Detection and Counting of Small Target Apples under Complicated Environments by Using Improved YOLOv7-tiny
Weather disturbances, difficult backgrounds, the shading of fruit and foliage, and other elements can significantly affect automated yield estimation and picking in small target apple orchards in natural settings. This study uses the MinneApple public dataset, which is processed to construct a datas...
| Main Authors: | , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2023-05-01
|
| Series: | Agronomy |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2073-4395/13/5/1419 |
| _version_ | 1797601457789206528 |
|---|---|
| author | Li Ma Liya Zhao Zixuan Wang Jian Zhang Guifen Chen |
| author_facet | Li Ma Liya Zhao Zixuan Wang Jian Zhang Guifen Chen |
| author_sort | Li Ma |
| collection | DOAJ |
| description | Weather disturbances, difficult backgrounds, the shading of fruit and foliage, and other elements can significantly affect automated yield estimation and picking in small target apple orchards in natural settings. This study uses the MinneApple public dataset, which is processed to construct a dataset of 829 images with complex weather, including 232 images of fog scenarios and 236 images of rain scenarios, and proposes a lightweight detection algorithm based on the upgraded YOLOv7-tiny. In this study, a backbone network was constructed by adding skip connections to shallow features, using P2BiFPN for multi-scale feature fusion and feature reuse at the neck, and incorporating a lightweight ULSAM attention mechanism to reduce the loss of small target features, focusing on the correct target and discard redundant features, thereby improving detection accuracy. The experimental results demonstrate that the model has an mAP of 80.4% and a loss rate of 0.0316. The mAP is 5.5% higher than the original model, and the model size is reduced by 15.81%, reducing the requirement for equipment; In terms of counts, the MAE and RMSE are 2.737 and 4.220, respectively, which are 5.69% and 8.97% lower than the original model. Because of its improved performance and stronger robustness, this experimental model offers fresh perspectives on hardware deployment and orchard yield estimation. |
| first_indexed | 2024-03-11T04:01:05Z |
| format | Article |
| id | doaj.art-e3dc65307f114c00b64f385081eb7e94 |
| institution | Directory Open Access Journal |
| issn | 2073-4395 |
| language | English |
| last_indexed | 2024-03-11T04:01:05Z |
| publishDate | 2023-05-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Agronomy |
| spelling | doaj.art-e3dc65307f114c00b64f385081eb7e942023-11-18T00:08:06ZengMDPI AGAgronomy2073-43952023-05-01135141910.3390/agronomy13051419Detection and Counting of Small Target Apples under Complicated Environments by Using Improved YOLOv7-tinyLi Ma0Liya Zhao1Zixuan Wang2Jian Zhang3Guifen Chen4College of Information Technology, Jilin Agricultural University, Changchun 130118, ChinaCollege of Information Technology, Jilin Agricultural University, Changchun 130118, ChinaCollege of Information Technology, Jilin Agricultural University, Changchun 130118, ChinaFaculty of Agronomy, Jilin Agricultural University, Changchun 130118, ChinaInstitute of Technology, Changchun Humanities and Sciences College, Changchun 130118, ChinaWeather disturbances, difficult backgrounds, the shading of fruit and foliage, and other elements can significantly affect automated yield estimation and picking in small target apple orchards in natural settings. This study uses the MinneApple public dataset, which is processed to construct a dataset of 829 images with complex weather, including 232 images of fog scenarios and 236 images of rain scenarios, and proposes a lightweight detection algorithm based on the upgraded YOLOv7-tiny. In this study, a backbone network was constructed by adding skip connections to shallow features, using P2BiFPN for multi-scale feature fusion and feature reuse at the neck, and incorporating a lightweight ULSAM attention mechanism to reduce the loss of small target features, focusing on the correct target and discard redundant features, thereby improving detection accuracy. The experimental results demonstrate that the model has an mAP of 80.4% and a loss rate of 0.0316. The mAP is 5.5% higher than the original model, and the model size is reduced by 15.81%, reducing the requirement for equipment; In terms of counts, the MAE and RMSE are 2.737 and 4.220, respectively, which are 5.69% and 8.97% lower than the original model. Because of its improved performance and stronger robustness, this experimental model offers fresh perspectives on hardware deployment and orchard yield estimation.https://www.mdpi.com/2073-4395/13/5/1419YOLOv7-tiny-Applesmall targetfruit detection and countingdigital agriculture |
| spellingShingle | Li Ma Liya Zhao Zixuan Wang Jian Zhang Guifen Chen Detection and Counting of Small Target Apples under Complicated Environments by Using Improved YOLOv7-tiny Agronomy YOLOv7-tiny-Apple small target fruit detection and counting digital agriculture |
| title | Detection and Counting of Small Target Apples under Complicated Environments by Using Improved YOLOv7-tiny |
| title_full | Detection and Counting of Small Target Apples under Complicated Environments by Using Improved YOLOv7-tiny |
| title_fullStr | Detection and Counting of Small Target Apples under Complicated Environments by Using Improved YOLOv7-tiny |
| title_full_unstemmed | Detection and Counting of Small Target Apples under Complicated Environments by Using Improved YOLOv7-tiny |
| title_short | Detection and Counting of Small Target Apples under Complicated Environments by Using Improved YOLOv7-tiny |
| title_sort | detection and counting of small target apples under complicated environments by using improved yolov7 tiny |
| topic | YOLOv7-tiny-Apple small target fruit detection and counting digital agriculture |
| url | https://www.mdpi.com/2073-4395/13/5/1419 |
| work_keys_str_mv | AT lima detectionandcountingofsmalltargetapplesundercomplicatedenvironmentsbyusingimprovedyolov7tiny AT liyazhao detectionandcountingofsmalltargetapplesundercomplicatedenvironmentsbyusingimprovedyolov7tiny AT zixuanwang detectionandcountingofsmalltargetapplesundercomplicatedenvironmentsbyusingimprovedyolov7tiny AT jianzhang detectionandcountingofsmalltargetapplesundercomplicatedenvironmentsbyusingimprovedyolov7tiny AT guifenchen detectionandcountingofsmalltargetapplesundercomplicatedenvironmentsbyusingimprovedyolov7tiny |