package Geo::Approx; use strict; use Carp; use Math::Trig; use vars qw ( $VERSION ); $^W = 1; $VERSION = 0.8; my %mask; BEGIN { for(my $i=0;$i<=32;$i++){ $mask{$i} = pack("B32",('0'x(32-$i) . '1'x$i)); } } sub new { my ($caller,$precision) = @_; my $class = ref($caller) || $caller; if(defined $precision){ unless(($precision =~ /^(\d+)$/)&&($1>=0)&&($1<=32)) { croak("optional argument to Geo::Aprox constructor must be a precision between zero and 32"); } } else { $precision = 32; } bless \$precision,$class; } sub latlon2int { my $prec = ${$_[0]}; my $result = _intint2int(_lat2int($_[1]),_lon2int($_[2])); $result = _setPrecision($prec,$result); return $result; } sub int2latlon { my $mask = \$_[0]; my ($lat_int,$lon_int) = _int2intint($_[1]); my $lat = _int2lat($lat_int); my $lon = _int2lon($lon_int); return ($lat,$lon); } sub _setPrecision { my ($prec,$num) = @_; return unpack("N",pack("N",$num)&$mask{$prec}); } sub _int2lat # 0 to 65535 { return rad2deg(asin(($_[0]+0.5)/32768-1.0)); } sub _lat2int # -90 to 90 inclusive { my $result = int((sin(deg2rad($_[0]))+1.0)*32768); $result = 65535 if ($result==65536); # special case for lat==90.000 return int($result); } # 0 to 65535 sub _lon2int # -180 to 180 inclusive { my $result = int(($_[0]/360 + 0.5)*65536); $result %= 65536; # wrap-around return $result; } # 0 to 65535 inclusive sub _int2lon # 0 to 65535 inclusive { return ($_[0]/65536 - 0.5)*360; } # -180 to <180 sub _intint2int # two 16-bit numbers { my $bina = substr(unpack("B32",pack("N",$_[0])),-16); my $binb = substr(unpack("B32",pack("N",$_[1])),-16); my $bbiinnab; for(my $i=0;$i<16;$i++){ $bbiinnab .= substr($bina,$i,1); $bbiinnab .= substr($binb,$i,1); } my $dec = unpack("V",pack("B32",$bbiinnab)); return $dec; } # one 32-bit number sub _int2intint # one 32-bit number { my $bbiinnab = substr(unpack("B32",pack("V",$_[0])),-32); my ($bina,$binb) = ('0000000000000000','0000000000000000'); for(my $i=0;$i<32;$i+=2){ $bina .= substr($bbiinnab,$i,1); $binb .= substr($bbiinnab,$i+1,1); } my $deca = unpack("N",pack("B32",$bina)); my $decb = unpack("N",pack("B32",$binb)); return ($deca,$decb); } # two 16-bit numbers 1; __END__ =head1 NAME Geo::Approx - represents an approximate global position by a single number =head1 SYNOPSIS use Geo::Approx; my $ga = Geo::Approx($precision); my $pos = $ga->latlon2int($lat,$lon); my ($approx_lat,$approx_lon) = $ga->int2latlon($pos); =head1 DESCRIPTION It is sometimes useful to condense the information present in a latitude and longitude into a single number (for example, when storing the position within a database). This module provides methods for this conversion. By default, the precision of the position is set at 32 bits (roughly Assuming the surface area of the earth is 5.1 x 10^8 sq km, the area represented by the single number at each precision is as follows: 0.12 sq km at 32 bits 0.24 sq km at 31 bits 0.47 sq km at 30 bits 0.95 sq km at 29 bits and so on (5.1E8 / 2^precision) These areas are constant across all latitudes and longitudes. Thus, if you want to be fairly precise about positions, you can store fairly large numbers, whereas if you want to be fairly imprecise, you can use small numbers. =head1 CONSTRUCTOR The constructor takes one optional argument - a number between zero and 32 indicating precision. my $ga = Geo::Aprox->new(24); By default, the precision is 32 bits. =head1 OBJECT METHODS =over 4 =item $pos = $ga-Elatlon2int($lat,$lon); Converts a latitude and longitude to an integer integer representing that position. Latitude must be between -90 and 90. Longitude must be between -180 and 180. =item ($lat,$lon) = $ga-Eint2latlon($pos); Converts a position represented by an integer into a latitude and longitude. =back =head1 SEE ALSO L - the post that raised my interest. =head1 COPYRIGHT Copyright (C) 2002,2003 Nigel Wetters. All Rights Reserved. NO WARRANTY. This module is free software; you can redistribute it and/or modify it under the same terms as Perl itself. =cut