foreword to the special issue on analysis of multitemporal ...doctoral research associate with the...

3
1 Foreword to the Special Issue on Analysis of Multitemporal Data and Applications Emmanuel Trouvé, Grégoire Mercier, Mathieu Fauvel, Lorenzo Bruzonne and Yifang Ban After 40 years of Earth Observation missions with both passive (multispectral, hyperspectral, etc.) and active (synthetic aperture radar, lidar, etc.) sensors, remote sens- ing data offer a unique opportunity to record, to analyze and to predict the evolution of our living planet. In the last decade, a large number of new satellite remote sensing missions have been launched, resulting in dramatic improvement in the image acquisition capabil- ities. The Landsat open archives program, the successful launching of the Sentinel-1 in 2014 and Sentinel-2 in 2015, with regular acquisition plans and free data access policy, result in new challenges for handling and processing such huge volume of data. This increasing number of Earth Observation systems involves an enhanced possibility to acquire multitemporal images of the Earth surface, with improved temporal and spatial resolution. Such new sce- nario significantly increases the interest of the time series processing in the remote sensing community. The develop- ment of novel data processing techniques to address new important and challenging applications is promising. From June 22 to 24, 2015, the 8th International Work- shop on the Analysis of Multitemporal Remote Sensing Images (MultiTemp 2015) has been held successfully in Annecy, France, with 150 participants from 29 countries. The General Chair, Prof. Emmanuel Trouvé from the University Savoie Mont Blanc, and the Techincal Chair Prof. Grégoire Mercier from the Télécom Bretagne had set up a technical program that covered issues related to multitemporal data processing, to the analysis of time series acquired by passive and active sensors and to their related applications. The 2nd EARSeL International Workshop on Temporal Analysis of Satellite Images (Temporal Analysis 2015) was chaired by Prof. Yifang Ban and held at KTH Royal Institute of Technology in Stockholm, Sweden during June 15-18, 2015. The workshop had 108 participants from 30 countries with 55 presentations. The workshop topics range from multitemporal analysis techniques and Change Detection methods to application in agriculture, forestry, urban, land-use/land-cover, landscape and vegetation dy- namics, as well as Glacier, Ice Sheet and Permafrost. This special issue of the IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing (JSTARS) presents extended contributions to MultiTemp 2015 and Temporal Analysis 2015. 21 submis- sions to the special issues as well as 25 regular submis- sions to JSTARS were selected for the special issue. The main topics addressed in this issue are change detection, parameters estimation and land-use/land-cover mapping (see Figure 1). Indeed, the detection of slow and abrupt changes occurring between two or several dates is the main application of the time series, since 18 papers deal with this issue. However, time series are also used to measure surface elevation changes or to infer surface parameters, such as temperature or normalized difference vegetation index (NDVI), that are used to environmental monitor- ing. Phenological variation is also used to derive land- cover/land-use map, for vegetation-type classes. Fig. 1. Word clouds of the 50 most frequent words in the titles of the papers. An analysis of the research themes shows that SAR data are the most popular source of data in this issues. SAR time series are used in approximately half of the paper in the special issues. They are used in various applications, ranging from urban area to vegetation monitoring. Optical data, such as MODIS or Landsat time series, are usually used through the NDVI. With the recent availability of Sentinel-2 time series, we expect that these optical time series will be more investigated in the future. The range of research presented in this special issues provides a picture of current methods and use-cases of multitemporal data. We hope this view will inspire new scientific and technical developments in the area of multi- temporal data. Acknowledgement The Guest Editors would like to acknowledge the assis- tance of Prof. J. Chanussot and Prof. Quian (Jenny) Du, respectively past and current Editor-in-Chief of JSTARS, and all the authors and reviewers for their outstanding contributions.

