|本期目录/Table of Contents|

[1]王鑫泽,何超,方国文,等.改进YOLO v8n的无人机高分辨率水稻幼苗数目检测方法[J].江苏农业科学,2025,53(5):95-104.
 Wang Xinze,et al.Detection of rice seedling number by high-resolution UAV with improved YOLO v8n[J].Jiangsu Agricultural Sciences,2025,53(5):95-104.
点击复制

改进YOLO v8n的无人机高分辨率水稻幼苗数目检测方法(PDF)
分享到:

《江苏农业科学》[ISSN:1002-1302/CN:32-1214/S]

卷:
第53卷
期数:
2025年第5期
页码:
95-104
栏目:
智能农业装备
出版日期:
2025-03-05

文章信息/Info

Title:
Detection of rice seedling number by high-resolution UAV with improved YOLO v8n
作者:
王鑫泽何超方国文黄立闩李熠璇张晓青
西南林业大学机械与交通学院,云南昆明 650000
Author(s):
Wang Xinzeet al
关键词:
水稻幼苗数量检测KW卷积WIoUYOLO v8n
Keywords:
-
分类号:
S126;TP391.41
DOI:
-
文献标志码:
A
摘要:
针对我国水稻机播率不足50%,高速机播出现操作不当、对秧池判断失误、天气引起存活降低等原因出现空穴率的问题,导致无法准确判断水稻密度并实施精确补插计划,提出了一种低成本的改进YOLO v8n算法的轻量化识别水稻幼苗数目的方法。首先,使用动态KWConv代替普通Conv,使卷积参数更加高效,卷积更加适合水稻幼苗数据集;其次,设计全新的C2f_KW模块代替部分C2f,减少模型的计算量,并提高了模型的检测精度;最后,使用WIoU边界损失函数解决了CIoU损失函数对样本质量较差的局限性。试验结果表明,改进YOLO v8n模型的mAP0.5mAP0.5~0.95分别为99.19%、72.56%;相比原模型YOLO v8n,mAP0.5mAP0.5~0.95分别提高0.50、4.13百分点,并且模型计算量从8.2 G降到5.8 G,实现了更广泛的无人机部署条件,改进YOLO v8n模型与主流模型Faster R-CNN、YOLO v5n、YOLO v5s、YOLO v7、YOLO v7-tiny、YOLO v8n-Swin Transformer等相比有着不同程度提升,提高35.27、5.01、2.69、9.54、13.14、3.45百分点。本研究方法对无人机监测统计水稻幼苗提供了有效的支持,为水稻种植相关研究提供了借鉴。
Abstract:
-

参考文献/References:

