基于改进RT-DETR的苹果病害检测算法

肖永帅; 陈若冰; 时雷; 郑光; 尹飞

doi:10.13300/j.cnki.hnlkxb.2026.03.008

您当前的位置：

首页 >

文章列表页 >

基于改进RT-DETR的苹果病害检测算法

农业人工智能专题 | 更新时间：2026-11-05

- 基于改进RT-DETR的苹果病害检测算法
- Apple disease detection algorithm based on improved RT-DETR
- 华中农业大学学报 2026年45卷第3期页码：87-97
- 作者机构：
  
  河南农业大学信息与管理科学学院，郑州 450046
- 作者简介：
  
  肖永帅，E-mail：17657318695@163.com
  尹飞，E-mail：yin.fei@henau.edu.cn
- 基金信息：
  
  河南省科技攻关项目(242102521027);河南省科技研发计划联合基金项目(222301420113)
- DOI：10.13300/j.cnki.hnlkxb.2026.03.008
  中图分类号： TP391.41
- 收稿：2025-10-10，
  
  纸质出版：2026-05-25
- 稿件说明：
移动端阅览
肖永帅，陈若冰，时雷，等.基于改进RT-DETR的苹果病害检测算法［J］.华中农业大学学报，2026，45（3）：87‐97.

XIAO Yongshuai,CHEN Ruobing,SHI Lei,et al.Apple disease detection algorithm based on improved RT-DETR[J].Journal of Huazhong Agricultural University,2026,45(03):87-97.
肖永帅，陈若冰，时雷，等.基于改进RT-DETR的苹果病害检测算法［J］.华中农业大学学报，2026，45（3）：87‐97. DOI： 10.13300/j.cnki.hnlkxb.2026.03.008.

XIAO Yongshuai,CHEN Ruobing,SHI Lei,et al.Apple disease detection algorithm based on improved RT-DETR[J].Journal of Huazhong Agricultural University,2026,45(03):87-97. DOI： 10.13300/j.cnki.hnlkxb.2026.03.008.

摘要

针对复杂果园环境下苹果病害检测中存在的背景干扰强、多尺度病斑识别困难以及模型轻量化部署需求，提出一种基于改进RT-DETR的轻量化苹果病害检测模型EGA-DETR。模型从3个方面进行了改进：设计RGLAN轻量化特征聚合模块，通过特征分流与重参数化卷积减少冗余计算并增强特征表达；构建BiFPN-GLSA多尺度特征融合模块，通过双向特征传递和全局-局部自注意力机制提升不同尺度病斑的表征能力；引入Inner-Shape-IoU损失函数，以增强模型对不规则病斑目标的定位能力。试验结果显示，EGA-DETR在苹果病害数据集上取得了91.8%的精确率、88.9%的召回率和92.1%的mAP

，较基线RT-DETR-18分别提升3.5、3.5、1.5百分点。同时，模型参数量降至11.8×10

，较基线模型减少40.4%，推理速度达到120帧/s。综上所述，EGA-DETR模型在检测精度与计算效率之间实现了较优平衡，可为苹果病害精准、实时检测提供技术支撑。

Abstract

To address the challenges of strong background interference，difficulty in recognizing multi-scale lesions，and the need for lightweight deployment in apple disease detection within complex orchard environments，an improved lightweight apple disease detection model named EGA-DETR was proposed based on RT-DETR. The model was enhanced in three key aspects. First，an RGLAN lightweight feature aggregation module was designed to reduce redundant computations and enhance feature representation through feature splitting and re-parameterized convolution. Second，a BiFPN-GLSA multi-scale feature fusion module was developed to improve the representation of lesions at various scales via b

idirectional feature transmission and a global–local self-attention mechanism. Third，an Inner-Shape-IoU loss function was introduced to improve model’s localization accuracy for irregularly shaped lesion targets. Experimental results showed that EGA-DETR achieved 91.8% precision，88.9% recall，and 92.1% mAP

on the apple disease dataset，with improvements of 3.5，3.5，and 1.5 percentage points over the baseline RT-DETR-18，respectively. Meanwhile，the number of model parameters was reduced to 11.8×10

，which is 40.4% fewer than that of the baseline model，and the inference speed reached 120 frames per second. In summary，the EGA-DETR model achieved a favorable balance between detection accuracy and computational efficiency，providing robust technical support for accurate and real-time apple disease detection.

关键词

Keywords

references

郑超杰，李少波，蒲睿强，等.基于轻量化卷积神经网络的番茄叶片病害识别［J］.江苏农业科学，2024，52（11）：225-231. ZHENG C J，LI S B，PU R Q，et al.Tomato leaf disease recognition based on lightweight convolutional neural network［J］.Jiangsu agricultural sciences，2024，52（11）：225-231 （in Chinese with English abstract） .

