Pub Date : 2020-11-20DOI: 10.1109/ICCSNT50940.2020.9304983
Zhiyan Bao, Chen Xing, Xi Liang
To detect trespassing in images captured by drones for water conservancy facilities inspection, this paper proposes a method that adapts Hight-Resolution Net to reserve high resolution features for improving detecting results. To detect trespassing target with small scale, this method parallels low-resolution and high-resolution conventical feature maps to reserve high-resolution features, besides that multi-scale fusions are conducted to enhance feature maps with different resolutions. Compare to Faster R-CNN, proposed method achieves 1.7% higher mAP on small targets.
{"title":"Object Detection on Aerial Image by Using High-Resolutuion Network","authors":"Zhiyan Bao, Chen Xing, Xi Liang","doi":"10.1109/ICCSNT50940.2020.9304983","DOIUrl":"https://doi.org/10.1109/ICCSNT50940.2020.9304983","url":null,"abstract":"To detect trespassing in images captured by drones for water conservancy facilities inspection, this paper proposes a method that adapts Hight-Resolution Net to reserve high resolution features for improving detecting results. To detect trespassing target with small scale, this method parallels low-resolution and high-resolution conventical feature maps to reserve high-resolution features, besides that multi-scale fusions are conducted to enhance feature maps with different resolutions. Compare to Faster R-CNN, proposed method achieves 1.7% higher mAP on small targets.","PeriodicalId":6794,"journal":{"name":"2020 IEEE 8th International Conference on Computer Science and Network Technology (ICCSNT)","volume":"7 1","pages":"111-114"},"PeriodicalIF":0.0,"publicationDate":"2020-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"72977484","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-11-20DOI: 10.1109/ICCSNT50940.2020.9304996
Hongwei Kan, Rengang Li, Dongdong Su, Yanwei Wang, Yanmei Shen, Wei Liu
To solve the problem of computing overload in cloud, we intend to design a trusted edge cloud computing model and method based on FPGA (Field Programmable Gate Array) clusters. Firstly, a device, named FPGA Box, with PCIe (Peripheral Component Interconnect Express) power supply capability is used to manage the FPGA accelerator in the model. Besides, the FPGA cluster provide heterogeneous accelerated computing services for the data center through the network. Furthermore, we proposed a trusted edge cloud computing method based on FPGA cluster. On the one hand, a bi-level encryption algorithm based on RSA is proposed to generate an authorized use code, which implied the FPGA accelerator IP (Internet Protocol) address and other information. On the other hand, based on the programmable features of the FPGA accelerators, we set FPGA registers as use status bits, which can control different working status of accelerator. Specifically, when the accelerators has been assigned, we also need upload the deadline of usage to it. Finally, the software activity process of the entire trusted edge cloud system is described in detail, including the process of generating authorized use code. Simulation results show that the edge cloud computing mechanism based on FPGA cluster is proved to be trusted and effective.
