Scientia Agricultura Sinica ›› 2025, Vol. 58 ›› Issue (18): 3598-3615.doi: 10.3864/j.issn.0578-1752.2025.18.003

• TILLAGE & CULTIVATION·PHYSIOLOGY & BIOCHEMISTRY·AGRICULTURE INFORMATION TECHNOLOGY • Previous Articles     Next Articles

Detection Method of Leaf Tip in Wheat Seedling Stage Based on Improved YOLOv8s

HE HaoXu2,3(), GAO Xiang1,3(), RAO Yuan2,3, ZHANG ZiRui3,4, WU Gong5, HOU YiTing2,3, HE Ye1, LI XinYi1   

  1. 1 College of Resources and Environment, Anhui Agricultural University, Hefei 230036
    2 College of Information and Artificial Intelligence, Anhui Agricultural University, Hefei 230036
    3 Key Laboratory of Agricultural Sensors, Ministry of Agriculture and Rural Affairs, Hefei 230036
    4 College of Engineering, Anhui Agricultural University, Hefei 230036
    5 College of Agronomy, Anhui Agricultural University, Hefei 230036
  • Received:2025-02-27 Accepted:2025-07-15 Online:2025-09-16 Published:2025-09-18
  • Contact: GAO Xiang

RichHTML

13

PDF

134

Abstract:

【Objective】In precision agriculture, the detection of crop seedlings can be interfered with by factors such as soil weeds, occlusion between seedling leaves, and multi-scale datasets. Based on the object detection algorithm, this paper improved the YOLOv8s algorithm and designed the wheat leaf tip detection model YOLO-Wheat to solve problems, such as leaf occlusion of wheat seedlings in the field, interference from soil weeds, and multi-view data with multiple scales, thereby enhancing the accuracy of wheat seedling leaf detection and providing a theoretical basis for wheat seedling detection at the seedling stage in precision agriculture. 【Method】Close-up and distant images of wheat seedlings were collected respectively through mobile phone cameras and on-board RGB cameras during the emergence period to construct a crop image dataset. In the network model, a pyramid structure of multi-scale feature fusion (high-level screening-feature fusion pyramid, HS-FPN) was adopted. This structure used high-level features as weights, filters low-level feature information through the channel attention module, and combined the screened features with the high-level features. Enhancing the feature expression ability of the model could effectively solve the problem of multi-scale data. Integrate the efficient local attention (ELA) local attention mechanism in the network model was used to enable the model to focus on the leaf tip information of wheat and to suppress the interference of soil background factors of weeds. Meanwhile, the loss function of YOLOv8s (complete IoULoss, CIoULoss) was optimized, and the inner-Iou Loss auxiliary bounding box loss function was introduced to enhance the network's attention to small targets and to improve the positioning accuracy of wheat leaf tips. In terms of training strategies, transfer learning was employed. The model was pre-trained using close-up images of wheat leaf tips, and then the parameters of the model were updated and optimized using distant images. 【Result】The YOLO-Wheat model was compared with five object detection models, namely Faster-RCNN, YOLOv5s, YOLOv7, YOLOv8s, and YOLOv9s. The YOLO-Wheat model was the best in wheat leaf tip detection, with a recognition accuracy rate of 92.7% and a recall rate of 85.1%, respectively. The mean Average Precision (mAP) values were 82.9%. Compared with the Faster-RCNN, YOLOv5s, YOLOv7, YOLOv8s and YOLOv9s models, the recognition accuracy mAP values of YOLO-Wheat have increased by 17.1%, 13.6%, 11.0%, 8.7% and 3.8% respectively; the recall rates increased by 13.1%, 6.7%, 4.5%, 1.8% and 1.3%, respectively. Compared with the Faw-RCNN, YOLOv5s, YOLOv7, YOLOv8s and YOLOv9s models, the mAP values of YOLO-Wheat have increased by 16.2%, 9.8%, 5.0%, 5.9% and 0.7%, respectively. 【Conclusion】This method could effectively solve the problem of multi-scale data, achieve precise detection of small targets at the leaf tips of wheat seedlings in complex field environments using unmanned aerial vehicle (UAV) images, and provide technical support and theoretical reference for intelligent leaf counting of wheat seedlings in complex fields.

