This appears to be an issue with Cesium 1.10. Swapping out with 1.9 seems to fix the problem.
// Mike
Thanks Mike. We may have broke this in Cesium when we fixed an issue with the sun. Could you submit an issue with instructions on how to reproduce this in your app?
Thanks,
Patrick
The issue for the ghost images is #2811. We’ll try to fix it for 1.11.
Patrick
I looked into the ECHO Documentation, and have understood the structure of the URL, that we would request for the meta-data. I will work on implementing the script for extracting the meta-data, today.
I have been looking into the issue of imagery not aligning, Patrick reported last week. I have observed that every 3rd and 4th imagery of an orbital track are being flipped, when viewed in the 3D View. I re-checked the coordinates in the meta-data.json, they match with the data from NASA’s Website.
Demo URL: http://abhishekvpotnis.me/data-curtains/
I have edited some imagery to, have letters/numbers, so that we could better understand the nature of the flipping. The letters can be observed on the first Day Time Orbital Track of Jan 1, 2015 data and the numbers on the Day Time Orbital Track of Jan 2, 2015.
I compared the First Day Time Orbital Track of Jan 2, 2015 in the Columbus View with the First Day Time Orbital Track on NASA’s Website. I think, it is the orientation of the sections(granules) thats causing the flipping. I have attached the screen-shot of the flipped imagery, for reference.
Hi Abhishek,
Thanks for carefully looking into this. I think the texture coordinates for the geometry are being flipped when the geometry crosses -180/180 longitude. See the attached screenshot.
Could you narrow this down to a self-contained Sandcastle example and submit an issue?
In the meantime, as a workaround, perhaps it is possible to generate two different geometries with two different images when crossing -180/180 longitude?
Thanks,
Patrick
I have filed an issue on Github for the image being flipped as the geometry coordinate crosses -180/180.
This week, I will be working on extracting imagery data from .hdf files by hacking on ccplot.
The ghost image fix is now in master thanks to #2824 from Ed. It will be in Cesium 1.11 on July 1.
Patrick
It’s looking really good guys. Thanks for all the hard work!
Would the HDF processor front-end benefit from embedding Cesium for a quick preview like the Cesium model converter?
Patrick
Hi All,
Here is the weekly report 9 of this project.
Report URL: https://goo.gl/djr1VV
@Patrick, The HDF Processor Front-end is currently being intended for use to the app-maintainers and not the end-users. It will generate the imagery from the .hdf files, that will be displayed in the Wall Geometry as curtain profiles using Cesium. I will discuss on this with mentors in today’s meeting.
In yesterday’s meeting, we decided to process about a month’s worth of Calipso data, generate images, and place them on a server for use by the client. I did about eight days and this is how it is looking:
http://map2.vis.earthdata.nasa.gov/data-curtains
One change that I made was that the directory structure is now by day/time instead of day/integer. Also, instead of working out of the current directory, HDFs are pulled from an “hdf” directory and output goes into an “images” directory which are both in the gitignore. Abhishek, take a look at this pull request and make sure everything looks fine:
https://github.com/nasa-gibs/data-curtains/pull/11
Let me know if the data is good and I’ll process some more imagery.
metadata.json is getting kind of large. We may want to break that up into a listing of available days and fetch a metadata file for each day as needed. Also, we need to upgrade to Cesium 1.11 at some point.
// Mike
Hey Mike,
The pull request and data look good. However the metadata.json currently being generated needs to be modified to be able to be used for by our Cesium based app. Currently for each .hdf file processed, a json object is simply being appended to the metadata.json.
Yes, the metadata.json file will increase in size, so I will divide the metadata.json into separate metadata files for each day as you suggested. I can do this by tomorrow, so you could process the HDFs after that.
I will also upgrade to Cesium 1.11.
Abhishek,
I merged in the PR. It does look like that metadata.json isn’t valid and is missing a comma or so. That should be a minor fix. It would probably be useful to split out the metadata gathering and the image processing into separate scripts. That would make it easy to regenerate the JSON as needed. I’ll go ahead and get some more data tomorrow or early next week.
// Mike
Hi Abhishek,
Have you created an outreach plan for the end of the summer? If not, can you add it to your schedule? If promoted well, I think your work could get a lot of attention, which would be great for you, GSoC, NASA, and Cesium.
Patrick
Hi Patrick,
No, I have not created an outreach plan yet. I have never made one before, so will have to read about it. Once done, I will add it to the wiki on Github, as suggested.
Thanks.
You don’t need to read the entire book, but browse through the Art of Community for ideas:
http://artofcommunityonline.org/Art_of_Community_Second_Edition.pdf
Of course, we can use standard NASA and Cesium channels to also help spread the word about your work.
Patrick
Here is the Weekly Report 10 for this project.
Report URL: https://goo.gl/kQxzBL
@Patrick, Thanks for pointing me to the book. I have jotted down some ideas for the outreach plan of this project in the report. I will discuss them with my mentors and work out a draft of the plan this week.