The use of remotely sensed data (satellite) with machine learning techniques to classify and analyse both commercial and environmental targets through time. Techniques will focus on both pixel classification and object detection. The primary platform for this will be the Python programming language and/or Google Earth Engine, with collaborative projects managed through git. Students will experience the latest in satellite imagery analysis with a focus on deriving actionable information. No formal prerequisites are required but it is highly advised for students to be comfortable with basic programming concepts and use. Students will gain a set of skills that will allow them to apply remote sensing data to a range of problem sets that they may meet in both academia and industry. Problem sets will be derived from both those set by the instructor and any which interest those taking part in the course. The course has a strong emphasis on teamwork and collaborative problem-solving. The use of AI as a tool to enable rapid implementation is explored and taught as part of this course.
Course Objectives:
(1) To understand the fundamental concepts and theories of machine learning algorithms as applied to satellite data.
(2) To learn how to use satellite data within a programmatic context in order to apply machine learning packages.
(3) To be able to examine a problem, select the appropriate data and interrogative approach, producing an analytical outcome.
(4) To develop skills in the presentation of code and analysis for both specialist and non-specialist audiences.
(5) To gain skills in small team leadership and management for scientific problem solving.
Course Structure: This course consists of scheduled lectures during which students can expect to engage with both the instructor and their peers throughout. Lectures and labs take place in the computer laboratory in which students will learn how to make use of relevant systems/packages, run the analyses taught in class and to interpret the results with the aim of producing actionable information for change. Project work will consist of small teams (2-3) working together to produce an analytical pipeline which they then write up and present to the course in an advocative manner (i.e., make the case).
In addition, relevant readings on topics will be provided for those who would like to gain in-depth knowledge on the topic throughout the semester.