Tea Leaf Disease Detection and Classification Using Modified Inceptionv3 with Transfer Learning Approach

KILLI BALA KIRAN, DR. P.V.V. SATYANARAYANA

1.           Introduction
1.1        Overview
Many tea plants get affected by diseases such as blister blight, grey blight, and algal leaf spot, all of which will damage the leafs and result in low-quality tea. In the past, farmers would check the leafs for signs of disease by manually inspecting each leaf. This type of inspection was tedious and could be inaccurate or misleading. Therefore, researchers created what is called an Intelligent System to assist in the disease detection process. The intelligent system is based on the Modified Inception V3 architecture that utilizes Transfer Learning on a pre-trained model that is effective at image classification and has learned to identify patterns within an image. The process of adapting and fine-tuning the model using images of both healthy and diseased tea leafs will give the farmer a much better chance at accurately identifying the disease in question. A farmer can easily take a photo of a leaf with his/her mobile phone and the smart system will quickly provide the information regarding the type of disease present on the leaf with a high level of confidence. This early detection allows farmers to treat diseases more quickly and effectively, and to grow higher quality tea with much less effort.