Real-Time Detection and Classification of <i>Scirtothrips dorsalis</i> on Fruit Crops with Smartphone-Based Deep Learning System: Preliminary Results
This study proposes a deep-learning-based system for detecting and classifying <i>Scirtothrips dorsalis</i> Hood, a highly invasive insect pest that causes significant economic losses to fruit crops worldwide. The system uses yellow sticky traps and a deep learning model to detect the pr...
Main Authors: | , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2023-06-01
|
Series: | Insects |
Subjects: | |
Online Access: | https://www.mdpi.com/2075-4450/14/6/523 |
_version_ | 1797594247025655808 |
---|---|
author | Gildas Niyigena Sangjun Lee Soonhwa Kwon Daebin Song Byoung-Kwan Cho |
author_facet | Gildas Niyigena Sangjun Lee Soonhwa Kwon Daebin Song Byoung-Kwan Cho |
author_sort | Gildas Niyigena |
collection | DOAJ |
description | This study proposes a deep-learning-based system for detecting and classifying <i>Scirtothrips dorsalis</i> Hood, a highly invasive insect pest that causes significant economic losses to fruit crops worldwide. The system uses yellow sticky traps and a deep learning model to detect the presence of thrips in real time, allowing farmers to take prompt action to prevent the spread of the pest. To achieve this, several deep learning models are evaluated, including YOLOv5, Faster R-CNN, SSD MobileNetV2, and EfficientDet-D0. EfficientDet-D0 was integrated into the proposed smartphone application for mobility and usage in the absence of Internet coverage because of its smaller model size, fast inference time, and reasonable performance on the relevant dataset. This model was tested on two datasets, in which thrips and non-thrips insects were captured under different lighting conditions. The system installation took up 13.5 MB of the device’s internal memory and achieved an inference time of 76 ms with an accuracy of 93.3%. Additionally, this study investigated the impact of lighting conditions on the performance of the model, which led to the development of a transmittance lighting setup to improve the accuracy of the detection system. The proposed system is a cost-effective and efficient alternative to traditional detection methods and provides significant benefits to fruit farmers and the related ecosystem. |
first_indexed | 2024-03-11T02:19:55Z |
format | Article |
id | doaj.art-f4dc80bb609048229c697d02441aa025 |
institution | Directory Open Access Journal |
issn | 2075-4450 |
language | English |
last_indexed | 2024-03-11T02:19:55Z |
publishDate | 2023-06-01 |
publisher | MDPI AG |
record_format | Article |
series | Insects |
spelling | doaj.art-f4dc80bb609048229c697d02441aa0252023-11-18T10:56:00ZengMDPI AGInsects2075-44502023-06-0114652310.3390/insects14060523Real-Time Detection and Classification of <i>Scirtothrips dorsalis</i> on Fruit Crops with Smartphone-Based Deep Learning System: Preliminary ResultsGildas Niyigena0Sangjun Lee1Soonhwa Kwon2Daebin Song3Byoung-Kwan Cho4Department of Smart Agricultural System, Chungnam National University, Daejeon 34134, Republic of KoreaDepartment of Biosystems Machinery Engineering, Chungnam National University, Daejeon 34134, Republic of KoreaCitrus Research Institute, Seogwipo 63607, Republic of KoreaDepartment of Biosystem Bio-Industrial Machinery Engineering, Gyeongsang National University, Jinju 52828, Republic of KoreaDepartment of Smart Agricultural System, Chungnam National University, Daejeon 34134, Republic of KoreaThis study proposes a deep-learning-based system for detecting and classifying <i>Scirtothrips dorsalis</i> Hood, a highly invasive insect pest that causes significant economic losses to fruit crops worldwide. The system uses yellow sticky traps and a deep learning model to detect the presence of thrips in real time, allowing farmers to take prompt action to prevent the spread of the pest. To achieve this, several deep learning models are evaluated, including YOLOv5, Faster R-CNN, SSD MobileNetV2, and EfficientDet-D0. EfficientDet-D0 was integrated into the proposed smartphone application for mobility and usage in the absence of Internet coverage because of its smaller model size, fast inference time, and reasonable performance on the relevant dataset. This model was tested on two datasets, in which thrips and non-thrips insects were captured under different lighting conditions. The system installation took up 13.5 MB of the device’s internal memory and achieved an inference time of 76 ms with an accuracy of 93.3%. Additionally, this study investigated the impact of lighting conditions on the performance of the model, which led to the development of a transmittance lighting setup to improve the accuracy of the detection system. The proposed system is a cost-effective and efficient alternative to traditional detection methods and provides significant benefits to fruit farmers and the related ecosystem.https://www.mdpi.com/2075-4450/14/6/523<i>Scirtothrips dorsalis</i>real timesmartphone applicationlighting conditionsobject detection |
spellingShingle | Gildas Niyigena Sangjun Lee Soonhwa Kwon Daebin Song Byoung-Kwan Cho Real-Time Detection and Classification of <i>Scirtothrips dorsalis</i> on Fruit Crops with Smartphone-Based Deep Learning System: Preliminary Results Insects <i>Scirtothrips dorsalis</i> real time smartphone application lighting conditions object detection |
title | Real-Time Detection and Classification of <i>Scirtothrips dorsalis</i> on Fruit Crops with Smartphone-Based Deep Learning System: Preliminary Results |
title_full | Real-Time Detection and Classification of <i>Scirtothrips dorsalis</i> on Fruit Crops with Smartphone-Based Deep Learning System: Preliminary Results |
title_fullStr | Real-Time Detection and Classification of <i>Scirtothrips dorsalis</i> on Fruit Crops with Smartphone-Based Deep Learning System: Preliminary Results |
title_full_unstemmed | Real-Time Detection and Classification of <i>Scirtothrips dorsalis</i> on Fruit Crops with Smartphone-Based Deep Learning System: Preliminary Results |
title_short | Real-Time Detection and Classification of <i>Scirtothrips dorsalis</i> on Fruit Crops with Smartphone-Based Deep Learning System: Preliminary Results |
title_sort | real time detection and classification of i scirtothrips dorsalis i on fruit crops with smartphone based deep learning system preliminary results |
topic | <i>Scirtothrips dorsalis</i> real time smartphone application lighting conditions object detection |
url | https://www.mdpi.com/2075-4450/14/6/523 |
work_keys_str_mv | AT gildasniyigena realtimedetectionandclassificationofiscirtothripsdorsalisionfruitcropswithsmartphonebaseddeeplearningsystempreliminaryresults AT sangjunlee realtimedetectionandclassificationofiscirtothripsdorsalisionfruitcropswithsmartphonebaseddeeplearningsystempreliminaryresults AT soonhwakwon realtimedetectionandclassificationofiscirtothripsdorsalisionfruitcropswithsmartphonebaseddeeplearningsystempreliminaryresults AT daebinsong realtimedetectionandclassificationofiscirtothripsdorsalisionfruitcropswithsmartphonebaseddeeplearningsystempreliminaryresults AT byoungkwancho realtimedetectionandclassificationofiscirtothripsdorsalisionfruitcropswithsmartphonebaseddeeplearningsystempreliminaryresults |