Scholarlink

Plant diseases are some of the most critical threats to global agriculture, inducing economic losses and food insecurity. Apples and maize rank among the most highly cultivated crops in the world, giving essential nutrition and financial stability to millions of farmers. These diseases on these crops reduce agricultural productivity and quality, hence affecting farmers’ incomes. Accurate and early disease detection is primarily important for effective crop management, thereby reducing losses. Conventionally, manual observation has been adopted, which besides being very time-consuming, is labor-intensive, expensive, and error prone. This paper presents an improved usage of artificial intelligence and deep learning for better performance and efficiency. Two varieties of plants have been taken, apples and maize. Using CNNs for both apple and maize, with VGG16 and MobileNetV2 for apple and ResNet50 and ResNet18 for maize. These models analyze the images of leaves to identify and classify diseases with high accuracy. Advanced pre-processing techniques, such as data augmentation, edge detection, and contrast enhancement, are included in the research to improve performance The inference time, calculated as the time taken by the model to process an input image and produce a classification result, is analyzed to ensure real-time disease detection and practical field application The proposed CNN model demonstrates efficiency for real-time disease detection. It processes apple images in 7.16 ms and maize images in 10.6 ms per input. In comparison, MobileNetV2 takes 14.73 ms, VGG16 takes 343.70 ms for apple images, ResNet18 takes 21.80 ms, and ResNet50 takes 55.52 ms for maize images. Experimental results have proven that deep learning models, especially MobileNetV2 and CNN, achieve superior classification accuracy and thus can be used in real-world applications. This integration of AI-powered disease detection provides real-time insights, thus enabling early intervention, with reduced use of chemical pesticides. This approach, leveraging AI for proper disease recognition, is meant to bring about environmental friendliness in agriculture, increased yield, and improved food security in Egypt and further afield.

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

Leveraging Deep Learning for Apple and Maize Plant Disease Detection for Sustainable Smart Agriculture Solution

  • 章节

  • 发表日期：

    2026年

查看全文
  • Bassant Ali,
  • Amany M. Sarhan,
  • Aboul Ella Hassanien

摘要

Plant diseases are some of the most critical threats to global agriculture, inducing economic losses and food insecurity. Apples and maize rank among the most highly cultivated crops in the world, giving essential nutrition and financial stability to millions of farmers. These diseases on these crops reduce agricultural productivity and quality, hence affecting farmers’ incomes. Accurate and early disease detection is primarily important for effective crop management, thereby reducing losses. Conventionally, manual observation has been adopted, which besides being very time-consuming, is labor-intensive, expensive, and error prone. This paper presents an improved usage of artificial intelligence and deep learning for better performance and efficiency. Two varieties of plants have been taken, apples and maize. Using CNNs for both apple and maize, with VGG16 and MobileNetV2 for apple and ResNet50 and ResNet18 for maize. These models analyze the images of leaves to identify and classify diseases with high accuracy. Advanced pre-processing techniques, such as data augmentation, edge detection, and contrast enhancement, are included in the research to improve performance The inference time, calculated as the time taken by the model to process an input image and produce a classification result, is analyzed to ensure real-time disease detection and practical field application The proposed CNN model demonstrates efficiency for real-time disease detection. It processes apple images in 7.16 ms and maize images in 10.6 ms per input. In comparison, MobileNetV2 takes 14.73 ms, VGG16 takes 343.70 ms for apple images, ResNet18 takes 21.80 ms, and ResNet50 takes 55.52 ms for maize images. Experimental results have proven that deep learning models, especially MobileNetV2 and CNN, achieve superior classification accuracy and thus can be used in real-world applications. This integration of AI-powered disease detection provides real-time insights, thus enabling early intervention, with reduced use of chemical pesticides. This approach, leveraging AI for proper disease recognition, is meant to bring about environmental friendliness in agriculture, increased yield, and improved food security in Egypt and further afield.