Upload: others

Post on 10-Feb-2020

1 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Foreword to the Special Issue on Analysis of Multitemporal ...doctoral research associate with the MISTIS Team of the National Institute for Research in Computer Science and Control

1

Foreword to the Special Issue on Analysis ofMultitemporal Data and Applications

Emmanuel Trouvé, Grégoire Mercier, Mathieu Fauvel, Lorenzo Bruzonne and Yifang Ban

After 40 years of Earth Observation missions withboth passive (multispectral, hyperspectral, etc.) and active(synthetic aperture radar, lidar, etc.) sensors, remote sens-ing data offer a unique opportunity to record, to analyzeand to predict the evolution of our living planet.

In the last decade, a large number of new satelliteremote sensing missions have been launched, resulting indramatic improvement in the image acquisition capabil-ities. The Landsat open archives program, the successfullaunching of the Sentinel-1 in 2014 and Sentinel-2 in 2015,with regular acquisition plans and free data access policy,result in new challenges for handling and processing suchhuge volume of data. This increasing number of EarthObservation systems involves an enhanced possibility toacquire multitemporal images of the Earth surface, withimproved temporal and spatial resolution. Such new sce-nario significantly increases the interest of the time seriesprocessing in the remote sensing community. The develop-ment of novel data processing techniques to address newimportant and challenging applications is promising.

From June 22 to 24, 2015, the 8th International Work-shop on the Analysis of Multitemporal Remote SensingImages (MultiTemp 2015) has been held successfully inAnnecy, France, with 150 participants from 29 countries.The General Chair, Prof. Emmanuel Trouvé from theUniversity Savoie Mont Blanc, and the Techincal ChairProf. Grégoire Mercier from the Télécom Bretagne hadset up a technical program that covered issues relatedto multitemporal data processing, to the analysis of timeseries acquired by passive and active sensors and to theirrelated applications.

The 2nd EARSeL International Workshop on TemporalAnalysis of Satellite Images (Temporal Analysis 2015) waschaired by Prof. Yifang Ban and held at KTH RoyalInstitute of Technology in Stockholm, Sweden during June15-18, 2015. The workshop had 108 participants from30 countries with 55 presentations. The workshop topicsrange from multitemporal analysis techniques and ChangeDetection methods to application in agriculture, forestry,urban, land-use/land-cover, landscape and vegetation dy-namics, as well as Glacier, Ice Sheet and Permafrost.

This special issue of the IEEE Journal of SelectedTopics in Applied Earth Observations and RemoteSensing (JSTARS) presents extended contributions toMultiTemp 2015 and Temporal Analysis 2015. 21 submis-sions to the special issues as well as 25 regular submis-sions to JSTARS were selected for the special issue. Themain topics addressed in this issue are change detection,parameters estimation and land-use/land-cover mapping

(see Figure 1). Indeed, the detection of slow and abruptchanges occurring between two or several dates is the mainapplication of the time series, since 18 papers deal withthis issue. However, time series are also used to measuresurface elevation changes or to infer surface parameters,such as temperature or normalized difference vegetationindex (NDVI), that are used to environmental monitor-ing. Phenological variation is also used to derive land-cover/land-use map, for vegetation-type classes.

Fig. 1. Word clouds of the 50 most frequent words in the titles ofthe papers.

An analysis of the research themes shows that SAR dataare the most popular source of data in this issues. SARtime series are used in approximately half of the paper inthe special issues. They are used in various applications,ranging from urban area to vegetation monitoring. Opticaldata, such as MODIS or Landsat time series, are usuallyused through the NDVI. With the recent availability ofSentinel-2 time series, we expect that these optical timeseries will be more investigated in the future.The range of research presented in this special issues

provides a picture of current methods and use-cases ofmultitemporal data. We hope this view will inspire newscientific and technical developments in the area of multi-temporal data.

AcknowledgementThe Guest Editors would like to acknowledge the assis-

tance of Prof. J. Chanussot and Prof. Quian (Jenny) Du,respectively past and current Editor-in-Chief of JSTARS,and all the authors and reviewers for their outstandingcontributions.

