Optimized routing algorithm with AlexNet-ShuffleNet for plant leaf disease and infectious classification in IoT

Abstract

Abstract In agriculture, utilizing images to detect plant leaf diseases is a vital area in precision farming. Typically, trained professionals physically inspect plant tissues to identify disease range. Nowadays, AI has made foremost paces in detecting and classifying plant diseases. Moreover, Internet of Things (IoT) has several applications, containing Agricultural-IoT (AIoT), which is considered to elevate agricultural yields. This paper intends to develop an approach in IoT for plant disease classification. Initially, simulation of IoT is done and the IoT nodes route sensed plant leaf images by proposed Serial Exponential Golf Optimization Algorithm (SEGOA), which is established by modifying Golf Optimization Algorithm (GOA) using Exponential Weighted Moving Average (EWMA) to the destination, where plant leaf disease detection is executed. To extract the RoI, CNN is used to discover diseased part in plant leaf. Then, plant leaves is classified as healthy and diseased subclasses by employing AlexNet-ShuffleNet. Moreover, the disease types is classified more into fungal/bacterial/viral infection using the AlexNet-ShuffleNet. Performance of adopted work is assessed by utilizing the metrics, such as energy, accuracy, sensitivity, and specificity. Overall outcome of AlexNet-ShuffleNet give a promising result, such as accuracy of 94.6%, sensitivity of 98.7% and specificity of 94%.

Plant phenotyping relevance

植物葉画像から病変領域を抽出し、健全・罹病状態および病原タイプを分類する画像解析手法を開発・評価しており、植物病害表現型の取得が中心である。

Code and data availability

植物フェノタイピング解析を再現する公開資産であることを、入力本文と直接リンクから確認できなかったため保留しました。

No evidence-backed public reproduction asset is currently recorded.