基于数据融合策略的灵芝孢子粉破壁率快速检测方法

A rapid detection method for wall breaking rate of Ganoderma lucidum spore powder based on data fusion strategy

关键词:  灵芝孢子粉  破壁率  近红外光谱  色度检测  数据融合

Key words:  ganoderma lucidum spore powder  sporoderm-broken rate  near-infrared spectroscopy  colorimetric detection  data fusion

摘要:

目的 针对现行灵芝孢子粉破壁率检测方法（血球计数法）存在检测耗时较长且依赖人工判断易引入误差等问题，提出一种基于近红外光谱与色度数据融合的快速检测方法，以实现对破壁过程中破壁率的高效、无损评估，从而保障产品质量。方法 首先，采集不同破壁率灵芝孢子粉样品的近红外光谱和色度数据，其次，分别对两类单一数据运用多种预处理方法建立初步定量分析模型；在此基础上，系统比较基于低级融合、中级融合和高级融合策略所构建的数据融合定量模型性能，最终筛选出预测效果最优的破壁率快速检测模型，为灵芝孢子粉破壁过程的实时监控提供技术支持。结果 最优破壁率预测模型为：LLF-SG-PLSR模型，其在测试集上预测破壁率的均方根误差（Root Mean Squared Error of Prediction，RMSEP）为0.0689，决定系数（R^2）为0.9420，性能与偏差之比（Ratio of Performance to Deviation，RPD）为4.154，预测值和真实值的误差较小。结论 优选模型可以实现对灵芝孢子粉破壁过程破壁率的快速、无损检测，为中药领域数据驱动的过程控制提供依据。

Abstract:

Aiming at the limitations of the current method (hemocytometry) for determining the sporoderm-broken rate of Ganoderma lucidum spore powder, such as time-consuming procedures and human error, this study proposes a rapid method based on the data fusion of near-infrared spectroscopy and colorimetry. This approach enables efficient and non-destructive assessment of the sporoderm-broken rate during the processing, thereby contributing to enhanced product quality. METHODS First, near-infrared (NIR) spectra and colorimetric data of Ganoderma lucidum spore powder samples with different sporoderm-broken rate were collected. Subsequently, preliminary quantitative models were established by applying various preprocessing methods to each type of single data individually. Then, the performance of quantitative models based on low-level, mid-level, and high-level data fusion strategies was systematically compared. Finally, the optimal model for rapid sporoderm-broken rate prediction was selected to provide technical support for the real-time monitoring of the sporoderm-broken rate process. RESULTS The LLF-SG-PLSR model was identified as the best-performing predictor, yielding an RMSEP of 0.0689, R^2 of 0.9420, and RPD of 4.154, confirming its precision in sporoderm-broken rate evaluation. CONCLUSION The optimized model enables rapid and non-destructive detection of the sporoderm-broken rate during Ganoderma lucidum spore powder processing, providing a data-driven approach for quality control in traditional Chinese medicine production.