International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 11 Issue: 04 | Apr 2024
p-ISSN: 2395-0072
www.irjet.net
Enhancing Maize Leaf Disease Detection using Transfer Learning Approach Nilakshi Devi1, Shakuntala Laskar2 1Nilakshi Devi: Research Scholar, Assam Don Bosco University, Assam, India
2 Shakuntala Laskar: Professor, Assam Don Bosco University, Assam, India
---------------------------------------------------------------------***--------------------------------------------------------------------and identifying disease symptoms as soon as they appear on Abstract – Maize is a staple food crop for many communities in India, especially in regions where it serves as a primary source of nutrition. It's consumed in various forms like cornmeal, popcorn, and as whole kernels. The agriculture sector is significantly important to India's economy. To stop crop loss and the spread of illness, early detection of plant diseases is essential. Farmers or plant pathologists will typically physically examine the plant leaf to determine the type of illness. Due of the various drawbacks of conventional inspection, scientists are focusing on implementing technology to streamline the procedure. To improve maize leaf disease detection precision, pre-trained models such as VGG16, VGG19, ResNet152 and EfficientNetB7 have been employed. The EfficientNetB7 model outperforms the others, with an accuracy rate of 98.56 %. This enhances crop productivity and guarantees the model's high degree of accuracy for illnesses of the maize leaf. Key Words: Agriculture, Artificial Intelligence, Deep Learning, Transfer Learning, Pre-Trained Models (PTMs), Visual Geometry Group (VGG)
1. INTRODUCTION In India maize is the third-most important cereal crop after wheat and rice[1]. Maize is extensively used as fodder for livestock such as poultry, cattle, and pigs. Its high nutritional content and digestibility make it a valuable component of animal feed. Maize is a versatile crop with numerous industrial applications. It's used in the production of cornstarch, corn syrup, ethanol, and other bio-fuels. Additionally, maize by-products like corn oil and corn gluten meal have various industrial uses [2]. Due to its many applications as food, feed, fodder, and raw materials for various industrial items, it is highly valued all over the world. Maize is very susceptible to a variety of plant diseases, despite its high nutritional value. Maize cultivation offers diversity to Indian agriculture, especially in regions where it's grown alongside other crops. Its resilience to various climatic conditions makes it an important crop choice for farmers, contributing to agricultural sustainability and resilience. Maize leaf diseases have a major effect on the amount and quality of products produced from maize. Diseases affecting maize cereals are influenced by soil type, temperature, and rainfall. It can be challenging to identify and categorize plant diseases in agriculture, even though they can be helpful in monitoring large agricultural fields © 2024, IRJET
|
Impact Factor value: 8.226
|
plant leaves. The classification and detection of crop diseases are the most crucial technical and economical elements in the agricultural sector. For diseases of the maize leaf to be treated and controlled early detection is essential. There are numerous traditional techniques for classifying and identifying diseases in maize leaves, such as using chemicals, human operators, etc. However, there are a number of drawbacks to the traditional methods for identifying maize leaf diseases, including the fact that they are expensive, prone to error, time-consuming, inconsistent, and ineffectual. They also require specialized tools and plant disease knowledge. Feature representation and extraction are the two main limitations of machine learning. In current days, deep learning has greatly advanced the identification of plant diseases [3]. Convolution neural networks (CNNs) are widely employed for illness diagnosis in academia and industry [4]. Deep learning has come a long way in the previous several years. It can now extract useful feature representations from a vast array of input pictures. Deep learning broadens the application of computer vision in precision agriculture by enabling detectors to diagnose agricultural illnesses fast and correctly, in addition to improving plant protection precision. Numerous industries, such as those that make food and drink, poultry, and animal feed, require maize. A major component in the low maize production is the numerous infections that wreak havoc on the crop, drastically reducing its overall yield [5]. Consequently, farmers stand to benefit greatly from a device that can identify plant illnesses based on the appearance of the plant and its telltale symptoms. The development of disease in the host plant is favoured by low temperatures, cloudy conditions, and high humidity. Southern Corn Leaf Blight (SCLB) and Maydis Leaf Blight (MLB), is a fungal disease that affects maize and is caused by the plant pathogen. Bipolaris maydis, or Cochliobolus heterostrophus, as it is sometimes called Bipolaris Maydis (Nisikado) Shoemaker is the cause of Southern maize leaf blight (SCLB), also known as Maydis leaf blight (MLB), a fatal foliar disease that affects maize and has a wide geographic spread. Globally, producing regions developed in warm, humid environments. Early indications of the disease are little regions of necrosis that hover like haloes. The cause of southern rust is the fungus Puccinia polysora. Even though southern rust is usually thought of as a tropical illness, it can appear in areas of the United States and Canada that are important for maize production. There are typical rust-like
ISO 9001:2008 Certified Journal
|
Page 108