Issue with tutorials/detecting-changes-in-sentinel-1-imagery-pt-2/index.ipynb

[guiattard]

Hello everyone,

I've been through the the python tutorial series focusing on sentinel-1. Thanks to the contributors and @mortcanty for this detailed and very interesting work. I would only have one recommandation to make easier the adaptation of the tutorial (part 2) to a user-specific area of interest (and different from the site proposed in the tutorial).

At the beginnig of the tutorial, the area of interest is defined as a polygon and just after that, the S1 image collection is filtered accordingly. In this procedure, the relative orbit number of interest (equal to 15) is specified but seems to come out of nowhere (see below).

im_coll = (ee.ImageCollection('COPERNICUS/S1_GRD_FLOAT')
                .filterBounds(aoi)
                .filterDate(ee.Date('2020-08-01'),ee.Date('2020-08-31'))
                .filter(ee.Filter.eq('orbitProperties_pass', 'ASCENDING'))
                .filter(ee.Filter.eq('relativeOrbitNumber_start', 15))
                .sort('system:time_start'))

To people not very familiar with the S1 product, it might be confusing and hard to understand how to adapt the number to the actual area of interest. Hence, I would recommand to add a line or a link to give an indication in that sence.

For example, I think that a reference to the Acquisition Plan Viewer might be helpfull to find relative orbit number associated to satellites, period and locations of interest.

Otherwise, the S1 series is great =)

Best regards,

[jdbcode]

Good suggestion. I wonder if it is necessary? It seems not. There is a date and aoi filter, which seems sufficient. I'll try without the relativeOrbitNumber_start and see if a different set of images are provided. In any case, it seems unnecessary for the purpose of the tutorial.

[jdbcode]

Best to keep the filter for relative orbit. I added a note where they can find relative orbit, see #768 (thanks for the suggestion about Acquisition Plan Viewer!)