基于YOLOv5s的轻量化朱鹮检测算法研究

张凡; 张鹏超; 王磊; 曹锐虎; 王晓鹏; 黄俊霖

doi:10.7652/xjtuxb202301011

您当前的位置：

首页 >

文章列表页 >

基于YOLOv5s的轻量化朱鹮检测算法研究

—————————————————————— | 更新时间：2025-07-09

- 基于YOLOv5s的轻量化朱鹮检测算法研究
- Research on Lightweight Crested Ibis Detection Algorithm Based on YOLOv5s
- 西安交通大学学报 2023年57卷第1期页码：110-121
- 作者机构：
  
  1. 陕西理工大学机械工程学院,陕西,汉中,723000
  2. 陕西省工业自动化重点实验室,陕西,汉中,723000
- 作者简介：
  
  : 2022-04-20。作者简介: 张凡(1996—),男,硕士生
  张鹏超(通信作者),男,教授,博士生导师。基金项目: 国家自然科学基金资助项目(62176146)。
- 基金信息：
  
  : 2022-04-20。作者简介: 张凡(1996—),男,硕士生;张鹏超(通信作者),男,教授,博士生导师。基金项目: 国家自然科学基金资助项目(62176146)。
- DOI：10.7652/xjtuxb202301011
  中图分类号： TP391.41
- 网络首发：2023-01-10，
  
  纸质出版：2023
- 稿件说明：
移动端阅览
张凡, 张鹏超, 王磊, 等. 基于YOLOv5s的轻量化朱鹮检测算法研究[J]. 西安交通大学学报, 2023,57(1):110-121.

ZHANG Fan, ZHANG Pengchao, WANG Lei, et al. Research on Lightweight Crested Ibis Detection Algorithm Based on YOLOv5s[J]. 2023, 57(1): 110-121.
张凡, 张鹏超, 王磊, 等. 基于YOLOv5s的轻量化朱鹮检测算法研究[J]. 西安交通大学学报, 2023,57(1):110-121. DOI： 10.7652/xjtuxb202301011.

ZHANG Fan, ZHANG Pengchao, WANG Lei, et al. Research on Lightweight Crested Ibis Detection Algorithm Based on YOLOv5s[J]. 2023, 57(1): 110-121. DOI： 10.7652/xjtuxb202301011.

摘要

针对当前朱鹮检测算法模型参数较多、计算量大的问题

本文设计研究了一种基于YOLOv5s的高性能轻量化网络模型。首先

结合EfficientNet网络中的MBConvBlock对原主干网络进行重构

大幅降低网络参数; 同时在浅层网络中采用Stem模块

提升浅层网络的特征提取能力; 然后改进卷积注意力模块(CBAM)

即将其中的通道注意力替换为高效通道注意力模块(ECA)

避免了降维操作

有效提取了邻近通道间的信息

且大幅降低了通道注意力的参数数量

并将其嵌入特征融合网络路径聚合网络(PANet)中

达到了引入微小参数数量而有效提升网络性能的目的

并将其命名为高效卷积注意力模块(ECBAM)。最后

在自建朱鹮数据集和公共数据集PASCAL VOC、COCO上进行实验

实验结果表明

与YOLOv5s算法相比

本文算法模型参数数量降低了52.37%

计算次数降低了54.55%

在自建朱鹮数据集上P

mAP

@0.5:0.95仅降低了约2个百分点

达到0.666

在公共数据集PASCAL VOC上P

mAP

@0.5达到0.792

在公共数据集COCO上P

mAP

@0.5:0.95达到0.298

证明了本文算法可在大幅降低网络计算量的同时保持较高的目标检测性能。

Abstract

Aiming at the problems of excessive parameters and heavy calculation workload in the current Crested Ibis detection algorithm model

this paper studies and designs a high-performance lightweight network model based on YOLOv5s. First

the original backbone network is reconstructed by referring to MBConvblock in the EfficientNet network to greatly reduce the network parameters; at the same time

the Stem module is used in the shallow network to i

mprove the feature extraction ability of the shallow network; then

the convolutional block attention module(CBAM)is improved

that is

the channel attention is replaced by the efficient channel attention module(ECA)

which avoids the dimensionality reduction operation

effectively extracts the information between adjacent channels

and greatly reduces the number of required parameters of channel attention; the ECA is embedded in the feature fusion network path aggregation network(PANet)for the purpose of introducing a small number of parameters to effectively improve the network performance

