Calculate distance between two latitude-longitude points? (Haversine formula)

This link might be helpful to you, as it details the use of the Haversine formula to calculate the distance.

Excerpt:

This script [in Javascript] calculates great-circle distances between the two points – that is, the shortest distance over the earth’s surface – using the ‘Haversine’ formula.

function getDistanceFromLatLonInKm(lat1,lon1,lat2,lon2) {
  var R = 6371; // Radius of the earth in km
  var dLat = deg2rad(lat2-lat1);  // deg2rad below
  var dLon = deg2rad(lon2-lon1); 
  var a = 
    Math.sin(dLat/2) * Math.sin(dLat/2) +
    Math.cos(deg2rad(lat1)) * Math.cos(deg2rad(lat2)) * 
    Math.sin(dLon/2) * Math.sin(dLon/2)
    ; 
  var c = 2 * Math.atan2(Math.sqrt(a), Math.sqrt(1-a)); 
  var d = R * c; // Distance in km
  return d;
}

function deg2rad(deg) {
  return deg * (Math.PI/180)
}

Here is a C# Implementation:

static class DistanceAlgorithm
{
    const double PIx = 3.141592653589793;
    const double RADIUS = 6378.16;

    /// <summary>
    /// Convert degrees to Radians
    /// </summary>
    /// <param name="x">Degrees</param>
    /// <returns>The equivalent in radians</returns>
    public static double Radians(double x)
    {
        return x * PIx / 180;
    }

    /// <summary>
    /// Calculate the distance between two places.
    /// </summary>
    /// <param name="lon1"></param>
    /// <param name="lat1"></param>
    /// <param name="lon2"></param>
    /// <param name="lat2"></param>
    /// <returns></returns>
    public static double DistanceBetweenPlaces(
        double lon1,
        double lat1,
        double lon2,
        double lat2)
    {
        double dlon = Radians(lon2 - lon1);
        double dlat = Radians(lat2 - lat1);

        double a = (Math.Sin(dlat / 2) * Math.Sin(dlat / 2)) + Math.Cos(Radians(lat1)) * Math.Cos(Radians(lat2)) * (Math.Sin(dlon / 2) * Math.Sin(dlon / 2));
        double angle = 2 * Math.Atan2(Math.Sqrt(a), Math.Sqrt(1 - a));
        return angle * RADIUS;
    }

}

I post here my working example.

List all points in table having distance between a designated point (we use a random point - lat:45.20327, long:23.7806) less than 50 KM, with latitude & longitude, in MySQL (the table fields are coord_lat and coord_long):

List all having DISTANCE<50, in Kilometres (considered Earth radius 6371 KM):

SELECT denumire, (6371 * acos( cos( radians(45.20327) ) * cos( radians( coord_lat ) ) * cos( radians( 23.7806 ) - radians(coord_long) ) + sin( radians(45.20327) ) * sin( radians(coord_lat) ) )) AS distanta 
FROM obiective 
WHERE coord_lat<>'' 
    AND coord_long<>'' 
HAVING distanta<50 
ORDER BY distanta desc

The above example was tested in MySQL 5.0.95 and 5.5.16 (Linux).

The haversine is definitely a good formula for probably most cases, other answers already include it so I am not going to take the space. But it is important to note that no matter what formula is used (yes not just one). Because of the huge range of accuracy possible as well as the computation time required. The choice of formula requires a bit more thought than a simple no brainer answer.

This posting from a person at nasa, is the best one I found at discussing the options

http://www.cs.nyu.edu/visual/home/proj/tiger/gisfaq.html

For example, if you are just sorting rows by distance in a 100 miles radius. The flat earth formula will be much faster than the haversine.

HalfPi = 1.5707963;
R = 3956; /* the radius gives you the measurement unit*/

a = HalfPi - latoriginrad;
b = HalfPi - latdestrad;
u = a * a + b * b;
v = