Scholarlink

<p>One of the most significant factors that negatively impacts potato plant development is the occurrence of early and late blight diseases on the leaves. While experts can detect these diseases, the average farmer lacks sufficient knowledge. Therefore, the diagnosis process can take longer, and the accuracy rate can be reduced. This study examines studies on plant diseases where traditional methods are inadequate and proposes a novel model using Artificial Intelligence (AI) algorithms for the detection of diseases on potato leaves. Optimization was applied to select the most effective features for disease detection, thus ensuring accurate, effective, and high-performance disease detection. The Potato Leaf Dataset (PLD) from Pakistan was used as the dataset. The PLD contains a total of 4072 images: 1020 healthy, 1628 early blight, and 1424 late blight images. The images were separated into RGB color channels, and the features of each channel were extracted using the SqueezeNet CNN model. Then, the most significant features were selected using the Harris-Hawk Optimization (HHO) algorithm, and classification was performed using the Support Vector Machines (SVM) algorithm. The proposed model also improved its classification accuracy by applying the data augmentation technique, achieving an overall accuracy rate of 97.87%.</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

Optimizing images of normal and diseased potato leaves in RGB color space with squeezenet and harris hawk algorithms and determination by SVM

  • 期刊论文

  • 发表日期：

    2026年4月12日

查看全文
  • Gaffari Celik

摘要

One of the most significant factors that negatively impacts potato plant development is the occurrence of early and late blight diseases on the leaves. While experts can detect these diseases, the average farmer lacks sufficient knowledge. Therefore, the diagnosis process can take longer, and the accuracy rate can be reduced. This study examines studies on plant diseases where traditional methods are inadequate and proposes a novel model using Artificial Intelligence (AI) algorithms for the detection of diseases on potato leaves. Optimization was applied to select the most effective features for disease detection, thus ensuring accurate, effective, and high-performance disease detection. The Potato Leaf Dataset (PLD) from Pakistan was used as the dataset. The PLD contains a total of 4072 images: 1020 healthy, 1628 early blight, and 1424 late blight images. The images were separated into RGB color channels, and the features of each channel were extracted using the SqueezeNet CNN model. Then, the most significant features were selected using the Harris-Hawk Optimization (HHO) algorithm, and classification was performed using the Support Vector Machines (SVM) algorithm. The proposed model also improved its classification accuracy by applying the data augmentation technique, achieving an overall accuracy rate of 97.87%.