{"title":"Trusted Edge Cloud Computing Mechanism Based on FPGA Cluster","authors":"Hongwei Kan, Rengang Li, Dongdong Su, Yanwei Wang, Yanmei Shen, Wei Liu","doi":"10.1109/ICCSNT50940.2020.9304996","DOIUrl":"https://doi.org/10.1109/ICCSNT50940.2020.9304996","url":null,"abstract":"To solve the problem of computing overload in cloud, we intend to design a trusted edge cloud computing model and method based on FPGA (Field Programmable Gate Array) clusters. Firstly, a device, named FPGA Box, with PCIe (Peripheral Component Interconnect Express) power supply capability is used to manage the FPGA accelerator in the model. Besides, the FPGA cluster provide heterogeneous accelerated computing services for the data center through the network. Furthermore, we proposed a trusted edge cloud computing method based on FPGA cluster. On the one hand, a bi-level encryption algorithm based on RSA is proposed to generate an authorized use code, which implied the FPGA accelerator IP (Internet Protocol) address and other information. On the other hand, based on the programmable features of the FPGA accelerators, we set FPGA registers as use status bits, which can control different working status of accelerator. Specifically, when the accelerators has been assigned, we also need upload the deadline of usage to it. Finally, the software activity process of the entire trusted edge cloud system is described in detail, including the process of generating authorized use code. Simulation results show that the edge cloud computing mechanism based on FPGA cluster is proved to be trusted and effective.","PeriodicalId":6794,"journal":{"name":"2020 IEEE 8th International Conference on Computer Science and Network Technology (ICCSNT)","volume":"65 1","pages":"146-149"},"PeriodicalIF":0.0,"publicationDate":"2020-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"75601752","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-11-20DOI: 10.1109/ICCSNT50940.2020.9304990
R-Y Wang, Tim Qinsong Guo, L. Li, Julia Yutian Jiao, Lena Yiqi Wang
In this research, a quantitative model is built to predict people's susceptibility to COVID-19 based on their genomes. Identifying people vulnerable to COVID-19 infections is crucial in stopping the spread of the virus. In previous studies, researchers have found that individuals with comorbid diseases have higher chances of being infected and developing more severe COVID-19 conditions. However, these patterns are only observed through correlational analyses between patient phenotypes and the severity of their COVID-19 infection. In this study, genetic variants underlying the observed comorbidity patterns are analyzed through machine learning of COVID-19 data from GWAS studies, which may reveal biological pathways underlying COVID-19 contraction that are essential to the development of effective and targeted therapeutics. Furthermore, through combining genetic variants with the individual's phenotypes, this study built a Neural Network model and Random Forest classifier to predict an individual's likelihood of COVID-19 infection. The Random Forest Classifier in this study shows that on-going symptoms are generally better predictors of COVID-19 condition (higher impurity-based feature importance) than diseases or medical histories. In addition, when trained with genomic data, the comorbid disease impact ranking deduced by the resulting RF model is highly consistent with phenotypic comorbidity patterns observed in past studies.
{"title":"Predictions of COVID-19 Infection Severity Based on Co-associations between the SNPs of Co-morbid Diseases and COVID-19 through Machine Learning of Genetic Data","authors":"R-Y Wang, Tim Qinsong Guo, L. Li, Julia Yutian Jiao, Lena Yiqi Wang","doi":"10.1109/ICCSNT50940.2020.9304990","DOIUrl":"https://doi.org/10.1109/ICCSNT50940.2020.9304990","url":null,"abstract":"In this research, a quantitative model is built to predict people's susceptibility to COVID-19 based on their genomes. Identifying people vulnerable to COVID-19 infections is crucial in stopping the spread of the virus. In previous studies, researchers have found that individuals with comorbid diseases have higher chances of being infected and developing more severe COVID-19 conditions. However, these patterns are only observed through correlational analyses between patient phenotypes and the severity of their COVID-19 infection. In this study, genetic variants underlying the observed comorbidity patterns are analyzed through machine learning of COVID-19 data from GWAS studies, which may reveal biological pathways underlying COVID-19 contraction that are essential to the development of effective and targeted therapeutics. Furthermore, through combining genetic variants with the individual's phenotypes, this study built a Neural Network model and Random Forest classifier to predict an individual's likelihood of COVID-19 infection. The Random Forest Classifier in this study shows that on-going symptoms are generally better predictors of COVID-19 condition (higher impurity-based feature importance) than diseases or medical histories. In addition, when trained with genomic data, the comorbid disease impact ranking deduced by the resulting RF model is highly consistent with phenotypic comorbidity patterns observed in past studies.","PeriodicalId":6794,"journal":{"name":"2020 IEEE 8th International Conference on Computer Science and Network Technology (ICCSNT)","volume":"15 3 1","pages":"92-96"},"PeriodicalIF":0.0,"publicationDate":"2020-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"77598006","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-11-20DOI: 10.1109/ICCSNT50940.2020.9305016
Zhixiong Zhang, Yun Wu
Since the complexity of artificial design features in model training, the toll data features cannot be used reasonably and efficiently. Toll prediction of one single station based on toll historical data would ignore the interaction between stations in the highway network. Therefore, this paper constructs a highway toll forecast model, named DBN-MSVR combining deep belief network and multi-task learning for multi-station toll forecasting. This model uses the optimized deep belief network to perform feature learning on toll data, and combines multi-task learning and support vector regression on the top layer of the deep belief network to predict tolls. Experiments show that the DBN-MSVR toll prediction model has higher prediction accuracy than traditional methods.