and is named as efficient convolutional block attention module(ECBAM). Finally

experiments are carried out on the self-built Crested Ibis data set and the public data sets Pascal VOC and COCO. The experimental results show that compared with YOLOv5s algorithm

the proposed algorithm reduces the number of required model parameters by 52.37% and the calculation workload by 54.55%; it decreases the P

mAP

@0.5:0.95 by only about 2% to 0.666 on the self-built Crested Ibis data set

and reduces the P

mAP

@0.5:0.95 to be 0.792 on the public data set Pascal VOC and 0.298 on the public data set COCO. Experiments show that the proposed algorithm can reduce the amount of network calculation workload while maintaining high detection performance.

关键词

Keywords

references

闫鲁, 黄治学. 35年来朱鹮的保护经验与发展思路 [J]. 野生动物学报, 2021, 42(3): 890-896.

YAN Lu, HUANG Zhixue. Thirty-five years conservation experience of crested ibis(Nipponia nippon)and future perspectives [J]. Chinese Journal of Wildlife, 2021, 42(3): 890-896.

陈保家, 陈学力, 沈保明, 等. CNN-LSTM深度神经网络在滚动轴承故障诊断中的应用 [J]. 西安交通大学学报, 2021, 55(6): 28-36.

CHEN Baojia, CHEN Xueli, SHEN Baoming, et al. An application of convolution neural network and long short-term memory in rolling bearing fault diagnosis [J]. Journal of Xi'an Jiaotong University, 2021, 55(6): 28-36.

闫彬, 樊攀, 王美茸, 等. 基于改进YOLOv5m的采摘机器人苹果采摘方式实时识别 [J/OL]. 农业机械学报[2022-07-21].http://kns.cnki.net/kcms/detail/11.1964.S.20220719.1604.008.html.

YAN Bin, FAN Pan, WANG Meirong, et al. Real time recognition of apple picking mode of picking robot based on improved YOLOv5m [J/OL]. Transactions of the Chinese Society for Agricultural Machinery[2022-07-21].http://kns.cnki.net/kcms/detail/11.1964.S.20220719.1604.008.html.

GIRSHICK R, DONAHUE J, DARRELL T, et al. Rich feature hierarchies for accurate object detection and semantic segmentation [C]//2014 IEEE Conference on Computer Vision and Pattern Recognition. Piscataway, NJ, USA IEEE, 2014: 580-587.

GIRSHICK R. Fast R-CNN [C]//2015 IEEE International Conference on Computer Vision(ICCV). Piscataway, NJ, USA: IEEE, 2015: 1440-1448.

REN Shaoqing, HE Kaiming, 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.

REDMON J, DIVVALA S, GIRSHICK R, et al. You only look once: unified, real-time object detection [C]//2016 IEEE Conference on Computer Vision and Pattern Recognition(CVPR). Piscataway, NJ, USA: IEEE, 2016: 779-788.

REDMON J, FARHADI A. YOLO9000: better, faster, stronger [C]//2017 IEEE Conference on Computer Vision and Pattern Recognition(CVPR). Piscataway, NJ, USA: IEEE, 2017: 6517-6525.

REDMON J, FARHADI A. YOLOv3: an incremental improvement [J/OL]. arXiv, 2018[2020-03-16].DOI:10.48550/arXiv.1804.02767.

BOCHKOVSKIY A, WANG C Y, LIAO H Y M. YOLOv4: optimal speed and accuracy of object detection [J/OL]. arXiv, 2020[2020-03-16].DOI:10.48550/arXiv.2004.10934.

GitHub, Inc. YOLOv5 [EB/OL]. [2021-06-15]. https://github.com/ultralytics/yolov5.

LIU Wei, ANGUELOV D, ERHAN D, et al. SSD: single shot MultiBox detector [C]//Computer Vision - ECCV 2016. Cham,Switzerland: Springer International Publishing, 2016: 21-37.

HAN S, MAO H, DALLY W J. Deep compression: compressing deep neural networks with pruning, trained quantization and Huffman coding [J]. Fiber, 2015, 56(4): 3-7.

