PROBA-V Super Resolution

Nov. 1, 2018, 4 a.m. UTC

June 1, 2019, 4 a.m. UTC

Timeline

The competition is over.

Online meetup

avatar

marcusm

Dear community,

as already hinted, we will have an online meetup for all participants of the PROBA-V Super Resolution competition.

The date is: Friday, 8th of March, 16:00 MET.
The virtual meeting room will be https://talky.io/kelvins

Talky.io is a webapp that requires no registration: just click the link and make sure your microphone/video is ready. There will also be a text chat for the shy ones. Talky allows for up to 15 participants (which is how much I would anticipate at the moment).

The idea of the meetup is for the participants to exchange experiences, strategies and struggles. We will also make some announcements regarding the competition so far and what is still about to come. There will be a chance to ask questions and give feedback as well.

I am very curious to hear about your experiences and hope to see you soon!

p.s.: In the event of unforeseen issues (i.e. service down), we will keep you posted in this thread.

avatar

marcusm

Dear community,

it was a pleasure meeting you last Friday! As a follow-up I would like to highlight a few points that were raised during the meetup and summarize some announcement for participants who were unable to attend.

1. The PROBA-V Products User Manual provides some additional insights on how the satellite data was processed. You can find the manual here: (http://proba-v.vgt.vito.be/sites/proba-v.vgt.vito.be/files/products_user_manual.pdf). In particular, Section 2.2.3 deals with the cloud and cloud shadow detection. The product level of the data for this competition is L2A.

2. As highlighted in "Scoring", the current leaderboard reflects scores taken from a random subset of the test images. When the competition ends, we will reevaluate each team's best submission with the complete set. I did this already now and while the values of the score differs for obvious reasons, the ranking of the teams above the baseline was not changed by extending the test-set. Whether this also holds for the final ranking needs to be seen of course.

3. Knowing the "point spread function" of a measurement instrument is not easy, but I am currently investigating what we can share about this. I will put up a new thread as soon as I know more.

4. We plan to publish a selection of the best solution strategies from participants of this competition in a scientific venue. For this purpose, we plan to invite the three top-performing teams to select a representative to meet us at the European Space Agency at ESTEC, the Netherlands, for a one day workshop, where they would present their solution strategy, and meet and discuss these solutions with PROBA-V engineers and remote sensing experts. The goal of the workshop is to negotiate and develop next steps for a joint publication and discuss potential implementation / application paths. We will announce more details at a later point.

Let me thank you once more for putting your effort into this competition. I am looking forward to some more action on the leaderboard again and wish you all great success during the second half!

Please log in to leave a comment.