Irena Hajnsek: Catalogue data in Spring Semester 2021

Name Prof. Dr. Irena Hajnsek
FieldEarth Observation (Microwave Remote Sensing)
Address
Institut für Umweltingenieurwiss.
ETH Zürich, HIF D 89.2
Laura-Hezner-Weg 7
8093 Zürich
SWITZERLAND
Telephone+41 44 633 74 55
E-mailhajnsek@ifu.baug.ethz.ch
DepartmentCivil, Environmental and Geomatic Engineering
RelationshipFull Professor

NumberTitleECTSHoursLecturers
101-0522-10LDoctoral Seminar Data Science and Machine Learning in Civil, Env. and Geospatial Engineering Restricted registration - show details
Number of participants limited to 21.
1 credit2SB. Soja, E. Chatzi, F. Corman, O. Fink, I. Hajnsek, M. A. Kraus, M. Lukovic, K. Schindler, M. J. Van Strien
AbstractCurrent research in machine learning and data science within the research fields of the department. The goal is to learn about current research projects at our department, to strengthen our expertise and collaboration with respect to data-driven models and methods, to provide a platform where research challenges can be discussed, and also to practice scientific presentations.
Learning objective- learn about discipline-specific methods and applications of data science in neighbouring fields
- network people and methodological expertise across disciplines
- establish links and discuss connections, common challenges and disciplinespecific differences
- practice presentation and discussion of technical content to a broader, less specialised scientific audience
ContentCurrent research at D-BAUG will be presented and discussed.
Prerequisites / NoticeThis doctoral seminar is intended for doctoral students affiliated with the Department of Civil, Environmental and Geomatic Engineering. Other students who work on related topics need approval by at least one of the organisers to register for the seminar.

Participants are expected to possess elementary skills in statistics, data
science and machine learning, including both theory and practical modelling and implementation. The seminar targets students who are actively working on related research projects.
101-0523-11LFrontiers in Machine Learning Applied to Civil, Env. and Geospatial Engineering (FS21) Restricted registration - show details
Number of participants limited to 21.
1 credit2SM. Lukovic, E. Chatzi, F. Corman, O. Fink, I. Hajnsek, M. A. Kraus, K. Schindler, B. Soja, M. J. Van Strien
AbstractThis doctoral seminar organised by the D-BAUG platform on data science and machine learning aims at discussing recent research papers in the field of machine learning and analyzing the transferability/adaptability of the proposed approaches to applications in the field of civil and environmental engineering (if possible and applicable, also implementing the adapted algorithms).
Learning objectiveStudents will
• Critically read scientific papers on the recent developments in machine learning
• Put the research in context
• Present the contributions
• Discuss the validity of the scientific approach
• Evaluate the underlying assumptions
• Evaluate the transferability/adpatability of the proposed approaches to own research
• (Optionally) implement the proposed approaches.
ContentWith the increasing amount of data collected in various domains, the importance of data science in many disciplines, such as infrastructure monitoring and management, transportation, spatial planning, structural and environmental engineering, has been increasing. The field is constantly developing further with numerous advances, extensions and modifications.
The course aims at discussing recent research papers in the field of machine learning and analyzing the transferability/adaptability of the proposed approaches to applications in the field of civil and environmental engineering (if possible and applicable, also implementing the adapted algorithms).
Each student will select a paper that is relevant for his/her research and present its content in the seminar, putting it into context, analyzing the assumptions, the transferability and generalizability of the proposed approaches. The students will also link the research content of the selected paper to the own research, evaluating the potential of transferring or adapting it. If possible and applicable, the students will also implement the adapted algorithms The students will work in groups of three students, where each of the three students will be reading each other’s selected papers and providing feedback to each other.
Prerequisites / NoticeThis doctoral seminar is intended for doctoral students affiliated with the Department of Civil, Environmental and Geomatic Engineering. Other students who work on related topics need approval by at least one of the organisers to register for the seminar.

Participants are expected to possess elementary skills in statistics, data science and machine learning, including both theory and practical modelling and implementation. The seminar targets students who are actively working on related research projects.
102-0617-01LMethodologies for Image Processing of Remote Sensing Data3 credits2GI. Hajnsek, O. Frey, S. Li
AbstractThe aim of this course is to get an overview of several methodologies/algorithms for analysis of different sensor specific information products. It is focused at students that like to deepen their knowledge and understanding of remote sensing for environmental applications.
Learning objectiveThe course is divided into two main parts, starting with a brief introduction to remote sensing imaging (4 lectures), and is followed by an introduction to different methodologies (8 lectures) for the quantitative estimation of bio-/geo-physical parameters. The main idea is to deepen the knowledge in remote sensing tools in order to be able to understand the information products, with respect to quality and accuracy.
ContentEach lecture will be composed of two parts:
Theory: During the first hour, we go trough the main concepts needed to understand the specific algorithm.
Practice: During the second hour, the student will test/develop the actual algorithm over some real datasets using Matlab. The student will not be asked to write all the code from scratch (especially during the first lectures), but we will provide some script with missing parts or pseudo-code. However, in the later lectures the student is supposed to build up some working libraries.
Lecture notesHandouts for each topic will be provided.
LiteratureSuggested readings:
T. M. Lillesand, R.W. Kiefer, J.W. Chipman, Remote Sensing and Image Interpretation, John Wiley & Sons Verlag, 2008
J. R. Jensen, Remote Sensing of the Environment: An Earth Resource Perspective, Prentice Hall Series in Geograpic Information Science, 2000