Topic 5f - Practical Guide: How are data processed?

In this next guide Tobias Reinicke takes us through some steps involved in processing data.

The data collected by satellites are processed in different ways to make them useful to researchers. There are technical descriptions of the different levels of data, but you don’t need to know them all to be able to use products available online. For example, different algorithms are applied to data to allow useful images to be identified.

One such algorithm figures out whether there is cloud cover in a given image. It calculates the amount of the image covered by cloud using the brightness in the visible and the thermal channel. Clouds show very brightly in the visible, but so does snow; however, the thermal channels can help distinguish between the two as cloud tops are much colder than snowpack. The cloud screening process sometimes mis-categorises things as cloud, but you can see them on the image and work out what is there. Low clouds are much more difficult to distinguish by temperature, and can also generate cloud shadow which ruins your view of the surface beneath. The presence of cloud is just one of the challenges of using optical data, but processing of the data will warn you how much of an image is likely to be compromised, and you can select from 0% to 100% how much cover you will tolerate.

Most data will have undergone some level of processing before you download it while higher level products like leaf area index, and chlorophyll monitoring will have had a whole suite of models applied to it to make it useful. These processes all must be calibrated against and tested against ground data.

Featured Experts:

  • Tobias Reinicke, independent GIS and EO expert

Featured Data Products and Software Tools:

ESA Snap Toolbox

Don’t forget you can download the video, transcript and take any quizzes available with the links on the right.

Optional Further Reading