satellite imagery in linux

[seandepagnier]

I finally have found a cross-platform way to generate kap files from tile servers in 3 steps. It doesn't require using google earth, google api, or ge2kap.

1)

Enter the bounding coordinates of the area to gather into a mapproxy-seed configuration, eg:

seeds:
  level17:
    caches: [default]
    coverages: [grenada]
    levels:
      from: 17
      to: 17

coverages:
  grenada:
    bbox: [-61.75, 11.95, -61.6616, 12.05]
    srs: "EPSG:4326"

This simple example using a single bounding box, but it is possible to use many bounding boxes, or even polygons and shapefiles to define areas. Perhaps the draw plugin could export boundaries into a format supported by mapproxy.

If you use a higher level like 18 you get more resolution, but each level is 4x more data.

2)

Then you must make a working mapproxy config file. This can work for all the seeds and can define various tile servers for sources. I have an example in my new project:

https://github.com/seandepagnier/mbtiles-nautical-boxes

Run mapproxy seed to obtain the data in a single mbtiles database file

3)

Run mbtiles2kaps.py to convert it to kap files. I have included scripts in this project:

https://github.com/seandepagnier/mbtiles-nautical-tools


As a final note, kap and png are _not_ a good format for satellite imagery.

jpg is much better. mbtiles supports jpg, but so far reading mbtiles charts in opencpn is not yet possible, but in the future will be which will eliminate the last step and reduce network and disk requirements. (it is already possible to build jpg mbtiles databases with mapproxy if you have limited bandwidth)

[BlackSea]

thanx for your great job. I'm using another way for create a KAP Charts.

https://github.com/wellenvogel/mobac_convert

[seandepagnier]

I am playing with mobac. Apparently it can save as mbtiles also.

What I don't understand with the script is why it claims it can't match the ellipsoid model, when in fact it could if it embedded the wpx and wpy parameters into the kap.

In any case, the real issue is gathering the required areas and levels to build a complete nautical chart. Otherwise the data required is huge, covering areas on both land and sea (we only care about coastlines)

What did you do about this? Do you just manually select boxes and levels, or do you have a better system? So far I am building a database of bounding boxes at various levels with 20,000 boxes of various zoom levels. The highest levels are small areas around ports or narrow passes.


From here I should make a tool to filter them based on areas so you can for example, only grab the med or australia or what you like.