Precision Corn Pest Detection: Two-Step Transfer Learning for Beetles (Coleoptera) with MobileNet-SSD
Using neural networks on low-power mobile systems can aid in controlling pests while preserving beneficial species for crops. However, low-power devices require simplified neural networks, which may lead to reduced performance. This study was focused on developing an optimized deep-learning model fo...
| Main Authors: | , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-12-01
|
| Series: | Agriculture |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2077-0472/13/12/2287 |
| _version_ | 1797382389755805696 |
|---|---|
| author | Edmond Maican Adrian Iosif Sanda Maican |
| author_facet | Edmond Maican Adrian Iosif Sanda Maican |
| author_sort | Edmond Maican |
| collection | DOAJ |
| description | Using neural networks on low-power mobile systems can aid in controlling pests while preserving beneficial species for crops. However, low-power devices require simplified neural networks, which may lead to reduced performance. This study was focused on developing an optimized deep-learning model for mobile devices for detecting corn pests. We propose a two-step transfer learning approach to enhance the accuracy of two versions of the MobileNet SSD network. Five beetle species (Coleoptera), including four harmful to corn crops (belonging to genera <i>Anoxia</i>, <i>Diabrotica</i>, <i>Opatrum</i> and <i>Zabrus</i>), and one beneficial (<i>Coccinella</i> sp.), were selected for preliminary testing. We employed two datasets. One for the first transfer learning procedure comprises 2605 images with general dataset classes ‘Beetle’ and ‘Ladybug’. It was used to recalibrate the networks’ trainable parameters for these two broader classes. Furthermore, the models were retrained on a second dataset of 2648 images of the five selected species. Performance was compared with a baseline model in terms of average accuracy per class and mean average precision (mAP). MobileNet-SSD-v2-Lite achieved an mAP of 0.8923, ranking second but close to the highest mAP (0.908) obtained by MobileNet-SSD-v1 and outperforming the baseline mAP by 6.06%. It demonstrated the highest accuracy for <i>Opatrum</i> (0.9514) and <i>Diabrotica</i> (0.8066). <i>Anoxia</i> it reached a third-place accuracy (0.9851), close to the top value of 0.9912. <i>Zabrus</i> achieved the second position (0.9053), while <i>Coccinella</i> was reliably distinguished from all other species, with an accuracy of 0.8939 and zero false positives; moreover, no pest species were mistakenly identified as <i>Coccinella</i>. Analyzing the errors in the MobileNet-SSD-v2-Lite model revealed good overall accuracy despite the reduced size of the training set, with one misclassification, 33 non-identifications, 7 double identifications and 1 false positive across the 266 images from the test set, yielding an overall relative error rate of 0.1579. The preliminary findings validated the two-step transfer learning procedure and placed the MobileNet-SSD-v2-Lite in the first place, showing high potential for using neural networks on real-time pest control while protecting beneficial species. |
| first_indexed | 2024-03-08T21:04:42Z |
| format | Article |
| id | doaj.art-ab43d86f6562402c91352c28530ca736 |
| institution | Directory Open Access Journal |
| issn | 2077-0472 |
| language | English |
| last_indexed | 2024-03-08T21:04:42Z |
| publishDate | 2023-12-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Agriculture |
