Departmental Seminar
As environmental engineers, how can we use remote sensing?
Presenter Utku Berkalp Ünalan
Date 04 November 2026, Wednesday
Time 15:40
Abstract
Remote sensing is a valuable tool for environmental engineers as it aids in monitoring of spatial and temporal environmental conditions over large areas. It does not replace measurements, but can extend them.
Satellites cover the whole planet, often every few days, and much of this data is freely available. Moreover, several online tools now go one step further than just supplying remotely sensed data. They produce information obtained through the data such as maps of soil properties, water quality, or carbon uptake with one click or a single API call.
However, there is a trending question: can we trust these products in our projects? And how do we check?
In this talk, I share what I have learned from using satellite data in real projects. The focus is on practice, not theory. I go through four common mistakes. The first is using a product without checking it against field data. The second is combining datasets with very different pixel sizes, for example a 10 m map with a 1 km map. The third is relying on a map because it looks good. The fourth is using a product outside the region it was built for.
Along the way, I ask some practical questions. When is an existing product enough, and when do we need our own method? What does an index like NDVI actually measure, and what does it solely suggest? Finally, I will mention about new AI-based tools, such as satellite embedding, and where they fit. For this talk, there is no need to have a remote sensing background. The goal is to build an understanding of a simple way to decide when satellite data can support your work, and when it cannot.
Satellites cover the whole planet, often every few days, and much of this data is freely available. Moreover, several online tools now go one step further than just supplying remotely sensed data. They produce information obtained through the data such as maps of soil properties, water quality, or carbon uptake with one click or a single API call.
However, there is a trending question: can we trust these products in our projects? And how do we check?
In this talk, I share what I have learned from using satellite data in real projects. The focus is on practice, not theory. I go through four common mistakes. The first is using a product without checking it against field data. The second is combining datasets with very different pixel sizes, for example a 10 m map with a 1 km map. The third is relying on a map because it looks good. The fourth is using a product outside the region it was built for.
Along the way, I ask some practical questions. When is an existing product enough, and when do we need our own method? What does an index like NDVI actually measure, and what does it solely suggest? Finally, I will mention about new AI-based tools, such as satellite embedding, and where they fit. For this talk, there is no need to have a remote sensing background. The goal is to build an understanding of a simple way to decide when satellite data can support your work, and when it cannot.
Last updated