Abstract
This paper describes a computing backend for a water pipe monitoring system. Today, most such systems are divided into event-triggered and continuous monitoring, but they all lack systematic handling of data. Many systems simply store data in files with specific naming conventions and ad hoc formats, making them difficult to retrieve, maintain, disseminate, and analyze. To address these problems, our backend supports data management and dissemination. Unlike previous systems that store data in files or conventional databases before analysis, our modular architecture not only saves data in efficiently searchable ways by indexing as a baseline dataset but also detected events in discrete time manner and other processed data. To facilitate analysis, we design a plug-in structure to allow processing modules to perform inline processing and shorten detection time. For data dissemination, our architecture can compose multiple visualizations including geographical maps to create powerful tools to yield new insight into massive datasets. The backend system enables Internet web service for visualization, data management, and remote sensor control for better integration. Our system is applicable to not only water pipelines but also bridges and civil structures in general. Our proposed backend system has been implemented and validated through field deployment. One such system has been running for over 1.5 years and has collected millions of records to date. A Google Map integrated visualization service has been developed to demonstrate lively collected records in real-time. This is expected to be more helpful for better understanding of civil structuresa' behavior in the long term. © 2014 SPIE.