Page 2: Foreword to the Special Issue on Analysis of Multitemporal ...doctoral research associate with the MISTIS Team of the National Institute for Research in Computer Science and Control

2

Emmanuel Trouvé (M’00–SM’10) receivedthe Engineer degree in electrical engineeringfrom the École Nationale Supérieure de Tech-niques Avancées, Paris, France, in 1990, thePh.D. degree in signal and image processingfrom Ecole Nationale Supérieure des Télé-communications, Paris, France, in 1996, andthe “Habilitation à Diriger des Recherches”degree from Université of Savoie, Chambéry,France, in 2006. From 1996 to 1998, he workedwith Thomson Marconi Sonar in Underwater

Acoustic and Signal Processing. He has been working with Universityof Savoie, Polytech Annecy-Chambéry, Laboratoire d’Informatique,Systèmes, Traitement de l’Information et de la Connaissance, as anAssociate Professor (1998–2008), then as a Professor of signal and im-age processing. His research interests include synthetic aperture radar(SAR) image processing and data fusion in remote sensing. Prof.Trouvé has been coordinating several research projects, includingthe national multilaboratory project EFIDIR (Extraction and Fusionof Information for Displacement measurement from SAR Imagery,2008–2012). He was the General Chair of MultiTemp 2015 (8thInternational Workshop on the Analysis of Multitemporal RemoteSensing Images) and the treasurer of the French Chapter of the IEEEGeoscience and Remote Sensing Society from 2013 to 2016.

Grégoire Mercier was born in France in1971. He received the Engineer Degree fromthe Institut National des Télécommunications,Evry, France in 1993, his Ph.D. degree fromthe University of Rennes I, Rennes, Francein 1999 and his Habilitation à Diriger desRecherches from the University of Rennes I in2007. Since 1999, he has been with the EcoleNationale Supérieure des Télécommunicationsde Bretagne, where he has been an AssociateProfessor in the Image and Information Pro-

cessing department (ITI). He is now Professor since early 2010.His research interests are in remote sensing image compression andsegmentation, especially in hyperspectral and Synthetic ApertureRadar. Actually, his research is dedicated to change detection andcombating pollution. He was a visiting researcher at DIBE (Univer-sity of Genoa, Italy) from March to May 2006 where he developedchange detection technique for heterogeneous data. He was also avisiting researcher at CNES (France) from April to June 2007 to takepart of the Orfeo Toolbox development. Since 2012, he is an externalcollaborator of the AYIN research group of the INRIA. Prof. Mercierhas been an Associate Editor for the IEEE Geoscience and RemoteSensing Letters (GRSL) from 2007 to 2014. He has been President ofthe French Chapter of IEEE Geoscience and Remote Sensing Societyfrom 2010 to 2013. He is now Associate Editor for the IEEE Journal ofSelected Topics in Applied Earth Observations and Remote Sensing(JSTARS). He serves as Head of the Technical Commitee of the IEEEMultiTemp’2015 workshop.

Mathieu Fauvel graduated in electrical en-gineering from the Grenoble Institut of Tech-nology (Grenoble INP), Grenoble, France, in2004. He received the M.S and Ph.D. degreesin image and signal processing from the Greno-ble INP in 2004 and 2007, respectively. In2007, he was a teaching assistant in Greno-ble INP. From 2008 to 2010, he was a post-doctoral research associate with the MISTISTeam of the National Institute for Researchin Computer Science and Control (INRIA).

Since 2011, Dr. Fauvel has been an Associate Professor with theNational Polytechnic Institute of Toulouse (ENSAT - Universityof Toulouse) within the DYNAFOR lab (University of Toulouse- INRA). His research interests are remote sensing, data fusion,pattern recognition, multicomponent signal and image processing.From March 2013 to March 2016, he was the president of the FrenchChapter of the GRSS.

