Artificial Intelligence for Classification and Detection of Major Insect Pests of Brinjal Kumar N Sai1,*, Emmanuel N1, Krishna K Sri Phani2, Chinnabbai Ch1, Krishna K Uma1 1Dr YSR Horticultural University, Venkataramannagudem, West Godavari District 543101, Andhra Pradesh, India 2National Institute of Technology, Tadepalligudem543101Andhra Pradesh, India *Email: saikumar363539@gmail.com (corresponding author): ORCID ID 0000-0002-8629-7583
Online Published on 18 October, 2023. Abstract The present study was carried out during rabi 2020-21 at the College of Horticulture, Venkataramannagudem, West Godavari, Andhra Pradesh. Detection and monitoring of insect pests through Artificial intelligence (AI) were conducted using the Python software through Keras and Tensorflow frame works and for this purpose CNN VGG-16 model was used. A total of 204 insect images were loaded and included in four datasets. All the datasets were resized to 224_224 pixels. CNN VGG-16 model codes developed for automatic pest classification and detection of pest images were run through python language to retrieve the predicted output. In the pursuit of detecting insect classification an accuracy of 95-98% for 4 insect classes viz., brinjal shoot and fruit borer larvae and adults, Epilachna beetle grub and adults were predicted with F1 score of 0.89 which shows that the CNN (VGG) model is consistent in detecting the type of insect. Top Keywords Machine learning (ML), Convolutional neural networks, Python language, Keras, Tensor flow, Insect pest, Insect images, Prediction models, Insect boxing, Insect monitoring. Top |