黄 宝,阳学进,李 智,谢 聪

针对自然环境下柑橘检测易受遮挡、光照变化影响导致准确率低的问题，本文提出一种新的树上柑橘检测算法。在主干嵌入C2BRA稀疏模块，强化密集场景下的细粒度特征表达，提升密集果实的区分度；利用SOEP模块，更大程度地保留幼果信息；引入SEAMHead提升遮挡果实可见区域的特征响应，应对遮挡问题。实验结果表明，该模型在自建的柑橘数据集上mAP50可达93.4%，较YOLOv11n提升4.3个百分点，优于原始模型及对比算法，有助于推动柑橘产业的智能化发展。

YOLOv11；目标检测；特征提取；柑橘识别

[1] 刘妤，刘洒，杨长辉，等．无遮挡重叠柑橘目标分割与重建[J]．江苏农业学报，2019，35（6）：1441-1449. [2] Zheng Z, Xiong J, Lin H, et al. A method of green citrus detection in natural environments using a deep convolutional neural network[J]. Frontiers in Plant Science, 2021, 12: 705737. [3] Ren S, He K, Girshick R, et al. Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(6): 1137-1149. [4] Redmon J, Divvala S, Girshick R, et al. You only look once: unified, real-time object detection[C]//2016 IEEE Conference on Computer Vision and Pattern Recognition (cvpr). Las Vegas, NV, USA: IEEE, 2016: 779-788. [5] Liu W, Anguelov D, Erhan D, et al. SSD: single shot MultiBox detector[M]. 2016: 21-37. [6] 刘芳．基于YOLOV5的柑橘果实目标检测研究[J]．信息与电脑（理论版），2022，34（2）：152-154. [7] 李川红，付兴兰，傅晗文，等．基于无人机影像的丘陵山地柑橘果实实时检测与计数[J]．农业工程学报，2025，41（17）：153-161. [8] Khanam R, Hussain M. YOLOv11: an overview of the key architectural enhancements[A]. arXiv, 2024. [9] Wu C, Fang J, Wang X, et al. DMSF-YOLO: cow behavior recognition algorithm based on dynamic mechanism and multi-scale feature fusion[J]. Sensors, 2025, 25(11): 3479. [10] Zhang Y, Lin C, Chen G. Efficient multi-scale detection of construction workers and vehicles based on deep learning[J]. Journal of Real-Time Image Processing, 2025, 22(3): 127. [11] Li Z, Hu X, Zhao X, et al. Beyond obstacles: feather-light YOLO11-LES for real-time ripeness detection of occluded strawberries in greenhouses[J]. Journal of Real-Time Image Processing, 2025, 22(5): 172-186.

点击下载PDF