Lorenzo Bruzzone (S’95-M’98-SM’03-F’10)received the Laurea (M.S.) degree in electronicengineering (summa cum laude) and the Ph.D.degree in telecommunications from the Uni-versity of Genoa, Italy, in 1993 and 1998,respectively. He is currently a Full Professorof telecommunications at the University ofTrento, Italy, where he teaches remote sens-ing, radar, pattern recognition, and electricalcommunications. Dr. Bruzzone is the founderand the director of the Remote Sensing Labo-

ratory in the Department of Information Engineering and ComputerScience, University of Trento. His current research interests are in theareas of remote sensing, radar and SAR, signal processing, and pat-tern recognition. He promotes and supervises research on these topicswithin the frameworks of many national and international projects.Among the others, he is the Principal Investigator of the Radar foricy Moon exploration (RIME) instrument in the framework of theJUpiter ICy moons Explorer (JUICE) mission of the European SpaceAgency. He is the author (or coauthor) of 161 scientific publicationsin referred international journals (111 in IEEE journals), more than220 papers in conference proceedings, and 17 book chapters. He iseditor/co-editor of 16 books/conference proceedings and 1 scientificbook. His papers are highly cited, as proven form the total numberof citations (more than 15300) and the value of the h-index (63)(source: Google Scholar). He was invited as keynote speaker in 30international conferences and workshops. Since 2009 he is a memberof the Administrative Committee of the IEEE Geoscience and Re-mote Sensing Society. Dr. Bruzzone ranked first place in the StudentPrize Paper Competition of the 1998 IEEE International Geoscienceand Remote Sensing Symposium (Seattle, July 1998). Since thattime he was recipient of many international and national honors andawards. Dr. Bruzzone was a Guest Co-Editor of different SpecialIssues of international journals. He is the co-founder of the IEEEInternational Workshop on the Analysis of Multi-Temporal Remote-Sensing Images (MultiTemp) series and is currently a member ofthe Permanent Steering Committee of this series of workshops. Since2003 he has been the Chair of the SPIE Conference on Image andSignal Processing for Remote Sensing. Since 2013 he has been thefounder Editor-in-Chief of the IEEE Geoscience and Remote SensingMagazine. Currently he is an Associate Editor for the IEEE Transac-tions on Geoscience and Remote Sensing and the Canadian Journalof Remote Sensing. Since 2012 he has been appointed DistinguishedSpeaker of the IEEE Geoscience and Remote Sensing Society.

Page 3: Foreword to the Special Issue on Analysis of Multitemporal ...doctoral research associate with the MISTIS Team of the National Institute for Research in Computer Science and Control

3

Yifang Ban is the Director of the Division ofGeoinformatics, and Vice Chair of the Depart-ment of Urban Planning and Environment,KTH Royal Institute of Technology in Stock-holm, Sweden. Before joining KTH in 2004,Dr. Ban was a tenured Associate Professorat York University in Toronto, Canada. Shereceived a BSc and MSc from Nanjing Univer-sity, China in 1984 and 1987 respectively, anda PhD from University of Waterloo, Canadain 1996.

Professor Ban’s research interests include multitemporal remote sens-ing, SAR image analysis, multisensor data fusion, image segmen-tation and classification, change detection, urban and land covermapping and monitoring, as well as environmental impact assess-ment. She is the PI for the ESA INNOVATOR III EO4Urbanproject and the European Lead PI for the Urbanization and Environ-ment/Climate project within the ESA/China Dragon 2-3 Program,among others. She is a co-lead of Group on Earth Observations(GEO) sub-task ‘Global Urban Observation and Information’, andchairs the ISPRS Inter-Commission II/IV/VIII Working Group onGlobal Land Cover Mapping and Services. She is a co-chair for ICACommission on Sensor-Driven Mapping and a co-chair for EuropeanAssociation of Remote Sensing Laboratories (EARSeL) Special In-terest Group ‘Temporal Analysis of Satellite Images’. She has beena scientific committee member for a number of major internationalremote sensing conferences and serves as reviewer for key remotesensing and GIS journals. She served as an Associate Editor forIEEE Journal on Selected Topics in Applied Earth Observations andRemote Sensing (JSTARS) during 2013-2016. She was also a GuestEditor for the JSTARS Special Issue on the Analysis of Multitem-poral Remote Sensing Data and ISPRS Journal of Photogrammetryand Remote Sensing Special Issue on Global Land Cover Mappingduring 2014-2015.