[1]罗云涛,张志安.基于改进YOLO v8模型的轻量化梨果实检测[J].江苏农业学报,2026,42(03):582-590.[doi:doi:10.3969/j.issn.1000-4440.2026.03.016]
　LUO Yuntao,ZHANG Zhian.Lightweight pear fruit detection based on an improved YOLO v8 model[J].,2026,42(03):582-590.[doi:doi:10.3969/j.issn.1000-4440.2026.03.016]
点击复制

基于改进YOLO v8模型的轻量化梨果实检测()

江苏农业学报[ISSN:1006-6977/CN:61-1281/TN]

卷:
42
期数:
2026年03期
页码:
582-590
栏目:
农业信息工程
出版日期:
2026-03-31

文章信息/Info

Title:
Lightweight pear fruit detection based on an improved YOLO v8 model
作者:
罗云涛; 张志安
(南京理工大学机械工程学院,江苏南京210094)
Author(s):
LUO Yuntao; ZHANG Zhi’an
(School of Mechanical Engineering, Nanjing University of Science & Technology, Nanjing 210094, China)
关键词:
梨; 果实; YOLO v8模型; 检测精度; 检测速率; 计算量; 参数量
Keywords:
pear; fruit; YOLO v8 model; detection accuracy; detection speed; computational load; parameter count
分类号:
TP391.4
DOI:
doi:10.3969/j.issn.1000-4440.2026.03.016
文献标志码:
A
摘要:
为了解决梨果实采摘机器人目标检测模型在复杂果园环境下检测效果不佳，且难以部署于资源受限嵌入式平台的问题,本研究提出一种基于改进YOLO v8模型的轻量化梨果实检测模型:设计一种PR_Bottleneck模块,通过创新结构设计,有效削减模型参数量和计算量,显著降低模型运行负担；构建SPPF_CBAM模块,大幅增强模型对目标区域的聚焦能力,显著提升其在复杂场景下的目标识别精度；引入ADown下采样模块,降低模型复杂度和参数量,增强下采样过程的信息保留能力；引入WIoU损失函数,使模型在训练过程中的回归损失函数能够集中在平均质量的锚框,从而提高检测器的整体性能指标,增强模型的泛化能力。试验结果表明,与YOLO v8原始模型相比,改进模型精确率、召回率、平均精度均值（mAP50）分别提高了0.7个百分点、3.7个百分点、1.7个百分点,其参数量和计算量分别降低38.72%和30.86%,模型体积缩小38.33%，检测速率为1 s 85.4帧,较原始模型提升1 s 12.8帧。因此,改进YOLO v8轻量化模型不仅提高了检测精度和检测速率,而且明显降低了计算量和参数量,能够部署在嵌入式平台上快速有效进行梨果实的实时目标检测。
Abstract:
To address the challenges of poor detection performance of target detection models for pear-picking robots in complex orchard environments and the difficulty of deployment on resource-constrained embedded platforms, a lightweight pear fruit detection model based on an improved YOLO v8 is proposed in this study. First, a novel PR_Bottleneck module was introduced, which employed an innovative structural design to effectively reduce the model’s parameter count and computational load, significantly easing the operational burden. Second, the SPPF_CBAM module was proposed to greatly enhance the model’s focus on target regions, thereby significantly improving detection accuracy in complex scenarios. The model also incorporated an ADown downsampling module, which not only reduced model complexity and parameter count but also strengthened information retention during the downsampling process. Finally, the WIoU loss function was introduced, which enabled the model to concentrate on medium-quality anchor boxes during training, thus improving overall detection performance and enhancing the model’s generalization capability. Experimental results showed that compared with the original YOLO v8 model, the precision, recall, and mean average precision (mAP50) of the improved model increased by 0.7, 3.7, and 1.7 percentage points, respectively. The parameter count and computational load were reduced by 38.72% and 30.86%, respectively. The model size was reduced by 38.33%, and the detection speed reached 85.4 frames per second, an improvement of 12.8 frames per second over the original model. Therefore, the improved YOLO v8 lightweight model not only improves detection accuracy and speed, but also significantly decreases the computational load and parameter count, enabling rapid and effective real-time pear fruit detection on embedded platforms.

参考文献/References:

