An Intuitive Convolutional Neural Network to Perceive Tomato Leaf Ailment
Poonam Shourie, Vatsala Anand, Himakshi Gupta
Abstract
The biological properties of tomato leaves must be categorized to determine if the plant needs to be treated with pesticides or insecticides to produce a decent harvest. To maximize agricultural output, plant diseases must be quickly and accurately identified. A deep learning method that makes use of artificial intelligence may identify illnesses in plants using a huge number of photos of plant leaves. However, it is challenging to properly diagnose illnesses using deep learning with little datasets. A deep learning system that automatically extracts characteristics in a hierarchical fashion named convolutional neural network is suggested in this research. To categorize the characteristics of the leaves into 10 classes such as healthy, bacterial spot, and Septoria leaf spot—neural networks are used. The model detects illness using both saved and real-time photos of tomato plants. In order to determine the proposed model's high degree of its performance is further assessed using adaptive moment estimation (Adam), batch sizes of 32, and varied epochs. The metric for measuring performance that was acquired confirms the effectiveness of the technique. The technique is helpful for managing tomato plant disease and for remedial action.