Key words: wheat, YOLOv8s, loss function, transfer learning, drone

Fig. 1

Some images of wheat seedlings collected under natural conditions a, b, and c are close-up images of wheat seedlings taken by mobile phones, d, e, and f are distant images of wheat seedlings taken by drones"

Fig. 2

Network structure diagram of YOLO-Wheat model"

Fig. 3

HS-FPN structure"

Fig. 4

SFF module structure"

Fig. 5

ELA structure diagram"

Fig. 6

Inner-IoU"

Fig. 7

Comparison of detection results of Faster-RCNN，YOLOv5s，YOLOv7，and YOLOv9 a：Faster-RCNN；b：YOLOv9s；c：YOLOv7；d：YOLOv5s"

Fig. 8

Comparison of detection results between YOLOv8s and YOLO-Wheat"

Fig. 9

Comparison of mAP, accuracy and recall index of six models"

Table 1

Results of comparison between this research model and mainstream models"

模型 Model 准确率 Precision (%) R (%) mAP (%)
Faster-RCNN 75.6 70.7 66.7
YOLOv5s 79.1 77.1 73.1
YOLOv7 81.7 80.6 77.9
YOLOv8s 84.0 83.3 77.0
YOLOv9s 88.9 83.8 82.2
YOLO-Wheat 92.7 85.1 82.9

Table 2

The effects of data enhancement on detection results"

数据处理
Data processing		 准确率
Precision (P)
(%)		 平均精度
Average accuracy (AP)
(%)
无数据增强
Data-free enhancement		 82.6 80.8
有数据增强
Active data enhancement		 83.3 81.2

Table 3

Ablation results"

HS-FPN ELA Inner-IoU P (%) mAP (%)
× × × 84.0 78.0
√ × × 86.4 79.6
× √ × 85.3 78.8
× × √ 85.5 79.1
√ × √ 88.3 80.6
√ √ × 88.1 79.4
× √ √ 87.7 81.3
√ √ √ 92.7 82.9

Fig. 10

Detection results of the improved YOLOv8s and YOLOv9s models"

Table 4

Experimental results of different feature fusion strategies"

多尺度融合模块
Multi-scale fusion module		 AP
(%)		 mAP
(%)
FPN 83.7 79.1
BiFPN 85.0 79.2
PaFPN 85.3 79.9
HS-FPN 87.3 80.1

Fig. 11

Comparison of mAP and AP indicators of different feature fusion policies"

Table 5

Experimental results of different loss functions"

IoU类别
IoU category		 精确率
Precision (%)		 mAP
(%)
GIoU 89.5 77.1
DIoU 89.3 77.8
CIoU 88.6 75.2
Inner-IoU 90.1 79.6

Table 6

Experimental results of different attention mechanisms"

注意力机制 Attention mechanism mAP@50 (%) mAP@95 (%)
CA 68.20 63.07
SE 67.73 62.09
CBAM 66.38 62.48
ELA 71.58 63.58

Fig. 12

Comparison of different attention mechanisms mAP@50 and mAP@95"

Fig. 13

Heat map visualization of different attention mechanisms"

Fig. 14

The fitting results of the number of wheat leaf tips to the actual value"

