关键词:
classification;attention module;infection severity;color space
摘要:
In this paper, a lightweight convolutional neural network model is proposed to diagnose the disease severity of tomato infection. Different regions of tomato leaf image had obvious threshold differences in Lab color space, and the grading label of disease infection degree of each tomato leaf image was obtained. At the same time, in order to solve the problems of low efficiency and general recognition accuracy of artificial recognition of tomato leaf diseases, and unable to accurately judge the tomato disease grade, this paper proposed a new method based on lightweight convolutional neural network, which selected ShuffleNet V2 as the backbone and applied Attention mechanisms that coordinate channel and spatial bidirectional perception. The results of a large number of cross-validation experiments showed that the accuracy of the network structure in classifying the severity of four common tomato leaf diseases and one healthy leaf infection was 91.817%, and the average accuracy was 85.496%.
作者:
Yang Xinyan;Zhang Jijuan;Wang Yufan;Zhang Zhongfeng
作者机构:
[Zhang Jijuan; Wang Yufan; Zhang Zhongfeng; Yang Xinyan] Cent South Univ Forestry Sci & Technol, Coll Furniture & Art Design, Changsha 410004, Hunan, Peoples R China.;[Zhang Jijuan; Wang Yufan; Zhang Zhongfeng; Yang Xinyan] Natl Forestry & Grassland Adm, Green Furniture Engn Technol Res Ctr, Changsha 410004, Hunan, Peoples R China.;[Zhang Jijuan; Wang Yufan; Zhang Zhongfeng; Yang Xinyan] Hunan Green Home Engn Technol Res Ctr, Changsha 410004, Hunan, Peoples R China.
会议名称:
International Conference on Mechanical Design and Simulation (MDS)
会议时间:
MAR 18-20, 2022
会议地点:
Wuhan, PEOPLES R CHINA
会议主办单位:
[Yang Xinyan;Zhang Jijuan;Wang Yufan;Zhang Zhongfeng] Cent South Univ Forestry Sci & Technol, Coll Furniture & Art Design, Changsha 410004, Hunan, Peoples R China.^[Yang Xinyan;Zhang Jijuan;Wang Yufan;Zhang Zhongfeng] Natl Forestry & Grassland Adm, Green Furniture Engn Technol Res Ctr, Changsha 410004, Hunan, Peoples R China.^[Yang Xinyan;Zhang Jijuan;Wang Yufan;Zhang Zhongfeng] Hunan Green Home Engn Technol Res Ctr, Changsha 410004, Hunan, Peoples R China.
摘要:
This paper proposes a multi-situational FBS and Quality Function Deployment (QFD) integrated product design requirements model for more accurate product positioning, longer product life and improved product design rationality.A multi-situational FBS-QFD model is constructed to design products by analysing the relationship between the needs of realistic and ambiguous user groups and product functionality.Specific steps: clarify the design objectives, analyse the scenario-structure units of the design objectives, and analyse the mapping relationship between scenario-function-behaviour-structure after building the FBS model.Predicting scenario feasibility, identifying user requirements and translating them into design features.Determine design feature scores and their positive and negative correlations, construct a matrix of user requirements and design feature relationships, derive key design innovation areas and functions, and complete product design solutions.The validation of the model is carried out using rehabilitation products as an example The study shows that the model is scientific and can improve the quality of product design.Provides a theoretical basis and practical strategies for product design decisions
作者:
Songlin Wu;Hanqing Wang;Chengjun Li;Qiuxin Liu
期刊:
E3S Web of Conferences,2022年356:05012-null ISSN:2267-1242
通讯作者:
Wu, S.
作者机构:
[Wang H.; Li C.; Wu S.] Central South University of Forestry and Technology, Hunan, Changsha, 410082, China;[Liu Q.] Wuhan City College, Hubei, Wuhan, 430075, China
通讯机构:
[Wu, S.] C;Central South University of Forestry and Technology, Hunan, China
作者机构:
[Zuo, Xiaolon; Zhou, Xuepeng; Deng, Zhenhua; Chen, Tao; Luo, Jin] Cent South Univ, Sch Automat, Changsha 410075, Peoples R China.;[Liu, Yong-min] Cent South Univ Forestry & Technol, Coll Comp & Informat Engn, Changsha 410004, Peoples R China.;[Liu, Yong-min] Cent South Univ Forestry & Technol, Res Ctr Smart Forestry Cloud, Changsha 410004, Peoples R China.
会议名称:
41st Chinese Control Conference (CCC)
会议时间:
JUL 25-27, 2022
会议地点:
Hefei, PEOPLES R CHINA
会议主办单位:
[Chen, Tao;Luo, Jin;Deng, Zhenhua;Zuo, Xiaolon;Zhou, Xuepeng] Cent South Univ, Sch Automat, Changsha 410075, Peoples R China.^[Liu, Yong-min] Cent South Univ Forestry & Technol, Coll Comp & Informat Engn, Changsha 410004, Peoples R China.^[Liu, Yong-min] Cent South Univ Forestry & Technol, Res Ctr Smart Forestry Cloud, Changsha 410004, Peoples R China.
会议论文集名称:
Chinese Control Conference
关键词:
Resource allocation;inequality constraints;second-order multi-agent systems;distributed algorithms;cyberphysical systems
摘要:
In this paper, we investigate distributed resource allocation problems of second-order multi-agent systems, where the decisions of agents are subjected to inequality network resource constraints. In contrast to well-known resource allocation problems, the second-order dynamics of agents and the coupling inequality constraints are considered in our problem at the same time. In order to optimally allocate the network resource, a distributed algorithm is developed via state feedback and gradient descent. Moreover, the convergence of the algorithm is analyzed with the help of convex analysis and Lyapunov stability theory. By the algorithm, the second-order agents globally asymptotically converge to the optimal solution. Finally, the effectiveness of our method is verified by the numerical example.
