B5192 - MONITORAGGIO E TELERILEVAMENTO DELLE FORESTE

Learning outcomes

At the end of the course, the student has foundational knowledge in remote sensing and machine learning, as well as their applications in forest monitoring. In particular, the student will be able to: (i) pre-process and process multi-platform remote sensing data; (ii) integrate field measurements with satellite-based remote sensing observations; and (iii) combine multiple data sources using artificial intelligence methods.

Course contents

1. Course overview: syllabus, lesson structure, key topics, and final assessment.
2. Light, remote sensing, and spatial, spectral, temporal, and radiometric resolutions.
3. Introduction to Google Earth Engine (GEE) and modelling: regression and classification.
4. Satellite missions, available data, and fields of application.
5. Forest inventories: principles of inferential statistics and sampling strategies.
6. Introduction to programming: basic concepts and data structures in R.
7. R: file management and data handling.
8. R: iterative functions, model validation, overfitting, and cross-validation.
9. R: forest variable mapping by integrating remote sensing and field data.
10. Scientific communication, research careers, and how to structure a scientific paper.
11. GEE: time-series analysis, chart production, and data download.
12. GEE: change detection and satellite-based monitoring of forest disturbances.
13. GEE: monitoring post-disturbance vegetation recovery and forest resilience (C2C approach).
14. GEE: automated classification of satellite imagery.
15. In-class work on the final project.

Readings/Bibliography

Brivio P.A., Lechi G., Zilioli E. (2006) Principi e Metodi di Telerilevamento. CittàStudi Edizioni

Corona P. (2000). Introduzione al rilevamento campionario delle risorse forestali. CLUSF. Firenze

Dainelli N. (2011). L'osservazione della terra. Fotointerpretazione. Dario Flaccovio Editore

Casagrande L., Cavallini P., Frigeri A., Furieri A., Marchesini I., Neteler M. (2012). GIS Open Source. GRASS GIS, Quantum GIS e SpatiaLite. Elementi di software libero applicato al territorio. Dario Flaccovio Editore.

Cloud-Based Remote Sensing with Google Earth Engine. Jeffrey A. Cardille, Morgan A. Crowley, David Saah, Nicholas E. Clinton. https://link.springer.com/book/10.1007/978-3-031-26588-4

Hermosilla, T., Francini, S., Nicolau, A.P., Wulder, M.A.,White, J.C., Coops, N.C. et al. (2024) Clouds and imagecompositing. In: Cardille, J.A., Crowley, M.A., Saah, D. &Clinton, N.E. (Eds.) Cloud-based remote sensing with Googleearth engine: fundamentals and applications. Cham: SpringerInternational Publishing, pp. 279–302. Available from:https://doi.org/10.1007/978-3-031-26588-4_15

Teaching methods

The course is designed with a strong hands-on approach and is based on active student participation. Each student will use their own laptop to carry out practical exercises guided by the instructor. Throughout the course, students will become familiar with various tools for data analysis and visualization (particularly R, Google Earth Engine, and QGIS), with a focus on programming, satellite data processing, and the development of reproducible workflows.

A significant part of the course will be dedicated to identifying a research question, designing an appropriate analytical approach, and writing a short scientific paper, which will constitute an integral part of the final assessment.

Teaching activities will also be enriched by seminars with external experts on specific topics related to forest monitoring, remote sensing, and artificial intelligence.

Given the practical and interactive nature of the course, active participation is required. Students must also complete the mandatory safety training modules 1 and 2 in e-learning format, and attend module 3 on safety in study environments, as detailed in the relevant section of the degree program website.

Assessment methods

The exam consists of the presentation and discussion of a project agreed upon and initiated during the course, and subsequently developed by the student independently. The exam includes:

• the preparation of a short scientific article on a case study chosen by the students;

• a PowerPoint presentation of the article;

• the illustration and discussion of the main codes developed in R and/or Google Earth Engine, with particular attention to the clarity, correctness, and reproducibility of the workflow.

The project may be carried out in small groups; in such cases, the article must clearly indicate each group member’s individual contribution. During the oral exam, each student is expected to present their part of the work, answer questions related to the theoretical content covered during the course, and demonstrate critical understanding of the code presented.

The final grade will be based on the scientific quality of the project, the effectiveness of the presentation and argumentation, the student's proficiency with the tools used, and—above all—their understanding of the underlying theoretical concepts.

Teaching tools

Teaching activities take place in the classroom using students’ personal laptops, which are employed for hands-on exercises, satellite data analysis, and the development of code in programming environments.

The course makes extensive use of open-source and freely accessible tools, including Google Earth Engine, R, QGIS, and other freeware for remote sensing and spatial data analysis.

Lectures are supported by projected presentations and interactive classroom sessions guided by the instructor.

All teaching materials (slides, datasets, code scripts, and supplementary readings) are made available in advance through the Virtuale [https://virtuale.unibo.it] platform.

Office hours

See the website of Saverio Francini