[1]
LIU H, WANG Z H, YU R, LI F C, LI K Y, CAO H B, YANG N, LI M H, DAI J, ZAN Y L, et al. Optimal nitrogen input for higher efficiency and lower environmental impacts of winter wheat production in China. Agriculture, Ecosystems & Environment, 2016, 224: 1-11.
[2]
IGREJAS G, BRANLARD G. The importance of wheat. Wheat Quality for Improving Processing and Human Health. Cham: Springer International Publishing, 2020: 1-7.
[3]
杨锋, 姚晓通. 基于改进YOLOv8的小麦叶片病虫害检测轻量化模型. 智慧农业(中英文), 2024, 6(1): 147-157.
YANG F, YAO X T. Lightweight model of wheat leaf pest detection based on improved YOLOv8. Smart Agriculture, 2024, 6(1): 147-157. (in Chinese)
[4]
余卓炫. 基于Mask R-CNN的小麦叶片展开检测系统设计[D]. 雅安: 四川农业大学, 2023.
YU Z X. Design of detection system for wheat leaf unfoldment based on Mask R-CNN[D]. Yaan: Sichuan Agricultural University, 2023. (in Chinese)
[5]
张洁, 林立杰, 苏红玉, 杨予熙, 吴波明. 基于YOLOv8s-seg的水稻叶片检测研究//中国植物病理学会. 中国植物病理学会2024年学术年会论文集, 2024: 377.
ZHANG J, LIN L J, SUN H Y, YANG Y X, WU B M. Study on rice leaf detection based on YOLOv8s-seg //Plant Pathology Society of China. Proceedings of the 2024 Annual Conference of the Chinese Society of Plant Pathology, 2024: 377. (in Chinese)
[6]
安江勇, 黎万义, 李茂松. 基于Mask R-CNN的玉米干旱卷曲叶片检测. 中国农业信息, 2019, 31(5): 66-74.
AN J Y, LI W Y, LI M S. Identification of maize drought rolled leaves based on Mask R-CNN model. China Agricultural Informatics, 2019, 31(5): 66-74. (in Chinese)
[7]
凌慕菲, 杨冬风. 基于改进YOLOv5s模型的玉米叶片病害识别. 农业灾害研究, 2023, 13(8): 126-128.
LING M F, YANG D F. Maize leaf disease recognition based on improved YOLOv5s model. Agricultural Disaster Research, 2023, 13(8): 126-128. (in Chinese)
[8]
杨琳琳, 王建坤, 别书凡, 白振江, 常金攀, 贺小梅, 李文峰, 施杰. 基于计算机视觉的玉米叶片表型检测方法研究. 江苏农业科学, 2023, 51(16): 195-202.
YANG L L, WANG J K, BIE S F, BAI Z J, CHANG J P, HE X M, LI W F, SHI J. Study on maize leaf phenotype detection method based on computer vision. Jiangsu Agricultural Sciences, 2023, 51(16): 195-202. (in Chinese)
[9]
赵弋秋. 基于无人机影像的大豆苗情快速检测方法研究[D]. 郑州: 河南农业大学, 2022.
ZHAO Y Q. Soybean's rapid detection method based on the UAV image research[D]. Zhengzhou: Henan Agricultural University, 2022. (in Chinese)
[10]
侯依廷, 饶元, 宋贺, 聂振君, 王坦, 何豪旭. 复杂大田场景下基于改进YOLOv8的小麦幼苗期叶片数快速检测方法. 智慧农业(中英文), 2024, 6(4): 128-137.
HOU Y T, RAO Y, SONG H, NIE Z J, WANG T, HE H X. A rapid detection method for wheat seedling leaf number in complex field scenarios based on improved YOLOv8. Smart Agriculture, 2024, 6(4): 128-137. (in Chinese)
[11]
邢忠健. 基于改进YOLOv5s模型的玉米叶片病害检测方法研究与应用[D]. 合肥: 安徽农业大学, 2023.
XING Z J. Research and application of maize leaf disease detection method based on improved YOLOv5s model[D]. Hefei: Anhui Agricultural University, 2023. (in Chinese)
[12]
LI Y X, JIANG Z C, ZHANG Z, LI H, ZHANG M. SeedingsNet:Field wheat seedling density detection based on deep learning. Sensing Technologies for Field and In-House Crop Production. Singapore: Springer Nature Singapore, 2023: 77-88.
[13]
刘家俊, 苏润琦, 吴麒, 许吉如. 基于YOLO模型的玉米叶片病虫害检测与识别研究. 化学工程与装备, 2023(6): 31-34.
LIU J J, SU R Q, WU Q, XU J R. Study on detection and identification of maize leaf diseases and insect pests based on YOLO model. Chemical Engineering & Equipment, 2023(6): 31-34. (in Chinese)
[14]
MA H, ZHAO W, JI J, JIN X, SHI Y, ZHENG F, LI N. A quick counting method for winter wheat at the seedling stage in fields based on an improved yolov4 model. The Journal of Animal and Plant Sciences, 2022(6): 1666-1681.
[15]
ZHANG Y, WANG H F, XU R X, YANG X Y, WANG Y C, LIU Y L. High-precision seedling detection model based on multi-activation layer and depth-separable convolution using images acquired by drones. Drones, 2022, 6(6): 152.
[16]
GUO X Q, GE Y J, LIU F Q, YANG J J. Identification of maize and wheat seedlings and weeds based on deep learning. Frontiers in Earth Science, 2023, 11: 1146558.
[17]
LIU T, YANG T L, LI C Y, LI R, WU W, ZHONG X C, SUN C M, GUO W S. A method to calculate the number of wheat seedlings in the 1st to the 3rd leaf growth stages. Plant Methods, 2018, 14: 101.

