I was pleasantly surprised when I discovered the data from the European Satellite Mission: Sentinel 2. This satellite imagery is sure at low resolution compared to Google data or bing -> only 10m/pixel, but it has several advantages:
- First, not least is that these data are clearly free to access and use (Creative Commons).
- Second, these data are very homogeneous in color and at good quality.
- Third, all data are represented and therefore the sea is not cut to the bottom of the odds and all the shallow waters are visible even very off the coast.
- Finally, the last advantage that I I discovered later that thanks to their redundancy (about 1 pass every 5 days) it is possible to clean the clouds quite efficiently. The principle in summary consists in superimposing different pass as layer in photoshop, which allows to clear cloud on a layer and missing part will be replaced by the layer of the bottom, etc. in the end, if you spend a little time, except in a very cloudy region you can get an almost cloudless picture. Later, I discovered that there were also utilities with algorithm to do the work alone but most of the time it is less precise work and it feels a lot in the final rendering. so reserve it therefore only to treat very large areas.
In order to try to the use of these sentinel 2 data, I decided to treat some of the Caribbean, the goal was to connect Florida to my scene in Martinique to allow a pleasant transit to the eye between two area
After a lot of tests and manual work I am rather satisfied with the result. I completed the Elevation part with the ALOS World 3D 30m data also in open-data (https://www.eorc.jaxa.jp/ALOS/en/aw3d30/index.htm)
Attached are some animated gif with comparative in game screenshot :
Basic Photo Cover / Base Mesh <-> Cover Photo Sentinel 2 / Mesh ALOS
If it interests some, I would to try to pack all this and put it available on flight-sim.org
To improve all this it would be interesting to know if it is not possible to use these basic data (resolution 10m) and to couple them with higher resolution aerial data (1 or 2 meter / p) in order to be able to take advantage of the color homogeneity of sentinel2 data and the accuracy of the high-resolution aerial data ... I know that there is this type of processing for satellite data in order to couple satellite low resolution colorimetric data with a panchromatic band which stores only high resolution data this is called pansharpening but after some tests I have not yet managed to transpose this for sentinel data 2. To be continued ...