Image intelligent recognition method of crop pests

doi:10.6048/j.issn.1001-4330.2023.12.013

Abstract

Abstract:

【Objective】Aiming at the problems of few crop pest dataset samples, low accuracy of existing single model in crop pest identification and poor generalization ability, a pest identification model based on transfer learning and multi-model integration is proposed.【Methods】Experiments were carried out on the large-scale public crop pest dataset IP102. In this study, transfer learning is used to train 6 deep neural networks separately, and the combination of EfficientNet, Vision Transformer, Swin Transformer and ConvNeXt with better recognition performance was selected, and then different strategies were used to integrate the prediction results.【Results】The results showed that the recognition accuracy of the proposed method based on transfer learning and multi-model integration reached 75.75%, which was 1.34% higher than that of the best-performing single-model ConvNeXt, and was comparable to the current compared with the performance of the optimal algorithm (CA-EfficientNet) on the dataset, the recognition accuracy was 6.3% higher, 【Conclusion】It has better stability and generalization ability.

Key words: pest identification; IP102 dataset; transfer learning; ensemble learning; voting method

摘要：

【目的】针对作物害虫数据集样本较少、现有单一模型在作物害虫识别上的准确率不高以及泛化能力较差的问题,提出一种基于迁移学习与多模型集成的害虫识别模型。【方法】在大规模公开作物害虫数据集IP102上进行试验,使用迁移学习单独训练6个深层神经网络,选择识别性能较好的EfficientNet、Vision Transformer、Swin Transformer和ConvNeXt进行组合,采用不同策略集成预测结果。【结果】提出的基于迁移学习与多模型集成方法的识别准确率达到75.75%,比性能最好的单模型ConvNeXt提高了1.34%,与目前该数据集上最优算法(CA-EfficientNet)的性能相比,识别准确率高出了6.3%。【结论】害虫图像智能识别模型具有较好的稳定性与泛化能力。

关键词: 害虫识别, IP102数据集, 迁移学习, 集成学习, 投票法

CLC Number:

TP391.41
S41

LI Yuqing, CHEN Yanhong, LI Yongke, XIAO Tianci, LI Qingyuan. Image intelligent recognition method of crop pests[J]. Xinjiang Agricultural Sciences, 2023, 60(12): 2973-2981.

李雨晴, 陈燕红, 李永可, 肖天赐, 李清源. 作物害虫图像智能识别方法[J]. 新疆农业科学, 2023, 60(12): 2973-2981.

