显微图像中花粉颗粒的自动检测方法
Automatic Detection Methods for Pollen Granules in Microscopic Images
摘要: 针对显微图像花粉颗粒检测中人工效率低、主观性强及小目标边界模糊问题，本文提出高精度自动检测方法，构建了延安地区6类花粉显微图像数据集，完成912张原始图像、1351个目标标注，并通过亮度扰动、噪声注入、旋转翻转扩增至9120张；同时在YOLOv8基础上提出BiFPN-ATFL和Wavelet-CG-C2f两种改进模型。其中BiFPN-ATFL在多尺度检测中实现稳定增益，Wavelet-CG-C2f在复杂背景下平均检测精度达80.3%，召回率达82.1%，松科花粉mAP达98.0%。本研究利用小波卷积与上下文引导机制有效缓解高频细节丢失和密集粘连干扰，为花粉浓度预报与过敏原筛查提供了可靠技术路径。
Abstract: Addressing the inefficiency and strong subjectivity of manual inspection, and the problem of blurred boundaries in small-target pollen detection from microscopic images, this paper develops a high-precision automatic detection method. A microscopic image dataset of six pollen types from the Yan’an region is first constructed, involving 912 raw images with 1,351 annotated targets, and is expanded to 9120 samples through brightness perturbation, noise injection, rotation and flipping. Two improved YOLOv8-based models, BiFPN-ATFL and Wavelet-CG-C2f, are proposed. BiFPN-ATFL delivers reliable gains in multi-scale detection, while Wavelet-CG-C2f achieves an average detection accuracy of 80.3% and a recall of 82.1% under complex backgrounds, with the mAP for Pinaceae pollen reaching 98.0%. By integrating wavelet convolution and a context-guided mechanism, the approach effectively alleviates high-frequency detail loss and interferences from dense particle adhesion, offering a dependable technical pathway for pollen concentration forecasting and allergen screening.
文章引用：李子瑶, 何进荣. 显微图像中花粉颗粒的自动检测方法[J]. 计算机科学与应用, 2026, 16(6): 211-221. https://doi.org/10.12677/csa.2026.166221

参考文献

[1] 章初龙, 冯佳威, 赵蕊, 等. 禾本科植物内生真菌多样性及重要类群分类与系统发育研究进展[C]//中国菌物学会. 中国菌物学会2024年学术年会论文摘要. 2024: 34.
[2] 王斌功, 许海涛, 葛凤梅, 等. 热带生态区玉米自交系离体花粉活性的变化特征[J]. 热带农业科学, 2023, 43(8):56-66.
[3] 郑家华, 李健, 李清华, 等. 承德市区气传花粉浓度监测及意义[J]. 中国耳鼻咽喉头颈外科, 2021, 28(5): 301-304.
[4] 王莉, 涂晓娟, 马琳, 等. 北京八大处地区春季花粉过敏症的临床特点分析[J]. 中国急救复苏与灾害医学杂志, 2023, 18(2): 240-243.
[5] 王立志, 王德群. 安徽霍山产三种石斛花粉粒在扫描电子显微镜下的观察[J]. 安徽中医学院学报, 1989(1): 53.
[6] 李荣胜, 付庆帅, 郭俊红, 等. 中药松花粉的化学成分及现代应用研究进展[J]. 沈阳药科大学学报, 2024, 41(10): 1275-1286+1297.
[7] 王新娣, 石晓峰, 刘东彦, 等. 紫斑牡丹花粉乙酸乙酯部位化学成分研究[J]. 天然产物研究与开发, 2019, 31(11): 1912-1918.
[8] 肖梅, 张续德，高星星, 等. 2015年潍坊市春季气传花粉调查分析[J]. 中国临床医生杂志, 2016(6): 43-45.
[9] 胡伟倪, 朱丽, 谢立锋, 等. 北京市三年内每日花粉浓度与变应性鼻炎患者就诊比例的关联分析[J]. 中华耳鼻咽喉头颈外科杂志, 2017, 52(1): 31-36.
[10] Girshick, R. (2015) Fast R-CNN. 2015 IEEE International Conference on Computer Vision (ICCV), Santiago, 7-13 December 2015, 1440-1448. [Google Scholar] [CrossRef] 
[11] Ren, S., He, K., Girshick, R. and Sun, J. (2017) Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 39, 1137-1149. [Google Scholar] [CrossRef] [PubMed]
[12] Redmon, J., Divvala, S., Girshick, R. and Farhadi, A. (2016) You Only Look Once: Unified, Real-Time Object Detection. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, 27-30 June 2016, 779-788. [Google Scholar] [CrossRef] 
[13] Redmon, J. and Farhadi, A. (2017) YOLO9000: Better, Faster, Stronger. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, 21-26 July 2017, 6517-6525. [Google Scholar] [CrossRef] 
[14] Redmon, J. and Farhadi, A. (2018) YOLOv3: An Incremental Improvement. arXiv preprint arXiv:1804.02767. [Google Scholar] [CrossRef] 
[15] Wang, C.Y., Bochkovskiy, A. and Liao, H.Y.M. (2023) Yolov7: Trainable Bag-of-Freebies Sets New State-of-the-Art for Real-Time Object Detectors. 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Vancouver, 17-24 June 2023, 7464-7475. [Google Scholar] [CrossRef] 
[16] Reis, D., Kupec, J., Hong, J., et al. (2023) Real-Time Flying Object Detection with YOLOv8.
[17] Wang, G., Chen, Y., An, P., Hong, H., Hu, J. and Huang, T. (2023) UAV-YOLOv8: A Small-Object-Detection Model Based on Improved Yolov8 for UAV Aerial Photography Scenarios. Sensors, 23, Article 7190. [Google Scholar] [CrossRef] [PubMed]
[18] Chen, J., Mai, H., Luo, L., Chen, X. and Wu, K. (2021) Effective Feature Fusion Network in BIFPN for Small Object Detection. 2021 IEEE International Conference on Image Processing (ICIP), Anchorage, 19-22 September 2021, 699-703.
[19] Yang, B., Zhang, X., Zhang, J., Luo, J., Zhou, M. and Pi, Y. (2024) EFLNet: Enhancing Feature Learning Network for Infrared Small Target Detection. IEEE Transactions on Geoscience and Remote Sensing, 62, 1-11. [Google Scholar] [CrossRef] 
[20] Chen, R., Huang, X., Yang, L., et al. (2019) Intelligent Fault Diagnosis Based on CNN and Discrete Wavelet Transform. Computers in Industry, 106, 48-59.
[21] Guo, A., Jia, Z., Wang, J., Zhou, G., Ge, B. and Chen, W. (2024) A Lightweight Weed Detection Model with Global Contextual Joint Features. Engineering Applications of Artificial Intelligence, 136, Article 108903. [Google Scholar] [CrossRef]