def circle_radius(lat, lon, max_lat, max_lon, min_lat, min_lon):
"""
Compute the maximum distance, in meters, from a (lat, lon) point
to any of the extreme points of a bounding box.
"""
points = numpy.array([
(min_lat, min_lon),
(min_lat, max_lon),
(max_lat, min_lon),
(max_lat, max_lon),
], dtype=numpy.double)
radius = _geocalc.max_distance(lat, lon, points)
return int(round(radius))
def aggregate_position(circles, minimum_accuracy):
"""
Calculate the aggregate position based on a number of circles
(numpy 3-column arrays of lat/lon/radius).
Return the position and an accuracy estimate, but at least
use the minimum_accuracy.
"""
if len(circles) == 1:
return (float(circles[0][0]),
float(circles[0][1]),
max(float(circles[0][2]), minimum_accuracy))
points, _ = numpy.hsplit(circles, [2])
lat, lon = centroid(points)
# Bad approximation. This one takes the maximum distance from
# the centroid to any of the provided circle centers.
# It ignores the radius of those circles.
radius = _geocalc.max_distance(lat, lon, points)
return (lat, lon, max(radius, minimum_accuracy))