{"title":"Highway Toll Forecasting Model","authors":"Zhixiong Zhang, Yun Wu","doi":"10.1109/ICCSNT50940.2020.9305016","DOIUrl":"https://doi.org/10.1109/ICCSNT50940.2020.9305016","url":null,"abstract":"Since the complexity of artificial design features in model training, the toll data features cannot be used reasonably and efficiently. Toll prediction of one single station based on toll historical data would ignore the interaction between stations in the highway network. Therefore, this paper constructs a highway toll forecast model, named DBN-MSVR combining deep belief network and multi-task learning for multi-station toll forecasting. This model uses the optimized deep belief network to perform feature learning on toll data, and combines multi-task learning and support vector regression on the top layer of the deep belief network to predict tolls. Experiments show that the DBN-MSVR toll prediction model has higher prediction accuracy than traditional methods.","PeriodicalId":6794,"journal":{"name":"2020 IEEE 8th International Conference on Computer Science and Network Technology (ICCSNT)","volume":"7 1","pages":"81-85"},"PeriodicalIF":0.0,"publicationDate":"2020-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"82386626","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-11-20DOI: 10.1109/iccsnt50940.2020.9305007
Xianglin Meng, He Wang, Mu Lin, Yonghua Zhou
With the rapid development of urban rail transit and the improvement of machine learning technology, the application of deep reinforcement learning to train operation control has become a research hotspot. In this paper, the train operation control method based on deep reinforcement learning is established for urban rail transit. A subway line is employed to perform simulation, and the developed method is verified. The simulation results revealed the applicability and practicability of the method.
{"title":"Deep Reinforcement Learning for Energy-efficient Train Operation of Automatic Driving","authors":"Xianglin Meng, He Wang, Mu Lin, Yonghua Zhou","doi":"10.1109/iccsnt50940.2020.9305007","DOIUrl":"https://doi.org/10.1109/iccsnt50940.2020.9305007","url":null,"abstract":"With the rapid development of urban rail transit and the improvement of machine learning technology, the application of deep reinforcement learning to train operation control has become a research hotspot. In this paper, the train operation control method based on deep reinforcement learning is established for urban rail transit. A subway line is employed to perform simulation, and the developed method is verified. The simulation results revealed the applicability and practicability of the method.","PeriodicalId":6794,"journal":{"name":"2020 IEEE 8th International Conference on Computer Science and Network Technology (ICCSNT)","volume":"47 1","pages":"123-126"},"PeriodicalIF":0.0,"publicationDate":"2020-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"76236486","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-11-20DOI: 10.1109/iccsnt50940.2020.9304976
Xiaoqing Zhang, K. Jiang, Wenjie Hu, Yonghua Zhou
Compared with the pure electromagnetic suspension system, the hybrid suspension system has the characteristics of lower energy consumption, can indirectly increase the safety of the system, and reduce the construction difficulty and engineering cost. In this paper, an electromagnetic-permanent-magnet hybrid levitation model of a maglev train is a control object, and a constrained model predictive computer controller is utilized for the levitation control. The simulation results show that the constrained predictive controller can satisfy the multiple constraints, and real-time and anti-disturbance requirements in the suspension process for this kind of hybrid suspension system.
