Public Lab Research note


Cartoscope GRN Harvey: Requests for feedback -- not ready for use

by sspatharioti | October 10, 2017 16:50 10 Oct 16:50 | #15017 | #15017

Harvey Response: Oil Sheen Identification using Cartoscope

Responding to a request by the Gulf Restoration Network for identifying oil sheens in aerial images, we created a project on Cartoscope, an open-sourced crowdsourcing tool for analyzing images using various types of tasks such as image labeling.

You can follow the project here.

As this is still a work in progress, please do not publicly share the link until it is ready for use.


6 Comments

Hi, @sspatharioti - this looks great; should people click the link above to try out the system? And are you looking for specific kinds of feedback, like user interface for example?

Thanks!

Is this a question? Click here to post it to the Questions page.

Reply to this comment...


Thank you @warren ! At this stage, we are mainly interested in feedback from GRN about the task, but anyone is free to click on the link and try it out for general feedback as well.

Reply to this comment...


Hi Sophia, I had a chance to show Scott in person last week, and here's what he said:

  • it's great
  • use fullsize images
  • link back to where the images originally came from (in this case, flickr folder) at the end
  • in the survey, the question for performance "flipped" the scale so that "good" is on the left, unlike the scale on the other questions
  • on the heat map, try another color besides green to indicate where the oil is.

Thanks!

Reply to this comment...


This looks great @sspatharioti ! Did you see some of the work done on the Harvey Image sorting project? here's a link to a write up: https://publiclab.org/notes/gilbert/10-13-2017/hurricane-harvey-industrial-impact-image-sorting

Still thinking about coming to Barnraising? Hope you can make it! www.publiclab.org/barnraising

Is this a question? Click here to post it to the Questions page.

Reply to this comment...


@sspatharioti so nice to see this coming together. I took a try at it and I only had wordsmith stuff. Wasn't sure about the term "overflight observer" and also the word "HIT" for submitting at end. I think there still needs to be a photo example for when water appears rippled and shiny, but is not sheen. May have to ask GRN about that since that is my main problem in identifying sheen. I really liked the jellyfish example!

Reply to this comment...


Thanks so much for this project, so great to meet you! I'm excited to try the NOAA images.

here's an example of sheen: https://storms.ngs.noaa.gov/storms/harvey/index.html#17/29.81546/-95.66662

When we tried it in new orleans, leslie, there were others who had the problem of Sun Glint. We may have to consider that in any kind of training...

Reply to this comment...


Login to comment.