[1]常怡勇. 秋天食梨保健康[J]. 健康生活,2016(9):44-45.
[2]李孟哲,程少丽,孟照刚. 我国主要果树发展现状与对策建议[J]. 农业科技通讯,2023(9):11-16.
[3]谢铮辉,郑倩,姚伟. 智能农业发展回顾及中国智能农业展望[J]. 热带农业科学,2019,39(3):125-133.
[4]黄凯奇,任伟强,谭铁牛. 图像物体分类与检测算法综述[J]. 计算机学报,2014,37(6):1225-1240.
[5]SENGUPTA S, LEE W S. Identification and determination of the number of immature green Citrus fruit in a canopy under different ambient light conditions[J]. Biosystems Engineering,2014,117:51-61.
[6]GAO F F, FU L S, ZHANG X, et al. Multi-class fruit-on-plant detection for apple in SNAP system using Faster R-CNN[J]. Computers and Electronics in Agriculture,2020,176:105634.
[7]LODDO A, LODDO M, DI RUBERTO C. A novel deep learning based approach for seed image classification and retrieval[J]. Computers and Electronics in Agriculture,2021,187:106269.
[8]GIRSHICK R, DONAHUE J, DARRELL T, et al. Rich feature hierarchies for accurate object detection and semantic segmentation[C]//IEEE. 2014 IEEE conference on computer vision and pattern recognition. Columbus,OH,USA:IEEE,2014.
[9]GIRSHICK R. Fast R-CNN[C]//IEEE. 2015 IEEE international conference on computer vision. Santiago,Chile:IEEE,2015.
[10]REN S Q, HE K M, 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.
[11]闫建伟,赵源,张乐伟,等. 改进Faster-RCNN自然环境下识别刺梨果实[J]. 农业工程学报,2019,35(18):143-150.
[12]LIU W, ANGUELOV D, ERHAN D, et al. SSD:single shot MultiBox detector[C]//ECCV. Computer vision-ECCV 2016. Cham,Switzerland:ECCV,2016.
[13]REDMON J, DIVVALA S, GIRSHICK R, et al. You only look once: unified,real-time object detection[C]//IEEE. 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR). Las Vegas,NV,USA:IEEE,2016.
[14]REDMON J, FARHADI A. YOLO9000:better,faster,stronger[C]//IEEE. 2017 IEEE conference on computer vision and pattern recognition (CVPR). Honolulu,HI,USA:IEEE,2017.
[15]吕石磊,卢思华,李震,等. 基于改进YOLO v3-LITE轻量级神经网络的柑橘识别方法[J]. 农业工程学报,2019,35(17):205-214.
[16]LI X, PAN J D, XIE F P, et al. Fast and accurate green pepper detection in complex backgrounds via an improved Yolov4-tiny model[J]. Computers and Electronics in Agriculture,2021,191:106503.
[17]桂泽春,赵思健. 人工智能在农业风险管理中的应用研究综述[J]. 智慧农业(中英文),2023,5(1):82-98.
[18]WANG C Y, BOCHKOVSKIY A, LIAO H M. YOLO v7:trainable bag-of-freebies sets new state-of-the-art for real-time object detectors[C]//IEEE/CVF. 2023 IEEE/CVF conference on computer vision and pattern recognition (CVPR). Vancouver,BC,Canada: IEEE,2023.
[19]徐彦威,李军,董元方,等. YOLO系列目标检测算法综述[J]. 计算机科学与探索,2024,18(9):2221-2238.
[20]LIU S, QI L, QIN H F, et al. Path aggregation network for instance segmentation[C]//IEEE/CVF. 2018 IEEE/CVF conference on computer vision and pattern recognition. Salt Lake City,UT,USA: IEEE,2018.
[21]CHEN J R, KAO S H, HE H, et al. Run,don’t walk:chasing higher FLOPS for faster neural networks[C]//IEEE/CVF. 2023 IEEE/CVF conference on computer vision and pattern recognition (CVPR). Vancouver,BC,Canada:IEEE,2023.
[22]HE K M, ZHANG X Y, REN S Q, et al. Deep residual learning for image recognition[C]//IEEE. 2016 IEEE conference on computer vision and pattern recognition (CVPR). Las Vegas,NV,USA,IEEE:2016.
[23]WOO S, PARK J, LEE J Y, et al. CBAM:convolutional block attention module[C]//ECCV. Computer vision-ECCV 2018. Cham,Switzerland:ECCV,2018.
[24]周飞燕,金林鹏,董军. 卷积神经网络研究综述[J]. 计算机学报,2017,40(6):1229-1251.
[25]李嘉诚,陈中举,许浩然. 基于改进YOLO v8n的真实场景下草莓病害检测方法[J]. 中国农机化学报,2024,45(12):267-274.
[26]ZHENG Z H, WANG P, REN D W, et al. Enhancing geometric factors in model learning and inference for object detection and instance segmentation[J]. IEEE Transactions on Cybernetics,20,52(8):8574-8586.
[27]TONG Z, CHEN Y, XU Z, et al. Wise-IoU:bounding box regression loss with dynamic focusing mechanism[EB/OL]. (2023-04-08)
[2024-03-10]. https://arxiv.org/abs/2301.10051.

