Land monitoring by advanced satellite remote sensing

Teacher(s)

Bontemps Sophie; Defourny Pierre;

Language

English
> French-friendly

Prerequisites

LBIRE2102 Géomatique appliquée or equivalent introductory class in remote sensing   
Pogramming skills (R, python)
Related courses :
LBRAT2102 Modélisation spatiale des dynamiques territoriales
LBRES2101 Smart technologies for environmental engineering
LBRAI2221 Agriculture de précision et mécanisation

Main themes

This course aims to develop in-depth understanding and professional skills to process and interpret very high resolution UAV (drone) imagery and Earth Observation satellite time series. Advanced concepts related to signal acquisition, time series quality control and uncertainty characterization are introduced. Radiative transfer modeling and methods for biophysical variables estimation (Leaf Area Index, biomass, nitrogen status, surface temperature, evapotranspiration, soil moisture, height, etc.) and change detection methods are explained and illustrated through practical applications and the European Copernicus Services. Finally,open source tools and systems supporting already operational and forthcoming monitoring systems, including flood monitoring, fire monitoring, forest monitoring and crop monitoring, are discussed in details.
The objective of this course is to develop the necessary knowledge and technical skills to use advanced image processing methods (including machine learning and artificial intelligence) and to implement workflow for UAV or satellite monitoring applications in the field of agriculture, forestry, land use land cover change, and water resources management.

Learning outcomes

At the end of this learning unit, the student is able to :

At the end of this learning unit, the student is able to : 1 a. Contribution of the activity to the program learning outcomes Consistency of LO courses with those of the program M1.1., M2.1., M4.4., M4.5 b. Specific formulation for this AA activity of the program (maximum 10) At the end of this activity, the student is able to: practically mobilize the advanced concepts and methods of airborne and satellite remote sensing applied to the monitoring and the management of natural resources, to regional planning and to the environment in general; understand and critize in depth the operational services, the available products and the existing tools to get the best out of each; mastering specialized open source remote sensing softwares and developping processing chains including several tools; design and conduct rigorous digital analyzes of optical and radar time series to respond to specific issues belonging to the bioengineer fields and to formulate the related hypotheses and limits; be able to grasp technological developments in the field of remote sensing applied to the fields of the bioengineers.

Content

The course consists of lectures and computer-lab practical exercises, primarily based on open-source software, used to carry out a professional-level case study of the student's choice. The lessons cover the following topics:

• the stages of signal acquisition and pre-processing, including quality indicators and uncertainty management;
• radiative transfer modeling and estimation of biophysical variables;
• analysis of optical and SAR time series, feature extraction, and pixel-based or object-based metrics; advanced processing of radar time series, including at the polarimetric and interferometric levels;
• introduction to machine learning and artificial intelligence algorithms for mapping, land monitoring, and change detection;
• validation methods and algorithm performance metrics;
• case studies focused on the use of earth observation data related to the environmental monitoring (glaciers, fires, flooding, landslides, ...), agriculture, forestry, water resources, and land-use planning.

Teaching methods

The teachnig introduces the concepts and advanced methods while the praticals in computer lab mobilise them in the context of specific applications. The lessons are quite interactive and thre practicals relies on an inductive approach based on a case study of your choice.
The course and the praticals aims to develop on one hand advanced technical skills in Earth 0bservation data processing and on the other hand, the ability of critical analysis  with regards existing solutions, services and products. The student learns not only to use open source packages and Google Earth Engine environment but also to assess the quality and to review the validity of the proposed algorithms and datasets for a given application.
The practical training is closely linked to the course and includes the use of several open source libraries (including QGIS, SNAP, GDAL, ORFEO, Sen4CAP), the exploitation of the Jupyter notebook environment for quality control et time series analysis, and the workflow coding in Python or R.

Evaluation methods

The assessment is based on the presentation, discussion, and critical analysis of a case study (a concrete application of the student's choice) carried out from start to finish over the course of the term and presented orally in the form of a scientific poster. The entirety of the course material must be drawn upon in the critical analysis and in responding to questions during the examination.

Other information

This course is part of the Certificate in Applied Geomatics accessible to professionals as part of continuing training.

This course might be taught in English.

Online resources

The oral lecture, the lecture slides, and the practical exercise materials constitute the teaching materials for this course. The lecture slides and practical exercise materials are available on the course's Moodle page. Digital resources, such as open-source libraries (in R and Python), are also introduced and recommended during the practical exercises.

Teaching materials

• No paid reference textbook needs to be purchased for this course.

Faculty or entity

> AGRO