AI Deep Transfer Learning DenseNet Models for Plant Disease Classification

Abstract

These are the illnesses that significantly impair the agricultural productivity and quality of crop, resulting in substantial economic losses globally. Early disease identification is an important issue in managing crops and agricultural sustainability. This paper proposes a deep learning-based identification of plant diseases system that can automatically detects plant illnesses from plant photos using DenseNet architecture. The algorithm was trained and tested on a huge group of pictures of well then ill plants of different crops along with disease lines. Data preparation techniques such picture scaling, normalization, and smoothing were used to upsurge the generalization and resilience of the prototypical. The suggested DenseNet model encourages high collaboration between layers to improve vegetation coverage, which enables vegetation reuse and reduces the "vanishing gradient" problem. The experimental findings indicate that the suggested technique is superior to classical deep learning.