JIN Xiaojie, YUAN Xiaotong, FENG Jiashi, et al. Training skinny deep neural networks with iterative hard thresholding methods [J/OL]. arXiv, 2016[2021-11-13].DOI:10.48550/arXiv.1607.05423.

WEN Wei, WU Chunpeng, WANG Yandan, et al. Learning structured sparsity in deep neural networks [C]//Proceedings of the 30th International Conference on Neural Information Processing Systems. Red Hook, NY, USA: Curran Associates Inc., 2016: 2082-2090.

HOWARD A G, ZHU Menglong, CHEN Bo, et al. MobileNets: efficient convolutional neural networks for mobile vision applications [J/OL]. arXiv, 2017[2021-11-16].DOI:10.48550/arXiv.1704.04861.

SANDLER M, HOWARD A, ZHU Menglong, et al. MobileNetV2: inverted residuals and linear bottlenecks [C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway, NJ, USA: IEEE, 2018: 4510-4520.

HOWARD A, SANDLER M, CHEN Bo, et al. Searching for MobileNetV3 [C]//2019 IEEE/CVF International Conference on Computer Vision(ICCV). Piscataway, NJ, USA: IEEE, 2019: 1314-1324.

ZHANG Xiangyu, ZHOU Xinyu, LIN Mengxiao, et al. ShuffleNet: an extremely efficient convolutional neural network for mobile devices [C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway, NJ, USA: IEEE, 2018: 6848-6856.

TAN Mingxing, LE Q. EfficientNet: rethinking model scaling for convolutional neural networks [C]//Proceedings of the 36th International Conference on Machine Learning. Chia Laguna Resort, Sardinia, Italy: PMLR, 2019: 6105-6114.

LIU Shu, QI Lu, QIN Haifang, et al. Path aggregation network for instance segmentation [C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition. Piscataway, NJ, USA: IEEE, 2018: 8759-8768.

WANG C Y, MARK LIAO H Y, WU Y H, et al. CSPNet: a new backbone that can enhance learning capability of CNN [C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops(CVPRW). Piscataway, NJ, USA: IEEE, 2020: 1571-1580.

HE Kaiming, ZHANG Xiangyu, REN Shaoqing, et al. Spatial pyramid pooling in deep convolutional networks for visual recognition [J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2015, 37(9): 1904-1916.

WANG Qilong, WU Banggu, ZHU Pengfei, et al. ECA-Net: efficient channel attention for deep convolutional neural networks [C]//2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition(CVPR). Piscataway, NJ, USA: IEEE, 2020: 11531-11539.

SZEGEDY C, IOFFE S, VANHOUCKE V, et al. Inception-v4, Inception-ResNet and the impact of residual connections on learning [J/OL]. arXiv, 2016 [2021-10-14].DOI:10.48550/arXiv.1602.07261.

WONG A, FAMUORI M, SHAFIEE M J, et al. YOLO nano: a highly compact you only look once convolutional neural network for object detection [C]//2019 15th Workshop on Energy Efficient Machine Learning and Cognitive Computing-NeurIPS Edition(EMC2-NIPS). Piscataway, NJ, USA: IEEE, 2019: 22-25.

QIN Zheng, LI Zeming, ZHANG Zhaoning, et al. ThunderNet: towards real-time generic object detection on mobile devices [C]//2019 IEEE/CVF International Conference on Computer Vision(ICCV). Piscataway, NJ, USA: IEEE, 2019: 6717-6726.

LONG Xiang, DENG Kaipeng, WANG Guanzhong, et al. PP-YOLO: an effective and efficient implementation of object detector [J/OL]. arXiv, 2020[2021-12-16].DOI:10.48550/arXiv.2007.12099.

GE Zheng, LIU Songtao, WANG Feng, et al. YOLOX: exceeding YOLO series in 2021 [J/OL]. arXiv, 2021[2022-03-04].DOI:10.48550/arXiv.2107.08430.

ZHANG Yuming, LEE C C, HSIEH J W, et al. CSL-YOLO: a new lightweight object detection system for edge computing [J/OL]. arXiv, 2021[2022-01-06].DOI:10.48550/arXiv.2107.04829.

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

轻质高刚度超结构超低频带隙机理及设计方法

复杂天气下铁轨异物入侵目标检测研究

利用MPEG-4视频流的运动目标检测

一种高分辨率遥感图像视感知目标检测算法

一种改进的单步多框目标检测算法