Title: Effective IoT-based crop disease prediction using localise search traversing coupled with deep convolutional neural network classifier

Authors: B.V. Vani; C.D. Guruprakash

Addresses: Information Science and Engineering, Sri Siddhartha Institute of Technology, Tumkur, Karnataka, India ' Department Computer Science and Engineering, Sri Siddhartha Institute of Technology, Tumkur, Karnataka, India

Abstract: Predicting crop disease on the image obtained from the affected crop has been a potential research topic. In this research, the Localise Search Optimisation Algorithm (LSOA) enabled deep Convolutional Neural Network (deep CNN) is used to predict the crop disease for which the dominant statistical and texture features are utilised and LSOA as a training algorithm. The experiments were done on an apple data set and a corn data set, and the results show that the LSOA-deep CNN model attains 98.474% of accuracy, 92.837% of sensitivity and 99.00% of specificity in k-fold training data and 94.683% of accuracy, 95.489% of specificity and 99.00% of specificity with 80% training data for the corn data set. With the apple data set, the developed method achieves 94.587% of accuracy, 99.00% sensitivity and 99.00% specificity under k-fold training, while for the 80% of training, 97.959% accuracy, 96.233% sensitivity and 99.005% specificity are attained.

Keywords: deep convolutional neural network; optimisation; IoT sensor; wireless sensor network; smart irrigation.

DOI: 10.1504/IJWMC.2024.137173

International Journal of Wireless and Mobile Computing, 2024 Vol.26 No.2, pp.168 - 181

Received: 25 Nov 2022
Accepted: 24 Jul 2023
Published online: 04 Mar 2024 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article