| spelling | doaj.art-ab43d86f6562402c91352c28530ca7362023-12-22T13:45:43ZengMDPI AGAgriculture2077-04722023-12-011312228710.3390/agriculture13122287Precision Corn Pest Detection: Two-Step Transfer Learning for Beetles (Coleoptera) with MobileNet-SSDEdmond Maican0Adrian Iosif1Sanda Maican2Faculty of Biotechnical Systems Engineering, National University of Science and Technology Politehnica Bucharest, RO-060042 Bucharest, RomaniaFaculty of Biotechnical Systems Engineering, National University of Science and Technology Politehnica Bucharest, RO-060042 Bucharest, RomaniaInstitute of Biology Bucharest, Romanian Academy, RO-060031 Bucharest, RomaniaUsing neural networks on low-power mobile systems can aid in controlling pests while preserving beneficial species for crops. However, low-power devices require simplified neural networks, which may lead to reduced performance. This study was focused on developing an optimized deep-learning model for mobile devices for detecting corn pests. We propose a two-step transfer learning approach to enhance the accuracy of two versions of the MobileNet SSD network. Five beetle species (Coleoptera), including four harmful to corn crops (belonging to genera <i>Anoxia</i>, <i>Diabrotica</i>, <i>Opatrum</i> and <i>Zabrus</i>), and one beneficial (<i>Coccinella</i> sp.), were selected for preliminary testing. We employed two datasets. One for the first transfer learning procedure comprises 2605 images with general dataset classes ‘Beetle’ and ‘Ladybug’. It was used to recalibrate the networks’ trainable parameters for these two broader classes. Furthermore, the models were retrained on a second dataset of 2648 images of the five selected species. Performance was compared with a baseline model in terms of average accuracy per class and mean average precision (mAP). MobileNet-SSD-v2-Lite achieved an mAP of 0.8923, ranking second but close to the highest mAP (0.908) obtained by MobileNet-SSD-v1 and outperforming the baseline mAP by 6.06%. It demonstrated the highest accuracy for <i>Opatrum</i> (0.9514) and <i>Diabrotica</i> (0.8066). <i>Anoxia</i> it reached a third-place accuracy (0.9851), close to the top value of 0.9912. <i>Zabrus</i> achieved the second position (0.9053), while <i>Coccinella</i> was reliably distinguished from all other species, with an accuracy of 0.8939 and zero false positives; moreover, no pest species were mistakenly identified as <i>Coccinella</i>. Analyzing the errors in the MobileNet-SSD-v2-Lite model revealed good overall accuracy despite the reduced size of the training set, with one misclassification, 33 non-identifications, 7 double identifications and 1 false positive across the 266 images from the test set, yielding an overall relative error rate of 0.1579. The preliminary findings validated the two-step transfer learning procedure and placed the MobileNet-SSD-v2-Lite in the first place, showing high potential for using neural networks on real-time pest control while protecting beneficial species.https://www.mdpi.com/2077-0472/13/12/2287pest controlColeopterasmart agricultureneural networkMobileNettransfer learning |
| spellingShingle | Edmond Maican Adrian Iosif Sanda Maican Precision Corn Pest Detection: Two-Step Transfer Learning for Beetles (Coleoptera) with MobileNet-SSD Agriculture pest control Coleoptera smart agriculture neural network MobileNet transfer learning |
| title | Precision Corn Pest Detection: Two-Step Transfer Learning for Beetles (Coleoptera) with MobileNet-SSD |
| title_full | Precision Corn Pest Detection: Two-Step Transfer Learning for Beetles (Coleoptera) with MobileNet-SSD |
| title_fullStr | Precision Corn Pest Detection: Two-Step Transfer Learning for Beetles (Coleoptera) with MobileNet-SSD |
| title_full_unstemmed | Precision Corn Pest Detection: Two-Step Transfer Learning for Beetles (Coleoptera) with MobileNet-SSD |
| title_short | Precision Corn Pest Detection: Two-Step Transfer Learning for Beetles (Coleoptera) with MobileNet-SSD |
| title_sort | precision corn pest detection two step transfer learning for beetles coleoptera with mobilenet ssd |
| topic | pest control Coleoptera smart agriculture neural network MobileNet transfer learning |
| url | https://www.mdpi.com/2077-0472/13/12/2287 |
| work_keys_str_mv | AT edmondmaican precisioncornpestdetectiontwosteptransferlearningforbeetlescoleopterawithmobilenetssd AT adrianiosif precisioncornpestdetectiontwosteptransferlearningforbeetlescoleopterawithmobilenetssd AT sandamaican precisioncornpestdetectiontwosteptransferlearningforbeetlescoleopterawithmobilenetssd |