Scholarlink

<p>This study develops a machine vision-based approach to intelligently predict moisture content during <i>Lyophyllum decastes</i> (LD) drying by quantitatively characterizing morphological changes. Color, contour, and texture features were extracted from mushroom images, from which seven key features were selected using the Pearson correlation coefficient and random forest algorithm. Nine predictive models combining optimization algorithms and machine learning techniques were constructed. The genetic algorithm-optimized XGBoost model performed best, achieving a coefficient of determination (R²) of 0.89 and a mean absolute error (MAE) of 6.60. These findings provide a theoretical foundation for intelligent control of LD hot-air drying and contribute to the digital transformation of agricultural product processing.</p> Graphical abstract <p></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

Non-destructive prediction of moisture content in Lyophyllum decastes using image features and multi-optimization algorithms

  • 期刊论文

  • 发表日期：

    2025年10月14日

查看全文
  • Zhizhen Cao,
  • Tao Yu,
  • Chunyan Zhang,
  • Zhen Xu,
  • Huihui Zhao

摘要

This study develops a machine vision-based approach to intelligently predict moisture content during Lyophyllum decastes (LD) drying by quantitatively characterizing morphological changes. Color, contour, and texture features were extracted from mushroom images, from which seven key features were selected using the Pearson correlation coefficient and random forest algorithm. Nine predictive models combining optimization algorithms and machine learning techniques were constructed. The genetic algorithm-optimized XGBoost model performed best, achieving a coefficient of determination (R²) of 0.89 and a mean absolute error (MAE) of 6.60. These findings provide a theoretical foundation for intelligent control of LD hot-air drying and contribute to the digital transformation of agricultural product processing.

Graphical abstract