Mattia Stasolla

Robust Extraction of Urban Area Extents in HR and VHR SAR Images

This work is devoted to analyze the performance and adapt the parameters requested for the operational use of a methodology aimed at urban extent extraction. The procedure, initially proposed in a reduced version in a previous paper, has been expanded and improved to make it useful with different HR and VHR radar sensors, and extensive comparison of the results in many different parts of the world have been considered. In this work the approach is compared against reference settlement extents obtained from maps provided by the most relevant global mapping projects. Considerations about the robust ness of the approach to different spatial resolution, adaptiveness of the parameter range to the SAR sensor characteristics and other issues dealing with practical implementation of the whole procedures complete the research work discussed in these pages.

Unstructured Human Settlement Mapping with SAR Sensors

In this paper human settlement detection using SAR data is addressed, with stress on informal settlement analysis. We show that, even using coarse spatial resolution SAR data, it is possible to discriminate between structured and unstructured settlements. In particular, two approaches are proposed: a supervised procedure to exploit spatial patterns and backscatter intensity to discriminate between structured and unstructured kinds of human settlements, and an unsupervised procedure to detect human settlements in large area SAR images.

A detailed comparison between two fast approaches to urban extent extraction in VHR SAR images

This work is devoted to he comparison of two algorithms for human settlement map extraction from VHR SAR data. The two approaches have been recently proposed in literature, but extensive comparison of their performance in different situation and in different areas of the world was not available yet.

Characterization of lossy layers through monostatic radar measurements

This paper presents an inversion technique based on monostatic radar measurements for the e.m. characterization of lossy slabs. Under the hypotesis of low-loss materials, it is possible to analytically bound permittivity, conductivity and thickness of a slab to the amplitudes and time of delays of the detected echoes. The analysis of various materials has shown that the methodology can provide very good results. It has been anyway confirmed that too high values of conductivity will lead to a significant increase of the errors or even the unfeasibility of the inversion process.

Electromagnetic infrastructure monitoring: The exploitation of GPR data and neural networks for multi-layered geometries

In this paper, an inversion ANN-based algorithm for the estimation of geophysical properties (i.e. thickness and permittivity) of subsurface layers in stratified geometries is presented. The basic procedure for the analysis of GPR scans of single subsurface layers placed over a uniform background recently proposed by the authors has been here extended and inserted into a general framework where each stratum is recursively processed.

Humanitarian Aids Using Satellite Technology

One of the main topics the remote sensing community is interested in regards the monitoring of informal settlements for humanitarian aids, as proved by a number of international projects like the European RESPOND in the framework of GMES (Global Monitoring for Environment and Security) or United Nations’ UNOSAT. This chapter discusses not only the possibility of employing remote sensing imagery to this aim, but above all the capability of semi-automated procedures to analyze such data and to assist the work of Administrations and NGOs. Test areas are located in Darfur region, Sudan, which became in 2003 the scene of one of the worst humanitarian crises of our age. Optical images of those territories were acquired by SPOT-5 and Quickbird satellites between 2003 and 2005, and high resolution radar data by the Japanese PALSAR sensor on board of the ALOS satellite in 2006, after refugee camps were built up for accommodating hundreds of thousands of displaced people. The proposed algorithms intend to provide land-cover/use maps that can be useful to keep changes under control and/or to update existing charts.

GIS and Remote Sensing for Disaster Assessment in Urban Areas

This paper is devoted to the exploitation of remotely sensed and geographic information system (GIS) data in urban areas to obtain a rapid damage assessment after disastrous events, like earthquakes, occurred. The basic idea is that the joint use of these two sources of information will provide a better understanding of the urban scene and allow a faster characterization of the damage patterns, extensions and possibly levels. This is achievable at different scales, as presented in the paper, according to the spatial resolution of the remotely sensed data but also to the quality and quantity of GIS data available. Examples for the Golcuk (Turkey) and Bam (Iran) earthquakes help to understand the advantages and limits of the proposed methodology.

Towards the Detection of Multiple Reflections in Time-Domain EM Inverse Scattering of Multi-Layered Media

In this paper, a new theoretical approach for the classiflcation of multiple re∞ections in time-domain e.m. inverse scattering of multi-layered media is presented. The existence of multiples limits the capabilities of inversion algorithms, thus suitable identiflcation and suppression techniques should be applied to reduce this undesired efiect. Assuming a scenario composed of loss-less and non-dispersive media, and providing an accurate time delay estimation (TDE) of backscattered signals, the proposed method allows not only to evaluate the presence of multiples and discriminate them from primary re∞ections, but also to determine their propagation paths. Preliminary tests performed on FDTD simulated data have shown its potentialities to efiectively handle multiple re∞ections and therefore to enhance the e.m. signals backscattered by primary re∞ectors.

Fusion of SAR and optical data for urban extent extraction improvement

This paper presents two methods to fuse SAR and optical data for urban extent extraction. The two methodologies build over single sensors procedure in order to improve the efficiency of the characterization of the urban environment when more data is available. Results over Pavia and Al Fashir conform the effectiveness of the proposed procedures.

A neural network electromagnetic approach for GPR pavement diagnostic: A preliminary study

In this paper, a preliminary study, based on GPR data analysis by means of artificial neural networks, for automatic pavement diagnostic is addressed. The proposed solving solution models the pavement as a multi-layered medium composed of N parallel homogenous layers, which are separately analyzed through a recursive procedure able to reconstruct their permittivity and thickness. The basic processing module of the whole procedure, is here presented.