Processing and analysis of multitemporal and hypertemporal remote sensing images for environmental monitoring

COFIN 02 Bruzzone DIT-PRJ-02-016

Status NOT active project
DISI role Coordinator
Project type Research Project
Dimension National
Acquisition date 2003-01-01
Start date 2003-01-01
End date 2004-12-31
SAP code 40100554

Project details

Project astract This research program deals with the processing and analysis (by using automatic and semi-automatic techniques) of multitemporal and hypertemporal remote sensing data. The main project objective consists in developing, experimenting and validating a set of advanced methodologies that process remotely-sensed images and signals in the temporal domain. Some of the most critical problems related to the multi/hypertemporal analysis will be considered. The developed methodologies will be devoted to the analysis of various remote-sensing data, with a particular attention to data provided by recent or near future Earth observation missions. The applications considered for the project will target environmental monitoring.<br/>The research activity will focus on two strategic (and at the same time critical) aspects in the analysis of remote sensing data in the temporal domain. The first one is related to multitemporal data processing, whereas the second one is connected to hypertemporal data processing. The different temporal sampling rate of these two kinds of data involves the development of different analysis methods, which can fully exploit their intrinsic properties. In particular, the project will face the following research topics:<br/>- Detection of changes and identification of their typology in multitemporal images by using unsupervised and partially supervised techniques;<br/>- Classification of multitemporal image by using partially supervised approaches;<br/>- Change (anomaly) detection and tracking in hypertemporal images;<br/>- Compression of multitemporal and hypertemporal data, by using methods that can adequately exploit the correlation in the temporal domain and the peculiarities of remote-sensing data.<br/>The work will be organized into 4 temporal phases from a chronological viewpoint, and into 4 Work Packages (WPs) from the thematic viewpoint. Within the whole project activity, 5 coordination meetings among the 5 Research Units (RUs) involved in the project are planned, in addition to technical meetings within each WP. At the end of each temporal phase, a technical report (which will highlight the work progress of each WP) will be released. In particular, the reports will provide a description of the methodological developments, the implemented techniques and algorithms, the developed software and the obtained experimental results. At the end of the project, a software package (composed of the software modules coordinately developed by each RU during the project activity) will be supplied. <br/>The importance of this research program consists in reaching the following targets:<br/>- Developing advanced methodologies and analysis tools to process multi/hypertemporal remote-sensing data effectively;<br/>- Contributing to define effective tools for addressing environmental monitoring problems in relation to fundamentals political issues (e.g., the Kyoto Protocol) and, in general, to increase the use of remote sensing data in real applications;<br/>- Providing an integrated software package for the analysis of multi/hypertemporal remote sensing data within applications related to environmental monitoring;<br/>- Contributing to evolve the remote sensing research area, which is of primary importance for the Telecommunications field;<br/>- Consolidating the Italian position within the international scientific community in the area of remote-sensing data analysis. <br/>
Keywords MULTITEMPORAL REMOTE SENSING IMAGES ; HYPERTEMPORAL REMOTE SENSING IMAGES ; CHANGE DETECTION ; MULTITEMPORAL CLASSIFICATION ; COMPRESSION OF MULTITEMPORAL AND HYPERTEMPORAL REMOTE SENSING IMAGES ; REMOTE SENSING ; IMAGE PROCESSING ; PATTERN RECOGNITION ;
Fundings 246100 €
Partners
  • DIT
  • MIUR

DISI Sub-project details

Project astract The Research Unit (RU) of the University of Trento (UniTN) is in charge of the national coordination of the entire research project. In this context, the RU of Trento will be the responsible of the Work Package WP1 ("Coordination of the Research and Software Package Integration"). This Work Package mainly deals with the project coordination and integration of the software package that will be the main output of the research carried out within the present project. It is worth noting that, among the considered aspects related to the coordination, a particular attention will be given to the monitoring of the activity progress and to the monitoring of the development of both the software modules and the final software package.<br/>However, the most significant contribution of the RU of Trento will be in the research activity planned within the WP2 and devoted to the analysis of multitemporal remote sensing images. The scope of such an activity is to develop advanced methodologies capable to overcome the main limitations of the techniques (often very simple and empiric) that are presently used to detect changes in multitemporal remote sensing data. Such methodologies will be developed taking into account the constraints typically imposed by real applications.
Keywords MULTITEMPORAL REMOTE SENSING IMAGES ; CHANGE DETECTION ; MULTITEMPORAL CLASSIFICATION ; REMOTE SENSING ; PARTIALLY SUPERVISED TECHNIQUES ; UNSUPERVISED TECHNIQUES ; PATTERN RECOGNITION
Fundings 71100 €
Manager Lorenzo Bruzzone