International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 11 Issue: 03 | Mar 2024
p-ISSN: 2395-0072
www.irjet.net
LEAF DISEASE DETECTION USING MOBILENET PROF. HEENA PATIL1, PROF. RENUKA DINGORE2, HARDIK CHAVAN3, MOURYA B.N4, HARSH MISHRA5. 1Head of Department, Dept. of AIML Diploma, ARMIET ,Maharashtra, India 2Lecturer, Dept. of AIML Diploma, ARMIET ,Maharashtra, India 3Student, Dept. of AIML Diploma, ARMIET ,Maharashtra, India 4Student, Dept. of AIML Diploma, ARMIET ,Maharashtra, India 5Student, Dept. of AIML Diploma, ARMIET ,Maharashtra, India
---------------------------------------------------------------------***---------------------------------------------------------------------
Abstract: The agriculture industry is critical to ensuring
farming practices. Moreover, DL can be applied in remote sensing applications, where large agricultural areas can be monitored efficiently.[3]
global food security, hence early diagnosis of plant diseases is critical. Deep Learning (DL) algorithms have emerged as powerful tools for automating the detection of leaf diseases, resulting in better crop management and higher agricultural yields. Deep Learning models, particularly Convolutional Neural Networks (CNNs), have proven to be extremely effective at automating disease detection in plants. DL algorithms can learn complicated patterns and features that distinguish between diseased and healthy leaf photos, allowing them to accurately identify numerous plant illnesses. The study used a dataset full of field images that we pre-trained on deep convolutional neural network called MobilNet and a method known as transfer learning by demonstrating the potential of deep learning approaches in detecting leaf disease. To achieve this goal, the transfer learning model was fine-tuned using a variety of hyperparameters and achieved an 95% accuracy rate
1.1 Problem Statement In the most of studies, the Plant Village dataset was utilized to assess the performance of the DL models. Although this dataset contains many photos of many plant species and their illnesses, they were captured in a laboratory. As a result, it is projected to generate a substantial dataset of plant diseases in real-world settings. Although some research uses hyperspectral pictures of ill leaves, and various DL frameworks are used for early identification of plant leaf diseases, issues that impede the broad application of HSI in plant disease detection remain unresolved. That is, labeled datasets for early plant disease detection are difficult to get, and even experienced specialists are unable to pinpoint where the invisible disease symptoms are and designate totally invisible disease pixels, which is critical for HSI to detect plant disease.[2]
Key Words: Disease detection, Deep learning, MobileNet, Transformer etc
1.INTRODUCTION
1.2 Purpose
In recent years, the agricultural sector has witnessed a transformative shift in the way plant diseases are diagnosed and managed. Traditional methods of disease identification, which often involve visual inspection by human experts, are labour-intensive and may lead to delayed interventions, potentially causing significant crop losses. To address these challenges, the integration of Deep Learning (DL) techniques has emerged as a groundbreaking approach for the automated and accurate detection of leaf diseases in plants.
DL-based systems are designed to identify plant diseases at an early stage. This enables farmers and agricultural experts to take prompt action to mitigate the spread of diseases and minimize crop losses The primary purpose is to protect crops from diseases. Timely detection allows for targeted interventions, such as applying the right amount of pesticides or adopting disease-resistant crop varieties, to safeguard agricultural yields. By identifying and managing diseases effectively, DL helps increase crop productivity. Healthy plants are more likely to produce higher yields, contributing to food security and economic sustainability.
Deep Learning is a subset of machine learning that focuses on training artificial neural networks to automatically learn and extract intricate patterns from data. Specifically, Convolutional Neural Networks (CNNs) have proven to be highly effective for image-related tasks, making them particularly suited for leaf disease detection. The use of DL for leaf disease detection offers several advantages, including speed, accuracy, and scalability. It enables early disease identification, allowing for timely interventions, reducing crop losses, and promoting more sustainable
© 2023, IRJET
|
Impact Factor value: 8.226
Accurate disease detection and management reduce the need for excessive pesticide use, which can have negative environmental consequences. DL helps in using pesticides more judiciously, minimizing ecological harm. DL provides valuable data and insights to farmers and agricultural experts, empowering them to make informed decisions about disease management strategies, irrigation, and other
|
ISO 9001:2008 Certified Journal
|
Page 202