Figures/Tables 8

References 23

[1]	Strange R N, Scott P R. Plant disease: a threat to global food security[J]. Annual Review of Phytopathology, 2005, 43(1):83-116. DOI URL
[2]	傅隆生, 宋珍珍, 李瑞, 等. 深度学习方法在农业信息中的研究进展与应用现状[J]. 中国农业大学学报, 2020, 25(2):105-120.
	FU Longsheng, SONG Zhenzhen, LI rui, et al. Application and research progress of deep learning in agriculture[J]. Journal of China Agriculture University, 2020, 25(2):105-120.
[3]	Deng L, Wang Y, Han Z, et al. Research on insect pest image detection and recognition based on bio-inspired methods[J]. Biosystems Engineering, 2018, 169:139-148. DOI URL
[4]	张永玲, 姜梦洲, 俞佩仕, 等. 基于多特征融合和稀疏表示的农业害虫图像识别方法[J]. 中国农业科学, 2018, 51(11):2084-2093. DOI
	ZHANG Yongling, JIANG Mengzhou, YU Peishi, et al. Agricultural pest identification based on multi-Feature fusion and sparse representation[J]. Scientia Agricultura Sinica, 2018, 51(11):2084-2093. DOI
[5]	马鹏鹏. 基于图像的水稻灯诱害虫智能识别与计数技术的研究[D]. 杭州: 浙江理工大学, 2018.
	MA Pengpeng. The research on intelligent identification and counting technology of rice light-trap pests based on images[D]. Hangzhou: Zhejiang Sci-Tech University, 2018.
[6]	曾伟辉, 唐欣, 胡根生, 等. 基于卷积块注意力胶囊网络的小样本水稻害虫识别[J]. 中国农业大学学报, 2022, 27(3):63-74.
	ZENG Weihui, TANG Xin, HU Gensheng, et al. Rice pests recognition with small number of samples based on CBAM and capsule network[J]. Journal of China Agricultural University, 2022, 27(3):63-74.
[7]	Huang M L, Chuang T C, Liao Y C. Application of transfer learning and image augmentation technology for tomato pest identification[J]. Sustainable Computing: Informatics and Systems, 2022, 33:100646.
[8]	Reza M T, Mehedi N, Tasneem N A, et al. Identification of crop consuming insect pest from visual imagery using transfer learning and data augmentation on deep neural network[C]// 2019 22nd International Conference on Computer and Information Technology. IEEE, 2019:1-6.
[9]	华月珊, 王佳新, 戎洁庆, 等. 基于ResNet网络的小蠢虫识别研究[J]. 林业与环境科学, 2021, 37(6):124-129.
	HUA Yueshan, WANG Jiaxin, RONG Jieqing, et al. Research on scolytidae recognition and identification based on ResNet network[J]. Forestry and Environmental Science, 2021, 37(6):124 -129.
[10]	Wu X, Zhan C, Lai Y K, et al. Ip102: A large-scale benchmark dataset for insect pest recognition[C]//Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2019:8787-8796.
[11]	Pan S J, Yang Q. A survey on transfer learning[J]. IEEE Transactions on Knowledge and Data Engineering, 2009, 22(10):1345-1359. DOI URL
[12]	罗勇琛. 融合空间和通道注意力机制的害虫图像识别方法研究[D]. 长春: 吉林农业大学, 2021.
	LUO Yongchen. Pest recognition based on spatial and channel attention mechanismusing field images[D]. Changchun: Jilin Agricultural University, 2021.
[13]	Han K, Wang Y, Chen H, et al. A survey on vision transformer[J]. IEEE Transactions on Pattern Analysis and Machine Intelligence, 2022.
[14]	Tan M, Le Q. Efficientnet: Rethinking model scaling for convolutional neural networks[C]// International conference on machine learning. PMLR, 2019:6105-6114.
[15]	Dosovitskiy A, Beyer L, Kolesnikov A, et al. An image is worth 16x16 words: Transformers for image recognition at scale[J]. arXiv preprint arXiv: 2010.11929, 2020.
[16]	Liu Z, Lin Y, Cao Y, et al. Swin transformer: Hierarchical vision transformer using shifted windows[C]//Proceedings of the IEEE/CVF International Conference on Computer Vision. 2021:10012-10022.
[17]	Liu Z, Mao H, Wu C Y, et al. A ConvNet for the 2020s[J]. arXiv preprint arXiv: 2201.03545, 2022.
[18]	周志华. 机器学习[M]. 北京: 清华大学出版社, 2016.
	ZHOU Zhihua. Machine learning[M]. Beijing: Tsinghua University Press, 2016.
[19]	Nanni L, Maguolo G, Pancino F. Insect pest image detection and recognition based on bio-inspired methods[J]. Ecological Informatics, 2020, 57:101089. DOI URL
[20]	Ayan E, Erbay H, Varçin F. Crop pest classification with a genetic algorithm-based weighted ensemble of deep convolutional neural networks[J]. Computers and Electronics in Agriculture, 2020, 179:105809. DOI URL
[21]	甘雨, 郭庆文, 王春桃, 等. 基于改进EfficientNet模型的作物害虫识别[J]. 农业工程学报, 2022, 38(1):203-211.
	GAN Yu, GUO Qingwen, WANG Chuntao, et al. Recognizing crop pests using an improved EfficientNet model[J]. Transactions of the Chinese Society of Agricultural Engineering (Transactions of the CSAE), 2022, 38(1):203-211.
[22]	Yang G, Chen G, Li C, et al. Convolutional Rebalancing Network for the Classification of Large Imbalanced Rice Pest and Disease Datasets in the Field[J]. Frontiers in Plant Science, 2021, 12:1150.
[23]	Nanni L, Manfè A, Maguolo G, et al. High performing ensemble of convolutional neural networks for insect pest image detection[J]. Ecological Informatics, 2022, 67:101515. DOI URL

模型 Model	训练集准确率 Training set accuracy(%)	验证集准确率 Verification set accuracy(%)
EN	85.10	74.63
SN	52.33	54.19
MN	73.51	68.79
CN	91.19	74.91
VT	98.58	74.12
ST	84.64	73.27
平均准确率 Average accuracy	80.88	69.99

模型 Model	训练集准确率 Training set accuracy(%)	验证集准确率 Verification set accuracy(%)
EN	85.10	74.63
SN	52.33	54.19
MN	73.51	68.79
CN	91.19	74.91
VT	98.58	74.12
ST	84.64	73.27
平均准确率 Average accuracy	80.88	69.99

模型 Model	准确率 Accuracy(%)	F₁值 (%)
CN	74.41	67.84
EN	71.31	63.47
VT	62.35	51.83
ST	73.27	66.08

模型 Model	准确率 Accuracy(%)	F₁值 (%)
CN	74.41	67.84
EN	71.31	63.47
VT	62.35	51.83
ST	73.27	66.08

组合方式 Composition	相对多数投票法 Relative majority voting method		加权投票法Weighted voting method
	相对多数投票法 Relative majority voting method		(A)		(B)		(C)
	准确率 Accuracy(%)	F₁值 (%)	准确率 Accuracy(%)	F₁值 (%)	准确率 Accuracy(%)	F₁值 (%)	准确率 Accuracy(%)	F₁值 (%)
CN+ST+EN	75.16	68.37	75.09	68.53	-	-	75.37	68.64
CN+ST+VT	74.91	68.23	75.03	68.50	-	-	75.27	68.73
CN+EN+VT	74.00	66.81	74.73	68.04	74.33	67.16	74.73	68.04
EN+VT+ST	73.76	66.54	74.00	66.70	74.00	66.70	73.84	66.70
CN+EN+VT+ST	75.56	68.81	75.67	68.85	75.70	68.87	75.75	69.13