doi: 10.1186/s13007-018-0369-5 pmid: 30473722
[18]
LIU C, YIN Y Y, QIAN R, WANG S H, XIA J J, ZHANG J K, ZHAO L Q. Enhanced winter wheat seedling classification and identification using the SETFL-ConvNeXt model: Addressing overfitting and optimizing training strategies. Agronomy, 2024, 14(9): 1914.
[19]
LI Y L, ZHAN X H, LIU S Y, LU H, JIANG R B, GUO W, CHAPMAN S, GE Y F, SOLAN B, DING Y F, BARET F. Self-supervised plant phenotyping by combining domain adaptation with 3D plant model simulations: application to wheat leaf counting at seedling stage. Plant Phenomics, 2023, 5: 0041.
[20]
KONG S L, LI J, ZHAI Y T, GAO Z Y, ZHOU Y, XU Y L. Real-time detection of crops with dense planting using deep learning at seedling stage. Agronomy, 2023, 13(6): 1503.
[21]
CUI J R, ZHENG H, ZENG Z W, YANG Y L, MA R J, TIAN Y Y, TAN J W, FENG X, QI L. Real-time missing seedling counting in paddy fields based on lightweight network and tracking-by-detection algorithm. Computers and Electronics in Agriculture, 2023, 212: 108045.
[22]
许鑫, 李海洋, 冯洋洋, 马新明, 沈帅杰, 乔新昱. 基于K-means和Harris角点检测的麦苗识别研究. 河南农业科学, 2020, 49 (12): 164-171.
XU X, LI H Y, FENG Y Y, MA X M, SHEN S J, QIAO X Y. Wheat seedling identification based on K-means and Harris corner detection. Journal of Henan Agricultural Sciences, 2020, 49(12): 164-171. (in Chinese)
[23]
ZHUANG F Z, QI Z Y, DUAN K Y, XI D B, ZHU Y C, ZHU H S, XIONG H, HE Q. A comprehensive survey on transfer learning. Proceedings of the IEEE, 2021, 109(1): 43-76.
[24]
CHEN Y F, ZHANG C Y, CHEN B, HUANG Y Y, SUN Y F, WANG C M, FU X J, DAI Y X, QIN F W, PENG Y, GAO Y. Accurate leukocyte detection based on deformable-DETR and multi-level feature fusion for aiding diagnosis of blood diseases. Computers in Biology and Medicine, 2024, 170: 107917.
[25]
XU W, WAN Y. Ela: Efficient local attention for deep convolutional neural networks. arXiv Preprint arXiv: 2403.01123, 2024.
[26]
ZHANG H, XU C, ZHANG S J. Inner-IoU: More effective intersection over union loss with auxiliary bounding box. arXiv preprint arXiv:2311.02877, 2023.
[27]
REN S Q, HE K M, GIRSHICK R, SUN J. Faster R-CNN: Towards real-time object detection with region proposal networks. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2017, 39(6): 1137-1149.

