pointhi / pysplat Goto Github PK

Doing RF calculations using python and SPLAT

This project aims for a specific usecase, means calculation of signal propagation maps which can then be used as overlay on online maps, and doing some specific processing on those generated tiles.

This means we call SPLAT in a specific way to return the results as georeferenced image. This image the can be split into tiles which are usable as overlay tiles for libraries like Leaflet. Those tiles can furthermore be processed by merging multiple of those maps into a single one. Using those techniques allows us to create maps showing the overal coverage of radio transmitters with as many TX-sites as we want (without beeing limited by SPLAT).

• Calculation of specific SPLAT-Maps
• Split SPLAT-Maps into Tiles
• Merging of those Tiles into a single one

• splat

python dependencies located in ```requirements.txt`

(currently, you have to download and preprocess the required srtm files manually)

./PySplat/pysplat.py example/qth/OE5XGL.qth --out ./example/html/base --srtm ./path/to/srtm/folder

./PySplat/pysplat_split.py ./example/html/base/OE5XGL.ppm ./example/html/rendered/OE5XGL -z 6-12
./PySplat/pysplat_downsample.py ./example/html/rendered/OE5XGL 6 -z 0-5

Please note, there is currently some bug concerning tiles of zoom <=4 (based on my tests), which means the tile are located at the wrong latitude. As temporary fix I wrote the tool python_downsample which simply is able to downsample tiles (merge them and return the next lower zoom level). It's not pretty efficient yet, so it should only be used to create low zoom levels at the moment

Now we can open the leaflet map located in ./example/html/map.html and check out our new rendered RF map overlay.

./PySplat/pysplat_merge.py ./example/html/rendered/OE5*/ ./example/html/rendered_merged/OE5xxx --gpu

Please note, using the --gpu flag activates the OpenCL implementation, which is highly recommended. Even using OpenCL over CPU is more than 10 times faster compared to the native python implementation.