[1]李文博. 稻田综合种养对水稻产量和品质的影响[D]. 合肥:安徽农业大学,2021:10-21.
[2]Li Z Y,Wang R,Zhang W,et al. Multiscale features supported DeepLab v3+ optimization scheme for accurate water semantic segmentation[J]. IEEE Access,2019,7:155787-155804.
[3]汤以胜,孙晓敏,陈前,等. 基于遥感大数据云计算平台下的水稻识别研究[J]. 航天返回与遥感,2022,43(3):113-123.
[4]周莹. 基于深度学习的遥感影像去云方法研究[D]. 哈尔滨:东北林业大学,2022:21-29.
[5]辛明金,邬立岩,宋玉秋,等. VP6型水稻插秧机作业质量试验研究[J]. 中国农机化学报,2016,37(4):19-23,48.
[6]段小斌. 基于无人机遥感技术的水稻倒伏区域识别研究[J]. 农机化研究,2021,43(12):225-228.
[7]Sandler M,Howard A,Zhu M L,et al. MobileNet v2:inverted residuals and linear bottlenecks[C]//IEEE. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City,2018:4510-4520.
[8]栾鑫鑫,翟晨,安焕炯,等. 应用分子光谱信息融合判别不同产地大米[J]. 光谱学与光谱分析,2023,43(9):2818-2824.
[9]Haq M A,Rahaman G,Baral P,et al. Deep learning based supervised image classification using UAV images for forest areas classification[J]. Journal of the Indian Society of Remote Sensing,2021,49(3):601-606.
[10]胡田田,赵璐,崔晓路,等. 无人机多光谱数据可靠性分析与冬小麦产量估算研究[J]. 农业机械学报,2023,54(12):217-225.
[11]王怡婧,丁启东,张俊华,等. 基于无人机高光谱遥感和机器学习的土壤水盐信息反演[J]. 应用生态学报,2023,34(11):3045-3052.
[12]Sharpe P. Open tech workstation[J]. Industrial and Commercial Training,1985,17(3):18-20.
[13]Joos L,de Tender C,Holderbeke A,et al. Exploring the microbial response as a potential bio-indicator for soil health:insights from a controlled incubator experiment[J]. Agriculture,Ecosystems & Environment,2023,356:108634.
[14]Du J B,Cheng W J,Li S L. Joint task offloading and resource allocation in mixed edge/cloud computing and blockchain empowered device-free sensing systems[J]. Computer Communications,2023,209:38-46.
[15]朱伟,马立新,张平,等. 基于GoogLeNet和无人机图像的水稻秧苗形态识别[J]. 华南农业大学学报,2022,43(3):99-106.
[16]慕涛阳,赵伟,胡晓宇,等. 基于改进的DeepLab v3+模型结合无人机遥感的水稻倒伏识别方法[J]. 中国农业大学学报,2022,27(2):143-154.
[17]郭柏璋,牟琦,冀汶莉. 融合注意力机制的YOLO v5深度神经网络杂草识别方法[J]. 无线电工程,2023,53(12):2771-2782.
[18]慕君林,马博,王云飞,等. 基于深度学习的农作物病虫害检测算法综述[J]. 农业机械学报,2023,54(增刊2):301-313.
[19]曹英丽,赵雨薇,杨璐璐,等. 基于改进DeepLab v3+的水稻田间杂草识别方法[J]. 农业机械学报,2023,54(12):242-252.
[20]Lin T Y,Dollár P,Girshick R,et al. Feature pyramid networks for object detection[C]//IEEE. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).Honolulu,2017:936-944.
[21]Liu S,Qi L,Qin H F,et al. Path aggregation network for instance segmentation[C]//IEEE. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City,2018:8759-8768.
[22]Ge Z,Liu S,Wang F,et al. YOLO x:Exceeding yolo series in 2021[EB/OL]. (2018-08-06)[2024-10-20]. https://arxiv.org/abs/2107.08430.
[23]Law H,Deng J. CornerNet:detecting objects as paired keypoints[J]. International Journal of Computer Vision,2020,128(3):642-656.
[24]Li C,Yao A B,Li C,et al. KernelWarehouse:towards parameter-efficient dynamic convolution[EB/OL]. (2023-08-16)[2023-11-20]. https://arxiv.org/abs/2308.08361v1.
[25]Ding X H,Zhang X Y,Han J G,et al. Diverse branch block:building a convolution as an inception-like unit[C]//IEEE. 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).Nashville,2021:10881-10890.
[26]Zhu X Z,Hu H,Lin S,et al. Deformable ConvNets v2:more deformable,better results[C]//IEEE. 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).Long Beach,2019:9300-9308.
[27]Hu M,Feng J Y,Hua J S,et al. Online convolutional reparameterization[C]//IEEE. 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).New Orleans,2022:558-567.
[28]Li J F,Wen Y,He L H. SCConv:spatial and channel reconstruction convolution for feature redundancy[C]//IEEE. 2023 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR).Vancouver,2023:6153-6162.

相似文献/References:

[1]方淑梅,梁喜龙,纪伟波,等.外源NO对盐碱胁迫下水稻幼苗生长抑制的缓解作用[J].江苏农业科学,2013,41(08):67.
 Fang Shumei,et al.Mitigative effect of exogenous nitric oxide on growth inhibition of rice seedlings under saline-alkali stress[J].Jiangsu Agricultural Sciences,2013,41(5):67.
[2]侯丽霞.CaCl2浸种对低温胁迫下水稻幼苗生理指标的影响[J].江苏农业科学,2013,41(08):70.
 Hou Lixia,et al.Effect of soaking seeds with CaCl2 on physiological indicators of rice seedlings[J].Jiangsu Agricultural Sciences,2013,41(5):70.
[3]刘涛,郑欣,刘夏囡,等.超氧阴离子调节水稻幼苗不定根形成的分子机制[J].江苏农业科学,2017,45(13):54.
 Liu Tao,et al.Molecular mechanism of superoxide anion regulating adventitious root formation of rice seedlings[J].Jiangsu Agricultural Sciences,2017,45(5):54.
[4]付朝晋,王红春,张善学,等.新型植物诱抗剂12%寡糖·香草硫缩病醚微乳剂对水稻幼苗生长的影响[J].江苏农业科学,2021,49(16):80.
 Fu Chaojin,et al.Influences of 12% oligosaccharide·vanilla thioacetal ether microemulsion on growth of rice seedlings[J].Jiangsu Agricultural Sciences,2021,49(5):80.
[5]赵颖,何志刚,曲航,等.不同秸秆育苗基质对水稻幼苗生长和根际微环境的影响[J].江苏农业科学,2023,51(10):100.
 Zhao Ying,et al.Effects of different straw rearing substrates on rice seedling growth and rhizosphere microenvironment[J].Jiangsu Agricultural Sciences,2023,51(5):100.

备注/Memo

备注/Memo:
收稿日期:2023-12-21
基金项目:国家自然科学基金(编号:51968065);云南省科技厅农业联合专项(编号:202301BD070001-077);云南省高层次人才项目(编号:YNWR-QNBJ-2018-066、YNQR-CYRC-2019-001)。
作者简介:王鑫泽(1999—),男,安徽南陵人,硕士研究生,研究方向为图像识别算法在农业上的应用。E-mail:xinzewang0328@163.com。
通信作者:何超,博士,教授,博士生导师,研究方向为智能农机装备研究等。E-mail:hcsmile@163.com。
更新日期/Last Update: 2025-03-05