Data Sets

[r4space]

Discussion thread around selecting data sets we want to process

[hlynch]

High spatial resolution imagery should definitely be in the mix, perhaps starting with WV2 imagery and QB imagery, since the size of these files makes segmentation and classification quite slow. Lower resolution MODIS or Landsat imagery would also be good to have, since the programming might be quite different.

[hlynch]

PS- Anything under license (WV, QB) should not be posted on the public repo. I suggest posting thumbnails and bringing hard copy versions on drives to XSEDE.

[akpetty]

These are the IceBridge DMS images I talked about https://nsidc.org/data/iodms1b
I have Python DMS readers too to read in (and combine) the three spectral bands. They are geolocated too.

[phil-mcdowall]

In the absence of freely sharing example images I've added a doc and csv with some info on some of the sensors we've used in the past at Stony Brook which I thought may be useful to get an idea of what we're dealing with. If anyone would like to add info about datasets they may have available feel free to edit the csv file.

[andypbarrett]

Hi All

I've put QB images on a restricted (password protected) ftp at NSIDC. Email me privately if you want access to these before we meet in MIami.

I'll be getting MODIS imagery for these dates and locations.

I'll post a map of image locations and a list of dates early next week.

Andy