相似文献/References:

[1]王冬群,胡寅侠,华晓霞.慈溪市梨农药残留膳食摄入风险评估[J].江苏农业学报,2016,(03):698.[doi:10.3969/j.issn.1000-4440.2016.03.033]
　WANG Dong-qun,HU Yin-xia,HUA Xiao-xia.Dietary intake risk assessment of pesticide residue in pears in Cixi city[J].,2016,(03):698.[doi:10.3969/j.issn.1000-4440.2016.03.033]
[2]王春雷,高季平,赵宪坤.利用酶联免疫反应鉴定梨树S基因型[J].江苏农业学报,2015,(05):1124.[doi:doi:10.3969/j.issn.1000-4440.2015.05.028]
　WANG Chun-lei,GAO Ji-ping,ZHAO Xian-kun.S genotyping in pear by PCR-ELISA[J].,2015,(03):1124.[doi:doi:10.3969/j.issn.1000-4440.2015.05.028]
[3]许园园,李晓刚,李慧,等.梨CDPK基因家族全基因组序列鉴定分析[J].江苏农业学报,2015,(03):659.[doi:10.3969/j.issn.1000-4440.2015.03.031]
　XU Yuan-yuan,LI Xiao-gang,LI Hui,et al.Identification of calcium-dependent protein kinase(CDPK) gene family in pear（Pyrus bretschneideri Rehd）[J].,2015,(03):659.[doi:10.3969/j.issn.1000-4440.2015.03.031]
[4]王端,景晨娟,陈雪峰,等.华北杏不同品种（品系）果实发育阶段香气成分分析[J].江苏农业学报,2021,(04):974.[doi:doi:10.3969/j.issn.1000-4440.2021.04.021]
　WANG Duan,JING Chen-juan,CHEN Xue-feng,et al.Analysis on aroma components in fruits of different cultivars (lines) of North China apricot during developmental stages[J].,2021,(03):974.[doi:doi:10.3969/j.issn.1000-4440.2021.04.021]
[5]谢昶琰,王迪,安祥瑞,等.滴灌减量施肥对梨树体养分及果实产量、品质的影响[J].江苏农业学报,2021,(06):1526.[doi:doi:10.3969/j.issn.1000-4440.2021.05.022]
　XIE Chang-yan,WANG Di,AN Xiang-rui,et al.Effects of drip irrigation and reducing fertilization on tree nutrient, fruit yield and quality of pear[J].,2021,(03):1526.[doi:doi:10.3969/j.issn.1000-4440.2021.05.022]
[6]施露,高庆超,李亚辉,等.梨果实品质的研究进展与潜在技术应用展望[J].江苏农业学报,2022,38(02):567.[doi:doi:10.3969/j.issn.1000-4440.2022.02.033]
　SHI Lu,GAO Qing-chao,LI Ya-hui,et al.Research progress and prospects of potential technology application in pear fruit quality[J].,2022,38(03):567.[doi:doi:10.3969/j.issn.1000-4440.2022.02.033]
[7]高磊,李慧,洪奔,等.葡萄品种钟山红玉与钟山红杂交后代的SSR鉴定及葡萄果实性状遗传倾向分析[J].江苏农业学报,2023,(01):187.[doi:doi:10.3969/j.issn.1000-4440.2023.01.022]
　GAO Lei,LI Hui,HONG Ben,et al.SSR identification and genetic tendency analysis of fruit characters of hybrid offspring of Zhongshan Hongyu and Zhongshan Hong[J].,2023,(03):187.[doi:doi:10.3969/j.issn.1000-4440.2023.01.022]
[8]毕清芮,马新院,薛玉冉,等.基于SSR标记的MCID法鉴定新疆地方梨品种及其遗传多样性分析[J].江苏农业学报,2024,(10):1933.[doi:doi:10.3969/j.issn.1000-4440.2024.10.018]
　BI Qingrui,MA Xinyuan,XUE Yuran,et al.Identification of local pear cultivars in Xinjiang Uygur Autonomous Region by MCID method based on SSR markers and analysis of their genetic diversity[J].,2024,(03):1933.[doi:doi:10.3969/j.issn.1000-4440.2024.10.018]
[9]潘超,肖逸,耿献辉.劳动力要素对梨生产比较优势的影响机制及空间效应[J].江苏农业学报,2025,(02):372.[doi:doi:10.3969/j.issn.1000-4440.2025.02.018]
　PAN Chao,XIAO Yi,GENG Xianhui.Influence mechanism and spatial effect of labor factors on the comparative advantage of pear production[J].,2025,(03):372.[doi:doi:10.3969/j.issn.1000-4440.2025.02.018]
[10]乔慧兰,马铭泽,武爱园,等.基于代谢组揭示雪兔和妙香7号草莓果实香气物质的差异[J].江苏农业学报,2025,(04):744.[doi:doi:10.3969/j.issn.1000-4440.2025.04.013]
　QIAO Huilan,MA Mingnze,WU Aiyuan,et al.Aroma compounds differences in strawberry fruits of Snow Rabbit and Miaoxiang No.7 revealed at metabolome level[J].,2025,(03):744.[doi:doi:10.3969/j.issn.1000-4440.2025.04.013]

备注/Memo

备注/Memo:
收稿日期：2025-05-02基金项目：江苏省现代农机装备与技术示范推广项目(NJ2023-13)作者简介：罗云涛(2001-),男,湖北荆州人,硕士研究生,研究方向为机器视觉、机器人控制技术。(E-mail)luoyuntao0422@163.com通讯作者：张志安,(E-mail)zzayoyo@163.com
更新日期/Last Update: 2026-04-17