Scholarlink

<p>Mango leaf diseases have been among the most significant scourges in recent years, causing major economic losses to the mango yield. Conventional approaches to disease indication and categorization are frequently overly lengthy and need a lot of manpower. In this paper, we introduce DeepLeafNet, a unique, optimized deep-learning framework that segments and categorizes mango leaf diseases. The DeepLeafNet network combines cutting-edge deep learning models with optimization algorithms to provide excellent categorization and detection outcomes. We used U-Net to segment the image data. Then, NAS-Net was introduced to classify and predict the leaf disease categories. We employed a genetic algorithm to optimize the classification model hyperparameters. The approach used a heuristic search approach to tune the relevant hyperparameters, which, in turn, increased the model’s performance. The proposed model reached an accuracy of 98.56% in the classification of 9 distinct classes of mango leaf diseases, ensuring model performance novelty. The data set used (<a href="https://drive.google.com/drive/folders/1NjqembK757cjrAVI9UeDPxGHyyv1TBc?usp=drive_link">https://drive.google.com/drive/folders/1NjqembK757cjrAVI9UeDPxGHyyv1TBc?usp=drive_link</a>) with total images of 4873 which is collected from Kaggle database and horticulture college of Mysuru, Karnataka, India.</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

DeepLeafNet: A Novel Deep Learning Approach for Mango Leaf Disease Segmentation and Classification with Optimization Algorithms

  • 期刊论文

  • 发表日期：

    2025年11月8日

查看全文
  • C. P. Vijay,
  • K. Pushpalatha,
  • D. R. Janardhana,
  • H. S. Ranjan Kumar,
  • Kiran Puttegowda

摘要

Mango leaf diseases have been among the most significant scourges in recent years, causing major economic losses to the mango yield. Conventional approaches to disease indication and categorization are frequently overly lengthy and need a lot of manpower. In this paper, we introduce DeepLeafNet, a unique, optimized deep-learning framework that segments and categorizes mango leaf diseases. The DeepLeafNet network combines cutting-edge deep learning models with optimization algorithms to provide excellent categorization and detection outcomes. We used U-Net to segment the image data. Then, NAS-Net was introduced to classify and predict the leaf disease categories. We employed a genetic algorithm to optimize the classification model hyperparameters. The approach used a heuristic search approach to tune the relevant hyperparameters, which, in turn, increased the model’s performance. The proposed model reached an accuracy of 98.56% in the classification of 9 distinct classes of mango leaf diseases, ensuring model performance novelty. The data set used (https://drive.google.com/drive/folders/1NjqembK757cjrAVI9UeDPxGHyyv1TBc?usp=drive_link) with total images of 4873 which is collected from Kaggle database and horticulture college of Mysuru, Karnataka, India.