作者机构:
[Zhou, Mengdong; Liu, Shuai; Li, Jianjun] Cent South Univ Forestry & Technol, Changsha 410004, Peoples R China.
会议名称:
31st International Conference on Artificial Neural Networks (ICANN)
会议时间:
SEP 06-09, 2022
会议地点:
Univ W England, Bristol, ENGLAND
会议主办单位:
Univ W England
会议论文集名称:
Lecture Notes in Computer Science
关键词:
Fire detection;YOLOv5;Cosine annealing
摘要:
Forest fires have a very bad impact on the natural environment and human beings. To protect the environment and enhance human safety, it is important to detect the source of a fire before it spreads. The existing fire detection algorithms have a weak generalization and do not fully consider the influence of fire target size on detection. To enhance the ability of fire detection of different sizes, ground fire data and Unmanned Aerial Vehicle (UAV) forest fire data are combined in this paper. To improve the detection accuracy of the model, a cosine annealing algorithm, label smoothing, and multi-scale training are introduced. The experimental results show that the Improved-YOLOv5s model proposed in this paper has strong generalization and a good detection effect for different sizes of fires. The mean Average Precision (mAP) value reaches 88.7%, 8% higher than that of YOLOv5s mAP. The proposed model has the advantages of strong generalization and high precision.
摘要:
In this paper, a regression analysis method of compressive strength curve of recycled concrete is proposed. Firstly, the compressive strength of recycled aggregate concrete at different ages is measured through laboratory tests. Then, based on the regression analysis method, the cubic function is used to simulate and analyze the compressive strength change of recycled aggregate concrete, and the age and comprehensive strength development law of recycled aggregate concrete are obtained.
期刊:
E3S Web of Conferences,2021年269:02002-null ISSN:2267-1242
通讯作者:
Chen, Z.
作者机构:
[Tang R.] Central South University of Forestry and Technology, Changsha, 410000, China;[Chen Z.; Ou Y.; Chen X.; Xu Y.] Hunan Institute of Water Resource and Hydropower Research, Changsha, 410007, China
通讯机构:
[Chen, Z.] H;Hunan Institute of Water Resource and Hydropower ResearchChina
会议名称:
2021 International Conference on Environmental Engineering, Agricultural Pollution and Hydraulical Studies, EEAPHS 2021
会议时间:
29 May 2021 through 30 May 2021
摘要:
In this study, an integrated ecological system was constructed to treat small scattered aquaculture wastewater in southern rural areas of China. The water outlet of 4 level wetlands was continuously monitored from July to December in 2017. Results showed the average concentrations of total nitrogen (TN), ammonia nitrogen (NH<sub>4<sup>+<sup/><sub/>-N), nitrate nitrogen (NO<sub>3<sub/>-N), total phosphorus (TP) and chemical oxygen demand (COD) were 43.64mg/L, 17.53mg/L, 1.71mg/L, 1.66mg/L and 51.39mg/L in the average effluent concentration of grade I wetland, respectively, and 8.35mg/L, 4.42mg/L, 0.24mg/L, 0.26mg/L, 21.32mg/L in the average effluent concentration of grade IV wetland, respectively. The removal rates were 81%, 75%, 86%, 85% and 59% for TN, NH<sub>4<sup>+<sup/><sub/>-N, NO<sub>3<sub/>-N, TP and COD in the integrated ecological system, respectively. The effluents from the integrated ecological system met the requirements of “Discharge Standard of Pollutants for Livestock and Poultry Breeding” (GB 18596-2001) and achieved “Discharge Standard of Pollutants for Municipal Wastewater Treatment Plant” (GB 18918-2002) the center two levels to discharge the standard. Obviously, the integrated ecological system could work efficiently in treating the rural scattered aquaculture wastewater, and also possess merits of low construction and operation costs and simple management method, which will be benefited to its application in the southern rural regions of China.
摘要:
In order to get the best mix proportion of recycled concrete with super plasticizer, an experimental research method of mix proportion of recycled concrete with super plasticizer is designed. The whole experimental scheme and steps are described in detail. By adding a kind of additive, i.e. super plasticizer, into recycled concrete of different strength grades (prepared with waste concrete and mixed with part of natural concrete), and testing the performance of recycled concrete in laboratory, the best mix proportion of recycled concrete with super plasticizer is obtained, which lays a foundation for realizing the best parameters of high performance of recycled concrete.
摘要:
Accurate measurement of subcutaneous adipose tissue (SAT) and visceral adipose tissue (VAT) in the thorax is important for understanding the impact of body composition upon various clinical disorders. The aim of this paper is to explore a practical system for the automatic localization of the axial slices through the thorax at the T7 and T8 vertebral levels in computed tomography (CT), and automatic segmentation of VAT in T7 slice and SAT at T8 slice via deep learning (DL). The methodology mainly consists of two models: the localization model based on AlexNet and the segmentation model based on UNet. For the first one, two slices (T7 and T8) at the middle of the seventh and eighth thoracic vertebrae, respectively, from the full or partial body scan of each patient are automatically detected. For the second one, all the CT images and the associated adipose tissue ground truth segmentations are used for training, where just T7 and T8 slices are tested by the two-label Unet. The datasets from four universities (Penn, Duke, Columbia, and Iowa) are used for training and validation of the models. In the experiments, relevant statistical parameters including Mean Distance (MD), Standard Deviation (SD), True Positive Rate (TPR), and True Negative Rate (TNR) indicate that the proposed algorithm has high reliability and may be useful for fully automated body composition analysis with high accuracy.