Field surveys face the compromise between level of detail, extend and temporal repetition. One can study selected processes in detail on a very limited number of observation sites or focus on a landscape survey using generalized observations. The restrictions loosen when linking survey information to area wide remote sensing observations and the modelling of raster maps as basis for scientific analysis.
Remote sensing has a history of more than 150 years. It is well established and cost-effective but must be applied with caution to gain robust results. The advent of artificial intelligence has opened a new chapter in data mining and information retrieval. Profound skills in data handling, machine learning, team-based software development, workflow documentation and presentation and discussion of result are required.
Intended learning outcomes
At the end of this course you should be able
- to research remote sensing methods and design workflows suitable to solve common remote sensing problems,
- to deploy your remote sensing workflows using geo-information science tools, R scripts and collaborative code management platforms for task management and issue tracking,
- to critically evaluate your remote sensing analysis,
- to document and communicate your remote sensing workflow and analysis results,
- to reflect your project workflow for potential improvements.
The course is linked to the parallel Geo Information Systems and build upon the parallel Data Analysis courses. It is intended as a blended learning module in our study program although the provided introductions, explanations and examples might be useful for self-study, too. Data and examples in this course focus on Marburg Open Forest - the open research and learning forest of Marburg University - and are related to the LOEWE Priority Programm Nature 4.0.
|Remote sensing basics|
|1||Remote Sensing in environmental research and nature conservation||Get to know basic remote sensing principles, the Nature 4.0 project approach and the study area of this course|
|2||Understanding remote sensing data and how to handle it||Reflect the guiding principals of remote sensing and think about general data precprocessing|
|Project 1: A picture is worth a thousand words||Baseline image processing|
|3||Problem: Comprehensive image collection for tree delineation, species prediction, and heterogeneity mapping||Decompose the problem, identify research tasks and sketch a project worklfow|
|4||Spotlight: LiDAR data processing||Dealing with storage expensive LiDAR data|
|5||Spotlight: Artificial images||Calculation of artificial images|
|6||Spotlight: Communicating research||Preparation for writing a short article|
|7||Compilation of area wide dataset||Transfer data to parallel GIS course project|
|8||Project 1 deadline||Provide your preliminary project results in the form of introduction and method chapters to your peers|
|9||Peer feedback on project presentation||Evaluate the work of your peers, reflect your own workflow and discuss potential improvements|
|Project 2: Seeing the tree species for the wood||Machine-learning-based prediction of species|
|10||Problem: Predict tree species in space||Decompose the problem, identify research tasks and sketch a project worklfow|
|11||Spotlight: Land cover prediction and machine learning models||Deal with machine learning models and error estimation strategies for spatial data|
|12||Continue your project|
|13||Project 2 deadline||Provide your project results as final article to your peers|
|14||Feedback and goodbye||Get feedback from your peers and instructors, tell us how you self-assess your skills and happy holidays|
The graded course certificate will be based on a portfolio. The individual portfolio items are defined in the course assignments along with information if they will be marked or not. The marked portfolio item will be an article following the type “Practical Tools” (but up to 2,000 words) of Methods in Ecology and Evolution.