June 21, 2024 Thesis Open Access

PAULOS MEKONNEN SISAY

JSON-LD (schema.org) Export

{
  "description": "<p>Ethiopia possesses tremendous potential for the development of diverse crop varieties, including those used for food, shelter, medicine, and daily necessities. Maize, a crucial crop worldwide, plays a significant role in providing sustenance and income for both rural and urban communities. However, the productivity of maize is often hindered by diseases and pests. Early detection and effective control of these issues are essential for improving crop yield. Unfortunately, many farmers lack access to agricultural experts capable of accurately identifying and treating these problems. This research paper developed a deep-learning model for the identification of common rust diseases and Fall Armyworm pests in maize leaves. To develop and evaluate the proposed method, a comprehensive dataset comprising 3899 images was collected, serving the purposes of training, validation, and testing. The methodology involves several key steps, including image preprocessing, feature extraction, and the training of both a Sequential model and an EfficientNetB0 model. To enhance the quality of maize leaf images and remove noise, various filtering techniques such as Gaussian filtering, median filtering, and a hybrid approach combining the two were compared. The comparative analysis revealed that the hybrid approach, which combines Gaussian and median filtering, yielded superior results. Irfanview64 software was applied for denoising, resizing, and augmenting images, while a convolutional neural network (CNN) was used for feature extraction. In the evaluation of the proposed model, the performance of the Sequential model was compared with the EfficientNetB0 model. The EfficientNetB0 model demonstrated significantly superior performance across all evaluation metrics. Specifically, it achieved perfect accuracy of 100% in both training and validation, as well as 100% accuracy in testing. Furthermore, the model achieved an exceptional overall classification accuracy of 100% based on the f1 score. This deep learning model offers a cost-effective and accurate solution for diagnosing maize leaf diseases and pests, addressing the lack of agricultural experts. Early detection and treatment can significantly improve maize crop yield, contributing to food security.<br>\n&nbsp;</p>", 
  "license": "http://www.opendefinition.org/licenses/cc-by", 
  "creator": [
    {
      "affiliation": "Woldia University", 
      "@type": "Person", 
      "name": "PAULOS MEKONNEN SISAY"
    }
  ], 
  "headline": "DEVELOPMENT OF MAIZE LEAF DISEASES AND PESTS IDENTIFICATION MODEL USING DEEP LEARNING", 
  "image": "https://zenodo.org/static/img/logos/zenodo-gradient-round.svg", 
  "datePublished": "2024-06-21", 
  "url": "https://nadre.ethernet.edu.et/record/10581", 
  "keywords": [
    "deep learning, convolutional neural network, Sequential model, EfficientNetB0, maize leaf disease and pest classification"
  ], 
  "@context": "https://schema.org/", 
  "identifier": "https://doi.org/10.20372/nadre:10581", 
  "@id": "https://doi.org/10.20372/nadre:10581", 
  "@type": "ScholarlyArticle", 
  "name": "DEVELOPMENT OF MAIZE LEAF DISEASES AND PESTS IDENTIFICATION MODEL USING DEEP LEARNING"
}