{"title":"Model Predictive Computer Control for the Hybrid Levitation System of a Maglev Train","authors":"Xiaoqing Zhang, K. Jiang, Wenjie Hu, Yonghua Zhou","doi":"10.1109/iccsnt50940.2020.9304976","DOIUrl":"https://doi.org/10.1109/iccsnt50940.2020.9304976","url":null,"abstract":"Compared with the pure electromagnetic suspension system, the hybrid suspension system has the characteristics of lower energy consumption, can indirectly increase the safety of the system, and reduce the construction difficulty and engineering cost. In this paper, an electromagnetic-permanent-magnet hybrid levitation model of a maglev train is a control object, and a constrained model predictive computer controller is utilized for the levitation control. The simulation results show that the constrained predictive controller can satisfy the multiple constraints, and real-time and anti-disturbance requirements in the suspension process for this kind of hybrid suspension system.","PeriodicalId":6794,"journal":{"name":"2020 IEEE 8th International Conference on Computer Science and Network Technology (ICCSNT)","volume":"74 1","pages":"186-189"},"PeriodicalIF":0.0,"publicationDate":"2020-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"86318206","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-11-20DOI: 10.1109/ICCSNT50940.2020.9305002
Chen Xing, Xi Liang, Pengliang Zhang
This paper proposes a two-phase solution for aerial inspecting. First phase focuses on removing images with no abnormal, modified YOLOv3 is used in this phase. Second phase focuses on target locating and identifying, modified SSD is applied in this phase. The experiment result shows the modified YOLOv3 get 2.2% higher accuracy than original design, and the miss rate of detecting images with no abnormal is only 2.6%.
{"title":"A Two-Phase Object Detection Solution for Aerial Images","authors":"Chen Xing, Xi Liang, Pengliang Zhang","doi":"10.1109/ICCSNT50940.2020.9305002","DOIUrl":"https://doi.org/10.1109/ICCSNT50940.2020.9305002","url":null,"abstract":"This paper proposes a two-phase solution for aerial inspecting. First phase focuses on removing images with no abnormal, modified YOLOv3 is used in this phase. Second phase focuses on target locating and identifying, modified SSD is applied in this phase. The experiment result shows the modified YOLOv3 get 2.2% higher accuracy than original design, and the miss rate of detecting images with no abnormal is only 2.6%.","PeriodicalId":6794,"journal":{"name":"2020 IEEE 8th International Conference on Computer Science and Network Technology (ICCSNT)","volume":"163 1","pages":"119-122"},"PeriodicalIF":0.0,"publicationDate":"2020-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"79818531","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-11-20DOI: 10.1109/ICCSNT50940.2020.9305017
Dan Lv, Zhaohui Gao, Dejun Mu, Y. Zhong, Chengfan Gu
A new adaptive random weighted filtering algorithm is proposed. It is based on online estimation of limited memory measurement noise to overcome the problem of low filtering precision caused by arithmetic average estimation of measurement noise and its covariance matrix in the existing Kalman filtering algorithm of limited memory online estimation of measurement noise. This method establishes the stochastic weighting theory to estimate the measurement noise online and its covariance by adaptive adj usting the weights of measurement noise statistics. The weight of measurement noise statistics is used to suppress the influence of measurement noise on state estimation and improve the accuracy of filter estimation. Through simulations and analysis, the superiority of the proposed adaptive random weighted filtering algorithm based on online estimation of limited memory measurement noise algorithm is proved.
{"title":"Limited Memory Measurement Noise Adaptive Random Weighted Filtering Algorithm","authors":"Dan Lv, Zhaohui Gao, Dejun Mu, Y. Zhong, Chengfan Gu","doi":"10.1109/ICCSNT50940.2020.9305017","DOIUrl":"https://doi.org/10.1109/ICCSNT50940.2020.9305017","url":null,"abstract":"A new adaptive random weighted filtering algorithm is proposed. It is based on online estimation of limited memory measurement noise to overcome the problem of low filtering precision caused by arithmetic average estimation of measurement noise and its covariance matrix in the existing Kalman filtering algorithm of limited memory online estimation of measurement noise. This method establishes the stochastic weighting theory to estimate the measurement noise online and its covariance by adaptive adj usting the weights of measurement noise statistics. The weight of measurement noise statistics is used to suppress the influence of measurement noise on state estimation and improve the accuracy of filter estimation. Through simulations and analysis, the superiority of the proposed adaptive random weighted filtering algorithm based on online estimation of limited memory measurement noise algorithm is proved.","PeriodicalId":6794,"journal":{"name":"2020 IEEE 8th International Conference on Computer Science and Network Technology (ICCSNT)","volume":"21 1","pages":"127-132"},"PeriodicalIF":0.0,"publicationDate":"2020-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"91078559","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-11-20DOI: 10.1109/ICCSNT50940.2020.9304979
Chen Xing, Xi Liang, Rongjie Yang
The targets in aerial images captured by drones are difficult to detect due to their small size, those neural networks with better detecting accuracy are too complicated to run real-time job on drone-mounted computer. This paper proposes a network combined residual network and YOLOv3-Tiny, residual network is used to merge different level features for improving YOLOv3-Tiny's small object detecting performance. During the experiment, the proposed network gets 2.9 higher mAP than YOLOv3-Tiny.