doi: 10.1109/TPAMI.2016.2577031 pmid: 27295650
[28]
JOCHER G. (2020). YOLOv5 by Ultralytics (Version 7.0)[Computer software]. https://doi.org/10.5281/zenodo.3908559
[29]
WANG C Y, BOCHKOVSKIY A, LIAO H M. 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). June 17-24, 2023, Vancouver, BC, Canada. IEEE, 2023: 7464-7475.
[30]
JOCHER G, CHAURASIA A, QIU J. (2023). Ultralytics YOLO (version 8.0.0)[Computer software]. https://github.com/ultralytics/ultralytics.
[31]
WANG C Y, YEH I H, LIAO H Y M. YOLOv9: Learning what you want to learn using programmable gradient information. arXiv preprint arXiv:2402.13616, 2024.
[32]
LIN T Y, DOLLÁR P, GIRSHICK R, HE K M, HARIHARAN B, BELONGIE S. Feature pyramid networks for object detection. 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). July 21-26, 2017, Honolulu, HI, USA. IEEE, 2017: 936-944.
[33]
TAN M X, PANG R M, LE Q V. EfficientDet: Scalable and efficient object detection. // 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Seattle, WA, USA. IEEE, 2020: 10778-10787.
[34]
LIU S, QI L, QIN H F, SHI J P, JIA J Y. Path aggregation network for instance segmentation. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City, UT, USA. IEEE, 2018: 8759-8768.
[35]
HOU Q B, ZHOU D Q, FENG J S. Coordinate attention for efficient mobile network design. 2021 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR). Nashville, TN, USA. IEEE, 2021: 13708-13717.
[36]
HU J, SHEN L, SUN G. Squeeze-and-excitation networks. 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Salt Lake City, UT, USA. IEEE, 2018: 7132-7141.
[37]
WOO S, PARK J, LEE J Y, KWEON I S. CBAM: Convolutional block attention module.Computer Vision - ECCV 2018. Cham: Springer International Publishing, 2018: 3-19.
[38]
SELVARAJU R R, COGSWELL M, DAS A, VEDANTAM R, PARIKH D, BATRA D. Grad-CAM: Visual explanations from deep networks via gradient-based localization. International Journal of Computer Vision, 2020, 128(2): 336-359.
[1] ZHU Qi, JIA ZhenPeng, Tahir SHAH, XU ChenSheng, LI ZhiQi, LÜ HuiShuai, ZHU PengChao, WEI XiaoMin, HUANG DongLin, SUN YanNi, CAO WeiDong, GAO YaJun, WANG ZhaoHui, ZHANG DaBin. Green Manure Crops Combined with Enhanced-Efficiency Products Reduced Greenhouse Gas Emissions and Carbon Footprints in Dryland Wheat Fields [J]. Scientia Agricultura Sinica, 2026, 59(7): 1507-1522.
[2] LI WenHu, LI HaiFeng, DU YuPeng, DING YuLan, LUO YiNuo, LI YuKe, SHE WenTing, ZHANG Feng, TENG Yu, ZHANG SiQi, HUANG Cui, LI XiaoHan, LIU JinShan, WANG ZhaoHui. Regional Differences in Wheat Zinc Uptake and Translocation Responses to Soil Zinc Fertilization [J]. Scientia Agricultura Sinica, 2026, 59(5): 1034-1047.
[3] JIAO WenJuan, HE WanLong, GENG HongWei, BAI Bin, LI JianFeng, CHENG YuKun. Stripe Rust Resistance Evaluation and Molecular Characterization of Yr Genes for 155 Spring Wheat Varieties (Lines) [J]. Scientia Agricultura Sinica, 2026, 59(5): 937-950.
[4] CUI ShiYou, CHEN PengJun, MIAO YuanQing, HAN JiJun, SHEN JunMing. Development and Field Evaluation of Glyphosate-Resistant Wheat Germplasm Generated Through EMS Mutagenesis [J]. Scientia Agricultura Sinica, 2026, 59(4): 723-733.
[5] QIAN Jin, LI YingXue, WU Fang, ZOU XiaoChen. Improved Leaf Phosphorus Content Estimation of Winter Wheat Using Ensemble Hyperspectral Dimensionality Reduction Method [J]. Scientia Agricultura Sinica, 2026, 59(4): 781-792.
