Research on Detection and Counting Method of Green Walnut Based on YOLOv8n-RBP

In natural environments, green walnuts often experience occlusion by branches and leaves, fruit overlap, and varying lighting conditions. To address the issues of low detection accuracy, missed detections, and false positives in the YOLO model, this study proposes an improved YOLOv8n-based detection...

Full description

Saved in:
Bibliographic Details
Main Authors: Bangbang Chen, Keke Tan, Kun Li, Baojian Ma, Xiangdong Liu
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10906562/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850028723123781632
author Bangbang Chen
Keke Tan
Kun Li
Baojian Ma
Xiangdong Liu
author_facet Bangbang Chen
Keke Tan
Kun Li
Baojian Ma
Xiangdong Liu
author_sort Bangbang Chen
collection DOAJ
description In natural environments, green walnuts often experience occlusion by branches and leaves, fruit overlap, and varying lighting conditions. To address the issues of low detection accuracy, missed detections, and false positives in the YOLO model, this study proposes an improved YOLOv8n-based detection and counting model for green walnuts, named YOLOv8n-RBP. First, a receptive field-concentrated attention module (RFCBAM) is integrated into the backbone network to enhance feature extraction capabilities. Second, a BiFPN-GLSA module is introduced to replace the Path Aggregation Network (PANet) in the neck, improving the fusion of feature layers from the backbone and Neck networks and enhancing the model’s ability to capture both global and local spatial features. Lastly, to address the weak generalization and slow convergence issues of the CIoU loss function in detection tasks, the PIoUv2 loss function is employed to accelerate bounding box regression and improve detection performance. Experimental results demonstrate that the YOLOv8n-RBP model excels across multiple evaluation metrics. Specifically, the model achieves a mean average precision (mAP@0.5) of 82.2% and a recall rate of 72.4%, with a model size of only 4.65 MB, 2.2 million parameters, and 8.3 GFLOPs. Compared to the original YOLOv8n model, the recall rate and mAP@0.5 improve by 2.7% and 2.5%, respectively, while the number of parameters, FLOPs, and model size decrease by 26.7%, 0.6%, and 22.0%, respectively. Further deployment on the NVIDIA Jetson Xavier NX demonstrates the model’s robust performance under natural conditions, indicating its suitability for orchard operations.
format Article
id doaj-art-5f74df047e944a05800c9d7c2c824be7
institution DOAJ
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-5f74df047e944a05800c9d7c2c824be72025-08-20T02:59:45ZengIEEEIEEE Access2169-35362025-01-0113392753928810.1109/ACCESS.2025.354631310906562Research on Detection and Counting Method of Green Walnut Based on YOLOv8n-RBPBangbang Chen0https://orcid.org/0009-0009-2110-6393Keke Tan1Kun Li2Baojian Ma3Xiangdong Liu4School of Mechatronic Engineering, Xinjiang Institute of Technology, Aksu, ChinaSchool of Mechatronic Engineering, Xinjiang Institute of Technology, Aksu, ChinaSchool of Mechatronic Engineering, Xinjiang Institute of Technology, Aksu, ChinaSchool of Mechatronic Engineering, Xinjiang Institute of Technology, Aksu, ChinaSchool of Mechatronic Engineering, Xinjiang Institute of Technology, Aksu, ChinaIn natural environments, green walnuts often experience occlusion by branches and leaves, fruit overlap, and varying lighting conditions. To address the issues of low detection accuracy, missed detections, and false positives in the YOLO model, this study proposes an improved YOLOv8n-based detection and counting model for green walnuts, named YOLOv8n-RBP. First, a receptive field-concentrated attention module (RFCBAM) is integrated into the backbone network to enhance feature extraction capabilities. Second, a BiFPN-GLSA module is introduced to replace the Path Aggregation Network (PANet) in the neck, improving the fusion of feature layers from the backbone and Neck networks and enhancing the model’s ability to capture both global and local spatial features. Lastly, to address the weak generalization and slow convergence issues of the CIoU loss function in detection tasks, the PIoUv2 loss function is employed to accelerate bounding box regression and improve detection performance. Experimental results demonstrate that the YOLOv8n-RBP model excels across multiple evaluation metrics. Specifically, the model achieves a mean average precision (mAP@0.5) of 82.2% and a recall rate of 72.4%, with a model size of only 4.65 MB, 2.2 million parameters, and 8.3 GFLOPs. Compared to the original YOLOv8n model, the recall rate and mAP@0.5 improve by 2.7% and 2.5%, respectively, while the number of parameters, FLOPs, and model size decrease by 26.7%, 0.6%, and 22.0%, respectively. Further deployment on the NVIDIA Jetson Xavier NX demonstrates the model’s robust performance under natural conditions, indicating its suitability for orchard operations.https://ieeexplore.ieee.org/document/10906562/Natural environmentgreen walnutdetectionalgorithm improvementequipment deployment
spellingShingle Bangbang Chen
Keke Tan
Kun Li
Baojian Ma
Xiangdong Liu
Research on Detection and Counting Method of Green Walnut Based on YOLOv8n-RBP
IEEE Access
Natural environment
green walnut
detection
algorithm improvement
equipment deployment
title Research on Detection and Counting Method of Green Walnut Based on YOLOv8n-RBP
title_full Research on Detection and Counting Method of Green Walnut Based on YOLOv8n-RBP
title_fullStr Research on Detection and Counting Method of Green Walnut Based on YOLOv8n-RBP
title_full_unstemmed Research on Detection and Counting Method of Green Walnut Based on YOLOv8n-RBP
title_short Research on Detection and Counting Method of Green Walnut Based on YOLOv8n-RBP
title_sort research on detection and counting method of green walnut based on yolov8n rbp
topic Natural environment
green walnut
detection
algorithm improvement
equipment deployment
url https://ieeexplore.ieee.org/document/10906562/
work_keys_str_mv AT bangbangchen researchondetectionandcountingmethodofgreenwalnutbasedonyolov8nrbp
AT keketan researchondetectionandcountingmethodofgreenwalnutbasedonyolov8nrbp
AT kunli researchondetectionandcountingmethodofgreenwalnutbasedonyolov8nrbp
AT baojianma researchondetectionandcountingmethodofgreenwalnutbasedonyolov8nrbp
AT xiangdongliu researchondetectionandcountingmethodofgreenwalnutbasedonyolov8nrbp