{"title":"Compact One-Stage Object Detection Network","authors":"Chen Xing, Xi Liang, Rongjie Yang","doi":"10.1109/ICCSNT50940.2020.9304979","DOIUrl":"https://doi.org/10.1109/ICCSNT50940.2020.9304979","url":null,"abstract":"The targets in aerial images captured by drones are difficult to detect due to their small size, those neural networks with better detecting accuracy are too complicated to run real-time job on drone-mounted computer. This paper proposes a network combined residual network and YOLOv3-Tiny, residual network is used to merge different level features for improving YOLOv3-Tiny's small object detecting performance. During the experiment, the proposed network gets 2.9 higher mAP than YOLOv3-Tiny.","PeriodicalId":6794,"journal":{"name":"2020 IEEE 8th International Conference on Computer Science and Network Technology (ICCSNT)","volume":"64 6","pages":"115-118"},"PeriodicalIF":0.0,"publicationDate":"2020-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"91406591","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2020-11-20DOI: 10.1109/ICCSNT50940.2020.9305010
Xu Liang, Zhen Du
In order to solve the limitation of traditional genetic algorithm to solve the job shop scheduling problem, combined with the advantages of genetic algorithm (GA) and simulated annealing algorithm (SA), this paper proposes a kind of algorithm based on NSGA-II, which inserts simulated annealing algorithm during operation. A hybrid genetic algorithm simulated annealing algorithm (GASA) combining the advantages of the two algorithms is generated. The algorithm not only has the advantages of fast convergence speed of genetic algorithm and wide search area of simulated annealing algorithm, but also overcomes the problem of premature convergence of the former and slow convergence speed of the latter. In the operation details of the algorithm, adaptive function, non-dominated sorting and elite retention strategy are added to effectively improve the effectiveness of job shop scheduling.
{"title":"Genetic Algorithm with Simulated Annealing for Resolving Job Shop Scheduling Problem","authors":"Xu Liang, Zhen Du","doi":"10.1109/ICCSNT50940.2020.9305010","DOIUrl":"https://doi.org/10.1109/ICCSNT50940.2020.9305010","url":null,"abstract":"In order to solve the limitation of traditional genetic algorithm to solve the job shop scheduling problem, combined with the advantages of genetic algorithm (GA) and simulated annealing algorithm (SA), this paper proposes a kind of algorithm based on NSGA-II, which inserts simulated annealing algorithm during operation. A hybrid genetic algorithm simulated annealing algorithm (GASA) combining the advantages of the two algorithms is generated. The algorithm not only has the advantages of fast convergence speed of genetic algorithm and wide search area of simulated annealing algorithm, but also overcomes the problem of premature convergence of the former and slow convergence speed of the latter. In the operation details of the algorithm, adaptive function, non-dominated sorting and elite retention strategy are added to effectively improve the effectiveness of job shop scheduling.","PeriodicalId":6794,"journal":{"name":"2020 IEEE 8th International Conference on Computer Science and Network Technology (ICCSNT)","volume":"12 1","pages":"64-68"},"PeriodicalIF":0.0,"publicationDate":"2020-11-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"81830840","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：