The new constellation of remote sensing satellite COSMO/SkyMed will guarantee a combination of spatial and temporal resolution never reached by previously systems. The full exploitation of this system can allow the development of new applications, like these aiming at providing insight into the magnitude of a disaster and a detailed assessment of the damages as required by first responders for planning relief actions. The problem posed by the necessity of processing a huge number of images looking for given objects cannot be afforded by using visual approaches. This paper aims at describing the results obtained by applying some algorithms able to fully exploit the performances of the COSMO constellation. The technique herein described represent a generalization to radar images of the methods successfully applied to optical data. The techniques we are referring to are based on Mathematical Morphology and have been developed in the mainframe of the EU funded Network of Excellence (NoE) GMOSS (Global Monitoring for Stability and Security). In that case this technique has been applied to the problem of detecting objects, belonging to very different contexts, like dwelling units in refugee camps, roads of complex shapes and different background, main structures in nuclear plants, etc. In particular, techniques able to automatically: extract mademan structures, which could be present in mosaic of images, detect and counting dwelling units in refugee camps, extract roads of complex shape or for monitoring nuclear plants, have been developed. The purpose of the present study is the assessment of the suitability of the same mathematical morphology techniques for detecting automatically roads/streets, verifying their state after a disastrous event, in both urban and extra-urban areas on radar images. Actually the objects of interest are detected by exploiting mathematical morphology and some ancillary information regarding the shape and size of the required object. ©2009 IEEE.
Application of Mathematical Morphology to Automatically Extract roads on Radar Images / Laneve, Giovanni; Santilli, Giancarlo; Cadau, E. G.. - (2009). (Intervento presentato al convegno Urban Remote Sensing Joint Event tenutosi a Shanghai; China nel Maggio) [10.1109/URS.2009.5137732].
Application of Mathematical Morphology to Automatically Extract roads on Radar Images
LANEVE, Giovanni;SANTILLI, GIANCARLO;
2009
Abstract
The new constellation of remote sensing satellite COSMO/SkyMed will guarantee a combination of spatial and temporal resolution never reached by previously systems. The full exploitation of this system can allow the development of new applications, like these aiming at providing insight into the magnitude of a disaster and a detailed assessment of the damages as required by first responders for planning relief actions. The problem posed by the necessity of processing a huge number of images looking for given objects cannot be afforded by using visual approaches. This paper aims at describing the results obtained by applying some algorithms able to fully exploit the performances of the COSMO constellation. The technique herein described represent a generalization to radar images of the methods successfully applied to optical data. The techniques we are referring to are based on Mathematical Morphology and have been developed in the mainframe of the EU funded Network of Excellence (NoE) GMOSS (Global Monitoring for Stability and Security). In that case this technique has been applied to the problem of detecting objects, belonging to very different contexts, like dwelling units in refugee camps, roads of complex shapes and different background, main structures in nuclear plants, etc. In particular, techniques able to automatically: extract mademan structures, which could be present in mosaic of images, detect and counting dwelling units in refugee camps, extract roads of complex shape or for monitoring nuclear plants, have been developed. The purpose of the present study is the assessment of the suitability of the same mathematical morphology techniques for detecting automatically roads/streets, verifying their state after a disastrous event, in both urban and extra-urban areas on radar images. Actually the objects of interest are detected by exploiting mathematical morphology and some ancillary information regarding the shape and size of the required object. ©2009 IEEE.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
