyanin index calculation and 3D
root volume reconstruction,
quality score generation and closed-loop regulation of
tissue culture environment, solves the problems of
subjective quality evaluation, high low-quality
seedling waste rate, lagging
environmental regulation and inaccurate key organ quantification in traditional aralia elegans seedling breeding, improves the output rate,
resource utilization rate and
automation level of high-quality aralia elegans seedlings, and provides a complete technical solution for the standardized and intelligent breeding of aralia elegans and other woody
medicinal plants.