Detecting Pests From Light-Trapping Images Based on Improved YOLOv3 Model and Instance Augmentation
文献类型： 外文期刊
作者： Lv, Jiawei 1 ; Li, Wenyong 1 ; Fan, Mingyuan 3 ; Zheng, Tengfei 1 ; Yang, Zhankui 1 ; Chen, Yaocong 3 ; He, Guohuang 3 ; Yang, Xinting 1 ; Liu, Shuangyin 3 ; Sun, Chuanheng 1 ;
作者机构： 1.Natl Engn Res Ctr Informat Technol Agr, Beijing, Peoples R China
2.Natl Engn Lab Agriprod Qual Traceabil, Beijing, Peoples R China
3.Zhongkai Univ Agr & Engn, Coll Informat Sci & Technol, Guangzhou, Peoples R China
关键词： pests and diseases; pest detection; YOLOv3; image cropping; convolutional neural network
期刊名称：FRONTIERS IN PLANT SCIENCE （ 影响因子：6.627； 五年影响因子：7.255 ）
ISSN： 1664-462X
年卷期： 2022 年 13 卷
页码：
收录情况： SCI
摘要： Light traps have been widely used as effective tools to monitor multiple agricultural and forest insect pests simultaneously. However, the current detection methods of pests from light trapping images have several limitations, such as exhibiting extremely imbalanced class distribution, occlusion among multiple pest targets, and inter-species similarity. To address the problems, this study proposes an improved YOLOv3 model in combination with image enhancement to better detect crop pests in real agricultural environments. First, a dataset containing nine common maize pests is constructed after an image augmentation based on image cropping. Then, a linear transformation method is proposed to optimize the anchors generated by the k-means clustering algorithm, which can improve the matching accuracy between anchors and ground truths. In addition, two residual units are added to the second residual block of the original YOLOv3 network to obtain more information about the location of the underlying small targets, and one ResNet unit is used in the feature pyramid network structure to replace two DBL(Conv+BN+LeakyReLU) structures to enhance the reuse of pest features. Experiment results show that the mAP and mRecall of our proposed method are improved by 6.3% and 4.61%, respectively, compared with the original YOLOv3. The proposed method outperforms other state-of-the-art methods (SSD, Faster-rcnn, and YOLOv4), indicating that the proposed method achieves the best detection performance, which can provide an effective model for the realization of intelligent monitoring of maize pests.
- 相关文献
作者其他论文 更多>>
-
GCVC: Graph Convolution Vector Distribution Calibration for Fish Group Activity Recognition
作者：Zhao, Zhenxi;Zhao, Chunjiang;Zhao, Zhenxi;Yang, Xinting;Zhou, Chao;Zhao, Chunjiang;Zhao, Zhenxi;Yang, Xinting;Zhou, Chao;Zhao, Chunjiang;Zhao, Zhenxi;Yang, Xinting;Zhou, Chao;Zhao, Chunjiang;Liu, Jintao
关键词：Fish; Feature extraction; Activity recognition; Calibration; Adhesives; Training; Convolution; Graph convolution vector calibration; fish group activity; activity feature vector calibration; fish activity dataset
-
Porphyrin fluorescence imaging for real-time monitoring and visualization of the freshness of beef stored at different temperatures
作者：Liu, Huan;Zhu, Lei;Ji, Zengtao;Zhang, Min;Yang, Xinting;Liu, Huan;Zhu, Lei;Ji, Zengtao;Yang, Xinting;Zhang, Min;Liu, Huan;Ji, Zengtao;Yang, Xinting;Liu, Huan;Ji, Zengtao;Yang, Xinting
关键词：Porphyrin; Fluorescence imaging; Beef; Freshness; Visualization
-
FCFormer: fish density estimation and counting in recirculating aquaculture system
作者：Zhu, Kaijie;Ma, Pingchuan;Zhu, Kaijie;Yang, Xinting;Yang, Caiwei;Fu, Tingting;Ma, Pingchuan;Hu, Weichen;Zhu, Kaijie;Yang, Xinting;Yang, Caiwei;Fu, Tingting;Ma, Pingchuan;Hu, Weichen;Zhu, Kaijie;Yang, Xinting;Yang, Caiwei;Fu, Tingting;Ma, Pingchuan;Hu, Weichen
关键词：recirculating aquaculture systems; density estimation; fish counting; transformer; deep learning
-
Synchronous fluorescence detection of nitrite in meat products based on dual-emitting dye@MOF and its portable hydrogel test kit
作者：Deng, Siyang;Liu, Junmei;Han, Dong;Zhang, Chunhui;Yang, Xinting;Liu, Huan;Yang, Xinting;Liu, Huan;Deng, Siyang;Liu, Junmei;Blecker, Christophe
关键词：Luminescent metal organic framework; Dual-mode detection; On-spot visualization; NO2-; Rh6G@UIO-66-NH2
-
Automatic acquisition, analysis and wilting measurement of cotton 3D phenotype based on point cloud
作者：Hao, Haoyuan;Zhuang, Lvhan;Xu, Longqin;Li, Hongxin;Liu, Shuangyin;Hao, Haoyuan;Wu, Sheng;Li, Yuankun;Wen, Weiliang;Zhuang, Lvhan;Guo, Xinyu;Hao, Haoyuan;Wu, Sheng;Li, Yuankun;Wen, Weiliang;Zhuang, Lvhan;Guo, Xinyu;Hao, Haoyuan;Zhuang, Lvhan;Xu, Longqin;Li, Hongxin;Liu, Shuangyin;Li, Yuankun;Zhang, Yongjiang
关键词：Phenotypic analysis; Deep learning; Leaf wilting; Multi-view
-
Maize emergence rate and leaf emergence speed estimation via image detection under field rail-based phenotyping platform
作者：Zhuang, Lvhan;Hao, Haoyuan;Li, Jinhui;Xu, Longqin;Liu, Shuangyin;Zhuang, Lvhan;Wang, Chuanyu;Hao, Haoyuan;Guo, Xinyu;Zhuang, Lvhan;Wang, Chuanyu;Hao, Haoyuan;Guo, Xinyu;Zhuang, Lvhan;Hao, Haoyuan;Li, Jinhui;Xu, Longqin;Liu, Shuangyin;Zhuang, Lvhan;Hao, Haoyuan;Li, Jinhui;Xu, Longqin;Liu, Shuangyin
关键词：Field rail-based phenotyping platform; Emergence rate; Leaf emergence speed; Faster R-CNN; Mask R-CNN
-
Principles and applications of convolutional neural network for spectral analysis in food quality evaluation: A review
作者：Luo, Na;Xu, Daming;Xing, Bin;Yang, Xinting;Sun, Chuanheng;Luo, Na;Xu, Daming;Xing, Bin;Yang, Xinting;Sun, Chuanheng;Luo, Na;Xu, Daming;Xing, Bin;Yang, Xinting;Sun, Chuanheng;Sun, Chuanheng
关键词：Convolutional neural network; Spectroscopic technologies; Evaluation; Food quality
