Scholarlink

Purpose <p>Prompt detection of leaf wetness with high accuracy is essential for effective disease management in crops, especially strawberries, where diseases like Botrytis fruit rot are major threats to yield. Traditional flat-plate sensors have limitations that reduce the reliability of advisory systems like the Strawberry Advisory System (SAS). This study aimed to overcome these challenges by developing an innovative and reliable vision-based detection system.</p> Methods <p>This study proposed a hybrid deep learning model combining a Convolutional Neural Network (CNN) with a Long Short-Term Memory (LSTM) network. A ConvLSTM model was designed to classify surface wetness as either dry or wet by analyzing sequences of images from a high-resolution camera. The CNN effectively extracted spatial features to detect minute droplets, while the LSTM captured temporal dependencies to understand patterns in droplet evolution. The model also employed Convolutional Block Attention Modules (CBAM) to enhance feature relevance and used Focal Loss to address the class imbalance problem.</p> Results <p>The model achieved 97% validation accuracy. This approach outperformed comparative models using a single CNN and the performance of traditional sensors.</p> Conclusion <p>The impact of this study is significant, as the model provides a highly accurate tool that can help farmers make informed decisions on moisture and dryness levels. This enables more precise timing of fungicide application, resulting in improved crop yields and reduced disease risk.</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

Time-Series Detection of Leaf Wetness Using a CNN-LSTM-Based Vision System in Strawberry Farming

  • 期刊论文

  • 发表日期：

    2026年3月13日

查看全文
  • Hemanth Reddy Sankaramaddi,
  • Won Suk Lee,
  • Natalia A. Peres

摘要

Purpose

Prompt detection of leaf wetness with high accuracy is essential for effective disease management in crops, especially strawberries, where diseases like Botrytis fruit rot are major threats to yield. Traditional flat-plate sensors have limitations that reduce the reliability of advisory systems like the Strawberry Advisory System (SAS). This study aimed to overcome these challenges by developing an innovative and reliable vision-based detection system.

Methods

This study proposed a hybrid deep learning model combining a Convolutional Neural Network (CNN) with a Long Short-Term Memory (LSTM) network. A ConvLSTM model was designed to classify surface wetness as either dry or wet by analyzing sequences of images from a high-resolution camera. The CNN effectively extracted spatial features to detect minute droplets, while the LSTM captured temporal dependencies to understand patterns in droplet evolution. The model also employed Convolutional Block Attention Modules (CBAM) to enhance feature relevance and used Focal Loss to address the class imbalance problem.

Results

The model achieved 97% validation accuracy. This approach outperformed comparative models using a single CNN and the performance of traditional sensors.

Conclusion

The impact of this study is significant, as the model provides a highly accurate tool that can help farmers make informed decisions on moisture and dryness levels. This enables more precise timing of fungicide application, resulting in improved crop yields and reduced disease risk.