| The repaid development of IoT has an important influence on realizing intensive agriculture, high yield and high quality, and it will provide a solid foundation for the development of agriculture information technologies. In this thesis, agriculture IoT-oriented multi-environment information fusion model for monitoring and recognition is established, providing a new way to access information for the farm.Sensor nodes are laid on targeted farmland areas and collected environmental signals are treated as a source of information. In this thesis, data mining and heterogeneous information fusion algorithms are studied, and information extraction and integration module and system platform module are designed. The system mainly includes three modules as follows:First module is called heterogeneous environmental information extraction and integration module, or heterogeneous database system information extraction and integration module. Steps are as follows:First, according to the information provided by the mapping file, extract all the environmental information from the source database system, and transfer the data into XML document. Second, according to the data table field names and field types in the destination database system, transfer the XML document into environmental information, matching the destination data table field names and field types, stored in the destination database system. This module consists of four sub-modules:heterogeneous database system field information extraction module, mapping file generation module, data extraction module and data conversion module. Realize data conversion between heterogeneous database systems, between different data table field names and field types.Second module is called multi-environment information fusion for monitoring and recognition module. Steps are as follows:First, association rules between environmental information and monitoring results are calculated according to the Apriori algorithm. Second, according to the fuzzy inference algorithm, real-time environmental information as input and association rules as rules, monitoring results are calculated. Finally, develop control strategies according to those monitoring results. Experimental results show that the monitoring and recognition rate is about eight percent, with good agricultural monitoring value and broad application prospects.Last module is called multi-environment information fusion for monitoring and recognition system platform module. Java and Matlab are used to realize this system platform. This system platform consists of four sub-modules:heterogeneous database system information extraction and integration module, multi-environment information fusion for monitoring and recognition module, environmental information condition display module and environmental information statistic display module. Since the first two are introduced before, system platform process and overall framework are focused in this module chapter. At the same time, how to realize environmental information condition display module and environmental information statistic display module, and how to use the interface are introduced.Conclusions and outlooks are made in the end. |
