Scholarlink

Agriculture plays a vital role in the economic success of every country, serving as a primary source of government revenue and contributing significantly to national income. Automating agricultural processes can help minimize resource consumption and improve the quality of food. Fruits are grown all throughout the world, with each having its own specific type. Fruit types are classified mostly based on external characteristics such as color, length, diameter, and shape. These visual traits play an important role in determining fruit type. However, correctly recognizing fruit kinds based on appearance frequently necessitates skill, which can be time-consuming and labor-intensive. The purpose of this chapter is to classify seven various types of date fruit (BERHI, DEGLET, ROTANA, SAFAVI, SOGAY, IRAQI and DOKOL) using Random Forest with Sequential Feature Selector for feature selection, Recurrent Neural Networks (RNN) and Long Short-Term Memory (LSTM) approaches for classification. The results demonstrated that the proposed RNN model achieved an accuracy of 93.47%, along with balanced performance metrics including an F1-score of 93.48, recall of 93.47, and precision of 93.49. Comparative analysis with existing methods, including Optimized Ensemble and Logistic Regression approaches, demonstrated that our model significantly outperforms previous techniques in terms of classification accuracy and reliability. These results highlight the effectiveness of combining advanced computational techniques for accurate fruit classification, contributing to the ongoing efforts in agricultural automation.

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

Sustainable Agriculture Through Automated Classification of Date Fruit Varieties Using Deep Learning Techniques

  • 章节

  • 发表日期：

    2026年

查看全文
  • Mamdouh Gomaa,
  • Heba Mamdouh Farghaly,
  • Ashraf Darwish,
  • Aboul Ella Hassanien

摘要

Agriculture plays a vital role in the economic success of every country, serving as a primary source of government revenue and contributing significantly to national income. Automating agricultural processes can help minimize resource consumption and improve the quality of food. Fruits are grown all throughout the world, with each having its own specific type. Fruit types are classified mostly based on external characteristics such as color, length, diameter, and shape. These visual traits play an important role in determining fruit type. However, correctly recognizing fruit kinds based on appearance frequently necessitates skill, which can be time-consuming and labor-intensive. The purpose of this chapter is to classify seven various types of date fruit (BERHI, DEGLET, ROTANA, SAFAVI, SOGAY, IRAQI and DOKOL) using Random Forest with Sequential Feature Selector for feature selection, Recurrent Neural Networks (RNN) and Long Short-Term Memory (LSTM) approaches for classification. The results demonstrated that the proposed RNN model achieved an accuracy of 93.47%, along with balanced performance metrics including an F1-score of 93.48, recall of 93.47, and precision of 93.49. Comparative analysis with existing methods, including Optimized Ensemble and Logistic Regression approaches, demonstrated that our model significantly outperforms previous techniques in terms of classification accuracy and reliability. These results highlight the effectiveness of combining advanced computational techniques for accurate fruit classification, contributing to the ongoing efforts in agricultural automation.