Scholarlink

<p>Hot air drying is a widely used method for the postharvest preservation of agricultural products. The accurate evaluation of dry matter content and choosing appropriate drying times are crucial for grading drying, reasonable packaging and storage. Until now, no relevant investigations have been reported in this area. In this study, a non-destructive prediction method for <i>Collybia radicata</i> using phenotype and machine learning, which involve a novel <i>Collybia radicata</i> phenotype collector (CRPC) method yielding 10 parameters, is proposed. The correlations between the morphological parameters and both the drying time and dry matter content under different temperatures were subsequently calculated through correlation analysis. Finally, the optimal input parameters of different regression models were screened using the maximum relevance minimum redundancy (MRMR) algorithm, and experiments were carried out at different temperatures. A comprehensive analysis was conducted on a total of 425 samples, utilizing the minimum redundancy maximum relevance (MRMR) method for feature selection. The experimental results demonstrated that for predicting both drying time and dry matter content, the neural network model with 3 hidden layers and 100–100–100 hidden nodes is the best option across all three temperatures. For dry matter content, the mushroom area (MA), stalk max thickness (SMT), cap diameter (CD) and stalk average thickness (SAT) were selected as the input parameters, and the R<sup>2</sup> values reached 0.9788 at 55&#xa0;°C, 0.9919 at 60&#xa0;°C, and 0.9890 at 65&#xa0;°C. For drying time, angle (KA), stalk max thickness (SMT), and cap area (CA) were selected as the input parameters for the best effect, and the R<sup>2</sup> values reached 0.9773 at 55&#xa0;°C, 0.9911 at 60&#xa0;°C, and 0.9914 at 65&#xa0;°C. Phenotypic parameters were automatically acquired in this study and correlated with drying characteristics, thereby providing valuable insights for optimising the drying process of <i>Collybia radicata</i> and showcasing its potential applicability to other agricultural products.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Evaluation of dry matter content and drying time by phenotype : a case study of collybia radicata

  • 期刊论文

  • 发表日期：

    2025年10月16日

查看全文
  • Hua Yin,
  • Jiamin Zhao,
  • Haijing Hu,
  • Jianpin Zhou,
  • Lu Yuan,
  • Hongruo Wang,
  • Yinglong Wang

摘要

Hot air drying is a widely used method for the postharvest preservation of agricultural products. The accurate evaluation of dry matter content and choosing appropriate drying times are crucial for grading drying, reasonable packaging and storage. Until now, no relevant investigations have been reported in this area. In this study, a non-destructive prediction method for Collybia radicata using phenotype and machine learning, which involve a novel Collybia radicata phenotype collector (CRPC) method yielding 10 parameters, is proposed. The correlations between the morphological parameters and both the drying time and dry matter content under different temperatures were subsequently calculated through correlation analysis. Finally, the optimal input parameters of different regression models were screened using the maximum relevance minimum redundancy (MRMR) algorithm, and experiments were carried out at different temperatures. A comprehensive analysis was conducted on a total of 425 samples, utilizing the minimum redundancy maximum relevance (MRMR) method for feature selection. The experimental results demonstrated that for predicting both drying time and dry matter content, the neural network model with 3 hidden layers and 100–100–100 hidden nodes is the best option across all three temperatures. For dry matter content, the mushroom area (MA), stalk max thickness (SMT), cap diameter (CD) and stalk average thickness (SAT) were selected as the input parameters, and the R2 values reached 0.9788 at 55 °C, 0.9919 at 60 °C, and 0.9890 at 65 °C. For drying time, angle (KA), stalk max thickness (SMT), and cap area (CA) were selected as the input parameters for the best effect, and the R2 values reached 0.9773 at 55 °C, 0.9911 at 60 °C, and 0.9914 at 65 °C. Phenotypic parameters were automatically acquired in this study and correlated with drying characteristics, thereby providing valuable insights for optimising the drying process of Collybia radicata and showcasing its potential applicability to other agricultural products.