Article Preview
TopIntroduction
Natural disasters like floods, landslides and fires require an increasing activity by civil protection agencies involved in territory monitoring. Territory monitoring is needed also in industrial and in intensive agriculture areas in order to ensure protection against the occurrence of water pollution and disposal of waste without a permit. The Advanced Monitoring System of the Territory (SMAT) is a distributed system for the monitoring of the territory that performs territory surveillance using a fleet of different Unmanned Aircraft Vehicles (UAVs). Each UAV, equipped with proper payload sensors, downloads to its ground control station (CS) video and images of the target territory. In turn, each ground control station communicates to the central Supervision and Coordination Station (SS&C) the received data in near-real-time. The SS&C performs data storage and near-real-time data fusion.
The software component described in this paper is one of the geo-spatial services provided by SS&C and is called Metadata Retrieval (MDR). MDR has been developed to enrich and enhance the information collected by UAVs’ missions and the information provided by authoritative geospatial data with data coming from Volunteered Geographical Information projects.
The presence of abundant structured and social data on the Web is providing new opportunities of searching for information. The ability of Web 2.0 applications to solicit feedback from crowds enables a new class of applications on the Web, which consist in methods for enhancing data and solving tasks by soliciting the contribution from the crowds. The resulting data are later aggregated, analyzed and used to solve some complex task, or they are published in a structured and semantic form on the Web performing the so-called mashup applications. VGI is closely related to the concept of crowdsourcing with the notion that information obtained from a crowd of many observers is likely to be closer to the truth than information obtained from one observer (Gooldchid, 2010).
Authoritative geospatial data are full of detailed information, but can be outdated and frequently very costly. Furthermore, they often do not contain all the information that is needed by final user. On the contrary, in Internet, a large amount of geo-referenced data is generated by the VGI projects through the everyday experience of people. This information is constantly updated by people that use handhelds, pocket PCs or mobile phones connected to the Web. For this reason we decided to use it to increase the information base of the SS&C database. Information from Internet is used to provide additional information related to the locations monitored by means of UAVs, without in any way replacing the valuable information collected by the UAVs payload sensors but integrating it in the resulting interactive map performing a sort of mashup.
Information on locations provided by VGIs will be henceforth referred as metadata.
Metadata are extracted from several XML-based (Bray, Paoli, Sperberg-McQueen, Maler & Yergeau, 1997) files, downloaded from the Web through Web Services provided by open, collaborative projects like OpenStreetMap (www.geonames.org). XML files are checked by the MDR system to be in a well-formed format. Metadata are also cleaned for the possible presence of duplicates and errors.
Finally, metadata can be used to produce a geo-summary of the target area. The geo-summary is composed by annotations filtered from the amount of metadata downloaded from the Web. The selection of the features of the geo-summary occurs by a statistical test, to guarantee statistical significance and as such, an increased level of reliability. The test aims to identify the typical features of a target area. The typical features are those that in the target area are surprising because in contrast to the surroundings.