| China is one of the countries with serious harmful algal blooms (HAB) disasters in the world. Frequent HAB disasters caused severe damage to the ecological environment and social economy. Efficient marine online monitoring and HAB disaster predicting system is needed to be designed urgently. In this paper, based on the advanced ocean monitoring technology, embedded system, XML, network communication, data warehouse support, dynamic web technologies are used to design an integrated system on marine online monitoring. Meanwhile, cluster analysis of data mining algorithms is introduced to do preliminary study on HAB warning model.In this paper the whole framework of marine online monitoring system is designed based on the need and the technique developing trend of HAB disaster predicting. Under this framework, we design the marine element online monitoring subsystem and marine information online managing subsystem. Marine element online monitoring subsystem is built on the platform of Linux embedded system. It solves the problem of transform medium between buoys and stations by distribution control method with GPRS wireless communication technology. Meanwhile, the Linux platform and Windows platform are seamlessly connected based on the XML data transformation so that the marine monitoring data can be uploaded in real time. The monitoring Network is consisted of sensors, buoys and stations to make it a three-level network c onstruction, which has a good ability of expansibility. M arine information online managing subsystem is built based on the data warehouse with the technologies of Microsoft IIS, Apache Tomcat Server and ASP.NET, JSP dynamic Website Browsing. The maintenance and update of marine data in data warehouse can be easily implemented through long-distance alternation between client and data warehouse on Internet. It makes marine data online publishing and management very practical and effective. The marine data dynamic curves publishing system is developed for research on regulation of marine data variety and for the convenience of marine element monitoring. The system can monitor the rules of marine data changing .The dynamic curves can be browsed on line in real time.Finally, the paper discussed the application of data mining algorithm in HAB disaster predicting based on cluster analysis. Lots of experiments are made based on the deep research with traditional FCM cluster algorithm and the rule of HAB disaster. An effective pretreatment weighted FCM algorithm named PW-FCM is proposed based on the traditional FCM cluster analysis algorithm. The new way shows more excellent performance compared with traditional FCM cluster algorithm. |
PDF Full Text Request