Scholarlink

<p>Strawberry surface disease and pest spot detection is of great significance in strawberry cultivation. Traditional methods for detecting strawberry surface diseases and pest spots suffer from low recognition accuracy and poor environmental adaptability. This study proposes an intelligent strawberry surface disease and pest spot detection model based on multi-feature fusion. The model uses Faster Region-based Convolutional Neural Network as the main framework. The model replaces the original feature extraction network with a&#xa0;50-layer Residual Network. It introduces Feature Pyramid Network to integrate feature information at different levels and to detect disease and pest spots at different scales. Simultaneously, it combines Convolutional Block Attention Module to build the strawberry surface disease and pest spot detection model. The proposed model achieves a&#xa0;detection accuracy of 97.07% for strawberry surface diseases and pest spots, with the Kappa coefficient reaching 0.9647. Both metrics outperform those of the baseline models. Application verification demonstrates that the model accurately detects these spots. Moreover, the classification accuracy for strawberry diseases and pest spots reaches 95.20%. The model can accurately detect strawberry surface disease and pest spots under both cloudy light and sunny light conditions. The model significantly improves the accuracy and environmental adaptability of strawberry surface disease and pest spot detection. It provides a&#xa0;new approach for modern strawberry cultivation and promotes the development of intelligent agriculture.</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

Intelligent Detection Method for Strawberry Surface Disease and Pest Spots Based on Multi-feature Fusion

  • 期刊论文

  • 发表日期：

    2026年7月30日

查看全文
  • Hang Wei,
  • Bing Li,
  • Yang Meng

摘要

Strawberry surface disease and pest spot detection is of great significance in strawberry cultivation. Traditional methods for detecting strawberry surface diseases and pest spots suffer from low recognition accuracy and poor environmental adaptability. This study proposes an intelligent strawberry surface disease and pest spot detection model based on multi-feature fusion. The model uses Faster Region-based Convolutional Neural Network as the main framework. The model replaces the original feature extraction network with a 50-layer Residual Network. It introduces Feature Pyramid Network to integrate feature information at different levels and to detect disease and pest spots at different scales. Simultaneously, it combines Convolutional Block Attention Module to build the strawberry surface disease and pest spot detection model. The proposed model achieves a detection accuracy of 97.07% for strawberry surface diseases and pest spots, with the Kappa coefficient reaching 0.9647. Both metrics outperform those of the baseline models. Application verification demonstrates that the model accurately detects these spots. Moreover, the classification accuracy for strawberry diseases and pest spots reaches 95.20%. The model can accurately detect strawberry surface disease and pest spots under both cloudy light and sunny light conditions. The model significantly improves the accuracy and environmental adaptability of strawberry surface disease and pest spot detection. It provides a new approach for modern strawberry cultivation and promotes the development of intelligent agriculture.