[6] KONG Yuan, CUI ShaSha, LI Mei, LI Jian, YANG SiYu, FANG Feng, LIU ShuaiShuai, LIU MingPing, ZENG Yan, GAO XingXiang, BAI LianYang. Spatiotemporal Distribution Dynamics of Five Grass Weed Species Including Lolium multiflorum in Winter Wheat Fields of the Huang- Huai-Hai Region [J]. Scientia Agricultura Sinica, 2026, 59(4): 807-823.
[7] WANG YongSheng, NIU Li, WANG ChangJie, MA LiHua, LIAN XiaoXiao, MENG YaXiong, MA XiaoLe, YAO LiRong, ZHANG Hong, YANG Ke, LI BaoChun, WANG HuaJun, SI ErJing, WANG JunCheng. Genome-Wide Association Study and Candidate Gene Identification for Thousand Grain Weight in Winter Wheat [J]. Scientia Agricultura Sinica, 2026, 59(3): 499-514.
[8] LI XinYi, LI JiaNing, YANG WenPing, XIA Qing, HUO YingRui, HAO ShiHang, HUANG TingMiao, REN YongKang, CHEN Jie, GAO ZhiQiang, YANG ZhenPing. Effects of Post-Anthesis Foliar Zinc Application on Zinc Nutrition in Colored-Grain Wheat [J]. Scientia Agricultura Sinica, 2026, 59(3): 515-527.
[9] XIAN QingLin, XIAO JianKe, GAO AQing, GAO LiChuang, LIU Yang. Effects of Planting Patterns Combined with Soil Moisture Measurement and Supplementary Irrigation on the Yield and Water Use Efficiency of Winter Wheat [J]. Scientia Agricultura Sinica, 2026, 59(3): 589-601.
[10] ZHANG ZhiYong, TAN ShiChao, XIONG ShuPing, MA XinMing, WEI YiHao, WANG XiaoChun. Effects of Annual Water and Nitrogen Optimization on Yield and Nitrogen Migration of Wheat-Maize Rotation System in Irrigation Area of Northern Henan [J]. Scientia Agricultura Sinica, 2026, 59(2): 336-353.
[11] LÜ XuDong, SUN ShiYuan, LI YaNan, LIU YuLong, WANG YanQun, FU Xin, ZHANG JiaYing, NING Peng, PENG ZhengPing. Effects of Intelligent Mechanized Layered Fertilization on Root-Soil Nutrient Distribution and Yield in Wheat Fields [J]. Scientia Agricultura Sinica, 2026, 59(1): 129-146.
[12] LU Hao, ZHANG MingLong, HAN Mei, YAN QingBiao, LI ZhengPeng, YIN Wen, FAN ZhiLong, HU FaLong, CHAI Qiang. Green Manure Returning via Sheep Digest with Nitrogen Fertilizer Reduction are Beneficial to Improve Wheat Yield and Soil Quality at Qinghai-Tibet Plateau [J]. Scientia Agricultura Sinica, 2026, 59(1): 147-160.
[13] YE MeiJin, CHEN JiaTing, ZHOU JieGuang, YIN Li, HU XinRong, LAN YuXin, CHEN Bin, SU LongXing, LIU JiaJun, LIU TianChao, LI XiaoYu, MA Jian. Identification, Validation and Genetic Effect Analysis of Major QTL for Spike Density in Wheat [J]. Scientia Agricultura Sinica, 2026, 59(1): 17-28.
[14] PU LiXia, ZHANG JiaRui, YE JianPing, HUANG XiuLan, FAN GaoQiong, YANG HongKun. The Combined Effects of 16, 17-Dihydro Gibberellin A5 and Straw Mulching on Tillering and Grain Yield of Dryland Wheat [J]. Scientia Agricultura Sinica, 2025, 58(9): 1735-1748.
[15] LI YunLi, DIAO DengChao, LIU YaRui, SUN YuChen, MENG XiangYu, WU ChenFang, WANG Yu, WU JianHui, LI ChunLian, ZENG QingDong, HAN DeJun, ZHENG WeiJun. Genome-Wide Association Study of Heat Tolerance at Seedling Stage in A Wheat Natural Population [J]. Scientia Agricultura Sinica, 2025, 58(9): 1663-1683.
Viewed
Full text


Abstract

Cited
  Shared   
  Discussed   
No Suggested Reading articles found!
京ICP备09089781号-30
Copyright 2018 © Editorial office of Scientia Agricultura Sinica
Tel: 86-010-82106279    E-mail: zgnykx@mail.caas.net.cn
Supported by Beijing Magtech Co.Ltd