The cross-domain data security of the visual Internet involves various links such as data transmission, storage, processing and destruction. This paper constructs a cross-domain sharing system composed of a basic resource layer, a shared resource layer and an application layer. The original measured data is converted into a common shared data format through sensors, and then converted into a digital quantity that can be processed by computers using an A/D converter. The error is corrected in the preprocessing stage. The key feature data is extracted, the data from different sensors are integrated, and the results are output in the required format. Privacy protection is achieved by calculating keyword weights, etc. In the experimental analysis, the scheduling time, resource generation time and search time of the method in this paper are the shortest, which are 303 ms, 1086 ms and 110 ms respectively; the number of shared resources calculated is the largest, whi1ch is 146, and the time for returning the calculation results is the shortest, which is 97 ms; when the number of system calculations reaches 100, the highest point of the calculation cost is the lowest point of all system calculation costs, which is 140 ms. Finally, it is concluded that the scheme in this paper has high privacy protection performance, can keep resource expenditure controllable, and has practicality.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Cross-Domain Sharing and Privacy Protection Method of Fused Multi-source Data in Visual Internet of Things

  • Longjie Zhu,
  • Xuming Fang,
  • Xinlei Yang,
  • Ming Li

摘要

The cross-domain data security of the visual Internet involves various links such as data transmission, storage, processing and destruction. This paper constructs a cross-domain sharing system composed of a basic resource layer, a shared resource layer and an application layer. The original measured data is converted into a common shared data format through sensors, and then converted into a digital quantity that can be processed by computers using an A/D converter. The error is corrected in the preprocessing stage. The key feature data is extracted, the data from different sensors are integrated, and the results are output in the required format. Privacy protection is achieved by calculating keyword weights, etc. In the experimental analysis, the scheduling time, resource generation time and search time of the method in this paper are the shortest, which are 303 ms, 1086 ms and 110 ms respectively; the number of shared resources calculated is the largest, whi1ch is 146, and the time for returning the calculation results is the shortest, which is 97 ms; when the number of system calculations reaches 100, the highest point of the calculation cost is the lowest point of all system calculation costs, which is 140 ms. Finally, it is concluded that the scheme in this paper has high privacy protection performance, can keep resource expenditure controllable, and has practicality.