About the project

Motivation

Processing of time series of remote sensing data is a challenging goal of both today and future research. Global Earth observation programmes provide extensive image archives dated back to several years (e.g. Sentinel) or even decades (e.g. Landsat or Spot). In a larger scale, aerial images, LiDAR and photogrammetric point clouds are acquired on the national level in two or three years cycles. Thanks to the increasingly supported and accepted open data policy, free access to these data has become a reality in several EU countries. Specific localities are monitored on regular (weekly, monthly, yearly) basis for research purposes tackling e.g. environmental issues such as resistance of plants to droughts, invasive species encroachment, melting of mountainous snow and glaciers, desertification, decrease of biodiversity, deforestation or urbanization.

Processing of time series of remote sensing data demands knowledge of handling large data sets (big data) and advanced statistical and machine learning methods. Nowadays, solutions offering access to opened archives of satellite images and cloud processing of the images exist (e.g. Google Earth Engine). To benefit from such services and to be able to document the achieved quality of the results, a user must have a deep knowledge about the used data and applied processing methods. On the other hand, these services are not suited for data from other sources than specific sensor(s) (e.g. Landsat or Sentinel). Thus, processing methods of the time series of remote sensing data of different time and spatial scales, the combination of heterogeneous and multimodal data sources and importantly the topics dealing with reliability and quality of the results obtained by these methods are becoming a key part of remote sensing and/or geo(infor)matics curricula.

Currently, the most common practical exercises are based on datasets acquired on single dates to present principles of remote sensing and field data collection, specific methods of extraction of information and validation of the results. Time series processing from beginning to the end of analysis are less frequent or optional in geoinformatics study programs. Our initiative aims at increasing our students’ digital literacy in time series analysis in remote sensing, which comprises the improvement in methodological and practical data handling skills. At the same time the students shall gain the skill for critical reflection and communication of complex data processing tasks, enabling them for transformative and interdisciplinary research missions.

Course structure

The e-learning course will consist of four modules, where the first one will provide a general overview on methods for remote sending data time series analysis and the other three will focus on specific processing steps connected to different types of data:

• Module 1: Methods of Time Series Analysis in Remote Sensing
• Module 2: Satellite Multispectral Images Time Series Analysis
• Module 3: 3D/4D Geographic Point Cloud Time Series Analysis
• Module 4: Airborne Imaging and Laboratory Spectroscopy Time Series Analysis

Each module will include three case studies with examples of datasets. Some of the case studies will be common to two all even all three modules in order to show how the different data sources are complementary with respect to extracted information. All used data sets and methodologies are connected to the past or ongoing research projects of the partners and are connected to the applications of remote sensing in environmental studies, specifically monitoring of vegetation (forest disturbances like windstorms, droughts, bark beetle attacks or change of relict arctic-alpine tundra vegetation as an possible indicator of climate change) and monitoring of geomorphological features (landslides, mountainous/sea glaciers, sandstone rocks).

Target groups

The primary target group are MSc and PhD students of geoinformatics and related geographical fields who specialize in remote sensing for monitoring Earth surface dynamic and changes. The secondary target group are MSc and PhD students in the fields related to environmental studies, ecology, geology and further potential users dealing with applications of remote sensing such as practitioners of national environmental and conservation agencies.

The project focuses on transferring the knowledge from research projects to education. All project partners focus on research in time series analysis but in different fields of applications (vegetation monitoring, monitoring of glaciers) and data sources (satellite and airborne images, LiDAR data) but all are connected to studying of environmental phenomena influence by human activities (climate change, air pollution) . Carrying out a transnational project brings an opportunity to create a comprehensive course content that shows similarities in applied methodologies as well as deep insight in variety of applications that would not be possible without synergy of all partners.