Changeset 9152 for src/django_gheat

Timestamp:

May 5, 2011, 11:30:48 AM (14 years ago)

Author:

rick

Message:

Signal is not normalized so make it so. Also the starting transparancy should
be related to the initial entry field.

While we are here, explaining in-depth the issues we are facing when plotting
the data.

File:

• src/django_gheat/website/tile.py (modified) (2 diffs)

src/django_gheat/website/tile.py

@@ -197 +197 @@
   # copy the raw reported values
   MAX_SIGNAL = 50
+  # XXX: The radius relates to the zoom-level we are in, and should represent
+  # a fixed distance, given the scale. Assume signal/distance to be lineair
+  # such that signal 100% = 100m and 1% = 1m.
+  # 
+  # XXX: The relation is not lineair but from a more logeritmic scape, as we
+  # are dealing with radio signals
+  #
+  MAX_RANGE = 100
   
   def dif(x,y):
@@ -211 +219 @@
     log.info(meting.accespoint.ssid, meting.latitude, meting.longitude, xcoord, ycoord)
 
-    # XXX: The radius relates to the zoom-level we are in, and should represent
-    # a fixed distance, given the scale. Assume signal/distance to be lineair
-    # such that signal 100% = 100m and 1% = 1m.
+    # TODO: Please note that this 'logic' technically does apply to WiFi signals,
+    # if you are plotting from the 'source'. When plotting 'measurement' data you
+    # get different patterns and properly need to start looking at techniques like:
+    # Multilateration,Triangulation or Trilateration to recieve 'source' points.
     # 
-    # XXX: The relation is not lineair but from a more logeritmic scape, as we
-    # are dealing with radio signals
+    # Also you can treat all points as seperate and use techniques like
+    # Multivariate interpolation to make the graphs. A nice overview at:
+    #     http://en.wikipedia.org/wiki/Multivariate_interpolation
     #
-    # TODO: Please note that this 'logic' technically does any apply to WiFi signals,
-    # if you are plotting from the 'source'. With measured data you get
-    # different patterns.
-    #
-    im.add_circle((xcoord,ycoord),float(meting.signaal) / meters_per_pixel,(255,0,0),MAX_SIGNAL - meting.signaal)
+    # One very intersting one to look at will be Inverse distance weighting
+    # with examples like this: 
+    # http://stackoverflow.com/questions/3104781/inverse-distance-weighted-idw-interpolation-with-python
+    signal_normalized = MAX_RANGE - (MAX_SIGNAL - meting.signaal)
+    im.add_circle((xcoord,ycoord),float(signal_normalized) / meters_per_pixel,(255,0,0), MAX_SIGNAL - meting.signaal)
   
   log.info("BoundingBox NW: %s" % nw_deg)