杜英杰，宗哲英，王祯，等.农作物病害诊断方法现状和展望［J］.江苏农业科学，2023，51（6）：16-23. DU Y J，ZONG Z Y，WANG Z，et al.Current situation and prospect of diagnostic methods for crop diseases［J］.Jiangsu agricultural sciences，2023，51（6）：16-23 （in Chinese with English abstract） .

BOCHKOVSKIY A，WANG C Y，LIAO H M.YOLOv4：optimal speed and accuracy of object detection［PP/OL］.arXiv， 2020 ：2004.10934［2025-10-10］.https：//arxiv.org/abs/2004.10934.

赵晓，杨梦婷，张懿丹.基于改进YOLO v10n的苹果叶片病害检测方法［J］.农业机械学报，2025，56（9）：657-666，728. ZHAO X，YANG M T，ZHANG Y D.Apple leaf disease detection method based on improved YOLO v10n［J］.Transactions of the CSAM，2025，56（9）：657-666，728 （in Chinese with English abstract） .

曾林涛，马嘉昕，丁羽，等.基于改进YOLO v8的苹果叶部病害检测方法［J］.江苏农业科学，2025，53（5）：147-156. ZENG L T，MA J X，DING Y，et al.An apple leaf disease detection method based on improved YOLO v8［J］.Jiangsu agricultural sciences，2025，53（5）：147-156 （in Chinese with English abstract） .

LIU L，OUYANG W L，WANG X G，et al. Deep learning for generic object detection： a survey［J］. International journal of computer vision，2020，128（2）： 261-318.

VASWANI A，SHAZEER N，PARMAR N，et al. Attention is all you need［C］//Proceedings of the 31st International Conference on Neural Information Processing Systems. Long Beach： Curran Associates，2017： 6000-6010.

CARION N，MASSA F，SYNNAEVE G，et al.End-to-end object detection with transformers［C］//Computer Vision–ECCV 2020.Cham：Springer，2020：213-229.

JI W，ZHAI K L，XU B，et al. Green apple detection method based on multidimensional feature extraction network model and transformer module［J/OL］. Journal of food protection，2025，88（1）： 100397［2025-10-10］. https：//doi.org/10.1016/j.jfp.2024.100397.

ZHANG L X，ZHAO Y，ZHOU C C，et al.JuDifformer：multimodal fusion model with transformer and diffusion for jujube disease detection［J/OL］.Computers and electronics in agriculture，2025，232：110008［2025-10-10］.https：//doi.org/10.1016/j.compag.2025.110008.

ZHAO Y A，LV W Y，XU S L，et al.DETRs beat YOLOs on real-time object detection［C］//2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition （CVPR）.Seattle：IEEE，2024：16965-16974.

PFE. Apple disease detection computer vision model［DS/OL］. （2022）［2025-10-10］. https：//universe.roboflow.com/pfe-ns5wl/apple-disease-detection-jspkx.

TAN M X，PANG R M，LE Q V.EfficientDet：scalable and efficient object detection［C］//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition （CVPR）.Seattle：IEEE，2020：10778-10787.

TANG F L，XU Z X，HUANG Q M，et al.DuAT：dual-aggregation transformer network forMedical image segmentation［C］//Pattern Recognition and Computer Vision.Singapore：Springer，2024：343-356.

REZATOFIGHI H，TSOI N，GWAK J，et al.Generalized intersection over union：a metric and a loss for bounding box regression［C］//2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition （CVPR）.Long Beach：IEEE，2020：658-666.

ZHANG H，XU C，ZHANG S J.Inner-IoU：more effective intersection over union loss with auxiliary bounding box［PP/OL］.2023：arXiv：2311.02877［2025-10-10］.https：//arxiv.org/abs/2311.02877.

ZHANG H，ZHANG S J.Shape-IoU：more accurate metric considering bounding box shape and scale［PP/OL］.arXiv，2023：2312.17663［2025-10-10］.https：//arxiv.org/abs/2312.17663.

XIAO J H，GUO H Q，ZHOU J，et al. Tiny object detection with context enhancement and feature purification［J/OL］. Expert systems with applications，2023，211：118665［2025-10-10］. https：//doi.org/10.1016/j.eswa.2022.118665.

ZHENG Z H，WANG P，LIU W，et al.Distance-IoU loss：faster and better learning for bounding box regression［J］.Proceedings of the AAAI conference on artificial intelligence，2020，34（7）：12993-13000.

SELVARAJU R R，COGSWELL M，DAS A，et al.Grad-CAM：visual explanations from deep networks via gradient-based localization［J］.International journal of computer vision，2020，128（2）：336-359.

浏览量

下载量

CSCD

文章被引用时，请邮件提醒。

提交

工具集

关联资源

不同杀菌剂对百合炭腐病菌的毒力和防效评价

生物质灰添加量对自蔓延合成基质块性能的影响

农业生物质增值化利用技术与装备团队

大豆灰斑病病原变异、抗性挖掘与智能防控

3个不同倍性油茶CoSAD家族的全基因组鉴定及表达分析