Massimo Zanetti in short. I received the MS’s degree in Mathematics (summa cum laude) from the University of Ferrara, Department of Mathematics and Computer Science (Ferrara, Italy), and the PhD (cum laude) in Communication and Information Technologies from the University of Trento, Deptartment of Information Engineering and Computer Science (Trento, Italy). Currently, I am a researcher at the Remote Sensing for Digital Earth Unit, Fondazione Bruno Kessler, (Trento, Italy). My research interests are mainly focused on image analysis in a wide sense. Some specific topics covered are multispectral/hyperspectral images change detection, time series analysis, variational methods for segmentation and numerical optimization methods for image processing problems.
Research activity. Since my MS’s degree in Mathematics at the University of Ferrara, I begun studying remote sensing image problems both from the theoretical and applicative points of view. My MS’s thesis was focused on developing novel mathematical models for the segmentation of very high resolution aerial images and LiDAR data and it was written in collaboration with the Department of Civil, Environmental and Mechanical Engineering at the University of Trento. I continued my studies on remote sensing topics related to image analysis as a PhD student in the Information Engineering and Computer Science Department at the University of Trento. During the three PhD years, I have further expanded the range of my research interests toward multi- and hyper-spectral image analysis with particular emphasis on change detection applications. My PhD thesis includes contributions mainly related to numerical methods for remote sensing image analysis. In particular, theoretical properties of tiling schemes applied to variational methods for large size images segmentation, and, statistical approaches to image modeling and related inference in the framework of change detection. The thesis has being awarded as best thesis of the year by two of the main Italian communities related to information engineering and remote sensing: Gruppo Telecomunicazioni e Tecnologie dell’Informazione (GTTI) and IEEE Geoscience and Remote Sensing Center-North Italy Chapter (IEEE GRS29-CNI).
As a post-doc, I continued my research in the remote sensing field finding applications also in the context of a number of projects at both national and European levels. These include development of advanced techniques for analysis of remote sensing images (founded by Italian Ministry of Education, University and Research), exploitation of satellite-borne operational missions for time-series analysis, change detection and land cover mapping (founded by European Space Agency) and the study of impact of land cover in climate changes through remote sensing data (founded by European Space Agency). In these contexts I have been able to develop and expand my skills in writing successful project proposals, working in a team, interact and establish fruitful partnership with a diversity of collaborators such as academics, engineers, climatologists, physicists, mathematicians as well as industrial partners.
Today's challenge. With the advent of the last generation satellite-borne multispectral sensors a new era is coming concerning the frequent and updated monitoring of our planet. The "open access" policy engaged by NASA USGS Landsat and ESA Sentinels for their entire archives of multispectral images has promoted and boosted a change of paradigm in the approach for Earth Observation through remote sensing. This is further valued considering the remarkable improvements in geometric and spectral resolution of latest images as well as their increased temporal coverage. Not surprisingly, all of this is having a major impact on the techniques related to the detection of changes on our planet via remote sensing image analysis. The capability of acquiring, at no cost, long and dense time series of images over the same geographical area is unprecedented and allows for improving the quality of products associated to specific necessities related to several monitoring/management applications. Given the huge amount of data that now potentially contribute to products definiton, today's main challenge is the development of methods/techniques independent as much as possible of ancillary information and user inputs. Furthermore, this principle must be applied at large scale in order to meet the ever challenging and specific requirements of accuracy and performance imposed by the different applicative contexts.
In order to address research actively, one usually first tries to devop (or find) suitable models that well describe the physics of the phenomena that are originating the data. To provide finest descriptions of what is not actually observable, one then attempts to exploit the mathematical properties that can be derived from the models. This process sometimes makes it possible to explain the solution of the target problem as solution of a coupled mathematical problem such as: an energy optimization problem, or a set of partial differential equations (any others?).