Querying the OSM data for peaks (or other features) which are nearby a GPX track

The Overpass API around filter queries for elements within a given radius (in meters) along a linestring when given a list of lat/lon pairs:

(around:radius,latitude_1,longitude_1,latitude_2,longitude_2,...,latitude_n,longitude_n)

For example (in Overpass Turbo):

node[natural=peak](around:500,{{latlons}});
(._;>;);
out meta;

{{latlons=47.492131,10.072515,47.497647,10.084685,47.497754,10.089965,47.502135,10.097873,47.502383,10.104472,47.504251,10.106848,47.504002,10.118661,47.510921,10.13158,47.510898,10.146653,47.517277,10.160911,47.516576,10.164531,47.520637,10.181296,47.528658,10.1929,47.531326,10.202335,47.530947,10.207266,47.537454,10.217361}}

There might be better ways, but here is a one-liner for the Linux/Ubuntu command-line to convert a GPX track to the expected linestring format:

ogr2ogr -f CSV /vsistdout/ brouter.gpx tracks -lco GEOMETRY=AS_WKT -simplify 0.001 | \
tail -n +2 | \
sed -e 's/.* (\+\([^)]*\))\+".*/\1/' -e 's/\([0-9\.]*\) \([0-9\.]*\),\?/\2,\1,/g' -e 's/,$//'

• example GPX track created using an online router
• uses GDAL ogr2ogr converter (requires gdal-bin package)
  with GPX and CSV with WKT (Well-known Text) formats, as this comes close to expected format
• outputs to console with "/vsistdout/"
• "simplify" reduces number of nodes, distance tolerance in degrees, 0.00001° about 1 meter, depending on latitude
• "tail" skips csv header
• "sed" extracts coordinates, flips lon/lat and uses all comma, removes trailing comma