Platform Construction Of Water Resources And Data Processing Based On Big Data

With the wide application of information technologies,such as internet of things,computer technology and network technology,the remote monitoring and real-time acquisition of water resources data also have gained further development.In the monitoring process,date variety and quantity are increasing,which will bring the question of solving the proper storage and efficiently analysis of water resources data.Using big data technologies forecast and analyse the water resources information,so as to make decisions for the rational use of water resources and prevent flood disaster.It is an important topic for the study of water resources big data.based on the national and international scientific and technological cooperation special project of Integrated supervision technology of multi-water sources in grassland-type watershed of the Internet of Things.Xilin River basin are chosen as the research area.The monitoring data of this area collected by the water information monitoring equipment used for the study of distributed storage and data prediction.On the basis of deep study of data collected by sensors and big data technologies,big data technologies and water resources data processing are combined to construct the data storage and analysis platform of water resources big data based on Hadoop,according to the collected data,the rainfall level and groundwater level are predicted based on the platform.With the quantity increasing of monitoring data,the platform use the distribution database HBase in order to store the data safely and stably.The monitoring data of research area are transmitted by the sensors to the matching data server.For preventing the data missing and abnormal data,the improved Kalman filtering algorithm was used to solve the problem.The accuracy of the data is ensured.Different prediction methods were used for the data of different monitoring cycles.According to the characteristics of rainfall and groundwater,different prediction methods were employed to predict.For the short monitoring cycle and the quick increasing data in the future,such as rainfall,the simple Naive Bayes algorithm based on Hadoop was used to predict rainfall level in case of massive data in the future.For the long monitoring cycle and the less amount of data accumulation,such groundwater,the AR algorithm was used to predict the groundwater level.The combination of distributed and traditional methods can improve the efficiency of data processing.The big data storage and analysis platform for water resources in Xilin River basin was constructed to collect and analyze water resources data,so as to provide the necessary decisions support for the rational use of water resources and flood prevention.