APPLE FOLIAR DISEASE DETECTION USING CONVOLUTIONAL NEURAL NETWORK BASED APPROACH

Abstract

Agriculture is the process of growing crops and raising livestock and cultivating other forms of food or fiber. It has been a fundamental activity for human societies through out history, providing food and other resources necessary for survival. It also provides employment, economic growth and environmental conservation. Some farming practices, such as sustainable agriculture, can promote environmental conservation and biodiver sity. Hence, plant illness can have a substantial effect on the economy, particularly in agricultural-dependent countries. Crop yield loss, trade restrictions, increased production costs, reduced agricultural productivity, and research and development costs are some of the ways plant diseases can affect the economy. It is essential to prevent and manage plant diseases to ensure food security and maintain a healthy agricultural sector. Farmers visually inspect their crops for symptoms of diseases, such as discoloration, spots, wilting, and deformities. Farmers can also use their sense of touch and smell to detect diseases, such as the sticky or slimy feel of plant leaves infected with fungal diseases and the foul smell of rotting or decaying plant material. However, traditional methods of plant disease detection do have limitations. Visual inspection and other traditional methods may not always detect diseases at an early stage, and there is a risk of misdiagnosis. Additionally, traditional methods may not be able to detect diseases that are not visible to the naked eye. The other alternatives are the use of Artificial Intelligence (AI) which includes training computers to detect plant diseases using image recognition technology. AI can analyze thousands of images to detect subtle changes in plant health that may indicate the pres ence of diseases. Some of the regular AI techniques used for plant disease detection are Convolutional Neural Networks (CNNs), Support Vector Machines (SVMs), Random Forest (RF), Deep Belief Networks (DBNs), Transfer Learning. These AI methods are increasingly being used for plant disease detection because they offer a fast, accurate, and cost-effective way to diagnose plant diseases, which can help prevent crop losses and iv increase yields. Here in this research work, a Multi-layered CNN model is introduced which is inspired from InceptionNet. The proposed model is trained on the “Plant Pathology 2020: FGVC7 dataset” and “Plant Pathology 2021: FGVC8 dataset”. This proposed model is compared with pertained models: InceptionV3. According to the findings, the suggested model outperforms other pre-existing models in terms of accuracy or performance and it’s able to detect the disease with a low error rate.