Automatic Matching Method for Ocean Observation Elements Based on Edge Computing
Zhijin Qiu, Tong Hu, Suiping Qi, Shiyong Liu, Xupeng Wang, Zhongwen Guo · 2018
In order to realize comprehensive observation and research on the ocean, large-scale data integration needs to be carried out on the existing heterogeneous ocean observing systems. However, the content and format of ocean observation data sources are extremely complicated, which causes problems such as large amount of calculation, long time for matching and high error rate in the process of ocean observation element conversion. By using edge computing and improved KNN algorithm, we propose a new automatic matching method for ocean observation elements based on edge computing. According to the characteristics and advantages of the local, edge and cloud, the different processes of matching observation elements are executed. In order to solve the problem of large amount of computation and occupying much space, KNN algorithm is improved by grouping fast search (G-KNN) and weighted Euclidean distance (ω-KNN). The prototype system was established by using the actual ocean observation data set to verify the method. The results show that this method can improve the execution efficiency by 2.67 times and the accuracy rate of 96%, which can further improve the efficiency while ensuring the accuracy of ocean observation element matching.