Geospatial Intelligence

Aim of the course:

To learn theory and practice, up to the level of libraries implementing, of artificial intelligence applied to the processing (data preparation, classification, interpolation) of geospatial data.

Topics:

A. Introduction

1. General introductions on AI and Geospatial Applications

2. Fundamentals of statistics and matrix algebra

3. Data Preparation and exploration

4. Classification and clustering: supervised and unsupervised approaches (general concepts)

5. Spatial interpolation and regression: from polynomial to kriging

B. Optimization and Machine learning

1. Single Objective Optimization (GA)

2. Multi-Objective Evolutionary Optimization (NSGA-II)

3. Classification (SVM and Decision Tree)

4. Artificial Neural network (ANN)

5. Deep Learning

6. Ensemble Learning and hyper parameter tuning

7. Geographic Weighted Machine Learning

Learning outcomes:

1. Knowledge and understanding

• To understand and explain the classical methods statistical processing, classification and interpolation
• To understand and explain the optimization methods and machine learning approaches listed in the topic
• To understand and explain the application of them to geospatial data and problems

2. Competences and skills

• To understand and explain the classical methods statistical processing, classification and interpolation
• To understand and explain the optimization methods and machine learning approaches listed in the topic
• To understand and explain the application of them to geospatial data and problems

3. Judgements and evaluations

• In processing of geospatial data, to decide algorithms and parameters suitable for processed data
• In the analysis of the results, to assess their accuracy, to a posteriori evaluate the correctness of the applied methods

Prerequisites:

• General knowledge of standard statistics: 5 ECTS
• General knowledge of remote sensing / earth observation: 5 ECTS
• General knowledge of GIS: 5 ECTS
• General skill in programming: 5 ECTS

Instructors:

Lecturers from the Cadeo Project, affiliated with Politecnico di Milano (Italy); Vietnamese-German University, Hanoi University of Mining and Geology, Hue University (Vietnam), Lund University (Sweden)