Great-circle distance computational forumlas in C++

geo_distance.cpp

/*
 * Great-circle distance computational forumlas
 *
 * https://en.wikipedia.org/wiki/Great-circle_distance
 */

#ifndef _USE_MATH_DEFINES
#define _USE_MATH_DEFINES
#endif
#include <cmath>

#ifndef M_PI
#define M_PI    3.1415926535897932384626433832795
#endif

using namespace std;

static const double earth_radius_km = 6371.0;

double deg2rad(double deg)
{
    return (deg * M_PI / 180.0);
}

double haversine_distance(double latitude1, double longitude1, double latitude2,
                          double longitude2)
{
    double lat1 = deg2rad(latitude1);
    double lon1 = deg2rad(longitude1);
    double lat2 = deg2rad(latitude2);
    double lon2 = deg2rad(longitude2);

    double d_lat = abs(lat1 - lat2);
    double d_lon = abs(lon1 - lon2);

    double a = pow(sin(d_lat / 2), 2) + cos(lat1) * cos(lat2) * pow(sin(d_lon / 2), 2);

    //double d_sigma = 2 * atan2(sqrt(a), sqrt(1 - a));
    double d_sigma = 2 * asin(sqrt(a));

    return earth_radius_km * d_sigma;
}

double vincenty_distance(double latitude1, double longitude1, double latitude2,
                         double longitude2)
{
    double lat1 = deg2rad(latitude1);
    double lon1 = deg2rad(longitude1);
    double lat2 = deg2rad(latitude2);
    double lon2 = deg2rad(longitude2);

    double d_lon = abs(lon1 - lon2);

    // Numerator
    double a = pow(cos(lat2) * sin(d_lon), 2);

    double b = cos(lat1) * sin(lat2);
    double c = sin(lat1) * cos(lat2) * cos(d_lon);
    double d = pow(b - c, 2);

    double e = sqrt(a + d);

    // Denominator
    double f = sin(lat1) * sin(lat2);
    double g = cos(lat1) * cos(lat2) * cos(d_lon);

    double h = f + g;

    double d_sigma = atan2(e, h);

    return earth_radius_km * d_sigma;
}

in code

double d_lat = abs(lat1 - lat2);
double d_lon = abs(lon1 - lon2);

the function

abs

should be changed to

fabs

, because abs returns the absolute value of an integer.

the function
abs
should be changed to
fabs
, because abs returns the absolute value of an integer.

No, it references std::abs() which is overloaded for all primitive types.