[1]周维,牛永真,王亚炜,等.基于改进的YOLOv4-GhostNet水稻病虫害识别方法[J].江苏农业学报,2022,38(03):685-695.[doi:doi:10.3969/j.issn.1000-4440.2022.03.014]
　ZHOU Wei,NIU Yong-zhen,WANG Ya-wei,et al.Rice pests and diseases identification method based on improved YOLOv4-GhostNet[J].,2022,38(03):685-695.[doi:doi:10.3969/j.issn.1000-4440.2022.03.014]
点击复制

基于改进的YOLOv4-GhostNet水稻病虫害识别方法()
分享到：

江苏农业学报[ISSN:1006-6977/CN:61-1281/TN]

卷:
38
期数:
2022年03期
页码:
685-695
栏目:
农业信息工程
出版日期:
2022-06-30

文章信息/Info

Title:
Rice pests and diseases identification method based on improved YOLOv4-GhostNet
作者:
周维; 牛永真; 王亚炜; 李丹
（东北林业大学信息与计算机工程学院，黑龙江哈尔滨150040）
Author(s):
ZHOU Wei; NIU Yong-zhen; WANG Ya-wei; LI Dan
(College of Information and Computer Engineering, Northeast Forestry University, Harbin 150040, China)
关键词:
水稻病虫害检测; GhostNet网络; YOLOv4; 轻量化; 迁移学习
Keywords:
rice diseases and pests detection; GhostNet network; YOLOv4; lightweight; transfer learning
分类号:
TP391.4
DOI:
doi:10.3969/j.issn.1000-4440.2022.03.014
文献标志码:
A
摘要:
针对水稻病虫害检测精度低、速度慢、模型复杂度高、部署困难等问题，改进了YOLOv4目标检测算法，结合轻量化GhostNet网络，提出了一种基于改进的YOLOv4-GhostNet水稻病虫害识别方法：1）利用幻象模块代替普通卷积结构，替换主干特征提取网络CSPDarkNet53，构建GhostNet模块进行图像的特征提取；2）改进YOLOv4网络的加强特征提取部分PANet结构；3）结合迁移学习与YOLOv4网络训练技巧。通过试验将YOLOv4及其MobileNet系列轻量化网络与Faster-RCNN系列网络和SSD系列网络进行对比，结果表明，改进的YOLOv4-GhostNet模型平均准确率达到79.38%，检测速度可达1 s 34.51 帧，模型权重大小缩减为42.45 MB，在保持检测精度达到较高水平的同时模型参数量大幅度降低，适用于部署在计算能力不足的嵌入式设备上。
Abstract:
Aiming at the problems of low accuracy, slow speed, high model complexity and difficult deployment in rice pests and diseases detection, the YOLOv4 target detection algorithm was improved. Combined with the lightweight GhostNet network, a rice pests and diseases recognition method based on the improved YOLOv4-GhostNet was proposed. The phantom module was used to replace the ordinary convolution structure, the backbone feature extraction network CSPDarkNet53 was replaced, and the GhostNet module was constructed for image feature extraction. The PANet structure of the enhanced feature extraction part of YOLOv4 network was improved. Transfer learning was combined with YOLOv4 network training skills. YOLOv4 and its MobileNet series lightweight networks were compared with Fast-RCNN series networks and SSD series networks. The results showed that the average accuracy of the improved YOLOv4-GhostNet model was 79.38%, the detection speed was 34.51 frames per second, and the weight of the model was reduced to 42.45 MB. While maintaining high detection accuracy, the number of model parameters is greatly reduced. It is suitable for embedded devices with insufficient computing power.

参考文献/References:

[1]商世吉,杨立群,赵庸洛,等. 水稻品种抗稻瘟病性鉴定方法的研究[J].东北农学院学报,1989(4):321-326.
[2]张亚玲,周万福,靳学慧. 黑龙江省稻瘟病菌与水稻品种的互作分析[J].东北农业大学学报,2011,42(1):28-33.
[3]张武，黄帅，汪京京，等. 复杂背景下小麦叶部病害图像分割方法研究[J].计算机工程与科学，2015，37（7）：1349-1354.
[4]马媛，冯全，杨梅. 基于HOG的酿酒葡萄叶检测[J].计算机工程与应用，2016，52（15）：158-161.
[5]鲍文霞,邱翔,胡根生,等. 基于椭圆型度量学习空间变换的水稻虫害识别[J].华南理工大学学报（自然科学版）,2020,48（10）:136-144.
[6]王映龙,戴香粮. 图像处理技术在水稻虫害系统中的应用[J].微计算机信息,2007（26）:274-275,256.
[7]XIA C, LEE J M , LI Y ,et al. Plant leaf detection using modified active shape models[J]. Biosystems Engineering, 2013,116（1）:23-35.
[8]YE M, CAO Z, YU Z , et al. Crop feature extraction from images with probabilistic superpixel Markov random field[J]. Computer & Electronics in Agriculture, 2015, 114:247-260.
[9]MOHANTY S P, HUGHES D P, MARCEL S. Using deep learning for image-based plant disease detection[J]. Frontiers in Plant Science, 2016,7: 14-19.
[10]FERENTINOS K P. Deep learning models for plant disease detection and diagnosis[J]. Computers and Electronics in Agriculture, 2018,145:311-318.
[11]房若民,沈凯文,李浩伟. MobileNet算法的嵌入式农业病虫害识别系统[J].单片机与嵌入式系统应用,2020,20(5):61-63.
[12]燕斌，周鹏，严利． 基于迁移学习的小样本农作物病害识别[J]. 现代农业科技，2019（6）:87-89.
[13]杨国国，鲍一丹，刘子毅． 基于图像显著性分析与卷积神经网络的茶园害虫定位与识别[J]．农业工程学报，2017,33（6）:156-162.
[14]张陶宁,陈恩庆,肖文福. 一种改进MobileNet_YOLOv3网络的快速目标检测方法[J].小型微型计算机系统,2021,42(5):1008-1014.
[15]李维刚,杨潮,蒋林,等. 基于改进YOLOv4算法的室内场景目标检测[J].激光与光电子学进展,2022,59（18）:1-19.
[16]曹远杰,高瑜翔. 基于GhostNet残差结构的轻量化饮料识别网络[J].计算机工程,2022,48(3):310-314.
[17]张官荣，陈相，赵玉，等． 面向小目标检测的轻量化 YOLOv3算法[J]．激光与光电子学进展, 2022,59(16):1-13.
[18]董艺威,于津. 基于SqueezeNet的轻量化卷积神经网络SlimNet[J].计算机应用与软件,2018,35(11):226-232.
[19]刘超军，段喜萍，谢宝文． 应用GhostNet卷积特征的ECO目标跟踪算法改进[J]．激光技术,2022,46(2):239-247.
[20]EVERINGHAM M, GOOL L V, WILLIAMS C, et al. The pascal visual object classes (VOC) challenge[J]. International Journal of Computer Vision, 2010, 88(2):303-338.
[21]WANG C Y, MARK LIAO H Y, WU Y H, et al. CSPNet: a new backbone that can enhance learning capability of CNN[C]//IEEE. 2020 IEEE/CVF conference on computer vision and pattern recognition workshops (CVPRW). Seattle, WA, USA: IEEE Press, 2020: 1571-1580.
[22]周睿璇,汪俊霖,孙宏,等. 一种改进YOLO v3的货车车轮检测算法[J].山西电子技术,2021(4):37-40.
[23]HE K, ZHANG X, REN S, 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.
[24]LIU S, QI L, QIN H, et al. Path aggregation network for instance segmentation[C]//IEEE. 2018 IEEE conference on computer vision and pattern recognition. Salt Lake City, UT, USA：IEEE Press, 2018: 8759-8768.
[25]LIN T Y, DOLLAR P, GIRSHICK R, et al. Feature pyramid networks for object detection[C]//IEEE. 2017 IEEE conference on computer vision and pattern recognition. Honolulu, HI, USA. IEEE Press, 2017: 936-944.
[26]张娜,贺兴时. 基于模拟退火的自适应正余弦算法[J].纺织高校基础科学学报,2021,34(1):84-90.
[27]HAN K, WANG Y, TIAN Q, et al. Ghostnet: More features from cheap operations[C]//IEEE. Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition. Washington, USA:IEEE Press, 2020: 1580-1589.
[28]李文婧,徐国伟,孔维刚,等. 基于改进YOLOv4的植物叶茎交点目标检测研究[J].计算机工程与应用,2022,58(4):221-228.

备注/Memo

备注/Memo:
收稿日期：2021-12-09基金项目：国家级大学生创新训练计划项目（41111214）作者简介：周维（2001-），女，吉林长春人，本科，研究方向为计算机视觉与病虫害目标检测。(E-mail)xingan_cangshu@nefu.edu.cn通讯作者：李丹，(E-mail)ld725725@126.com
更新日期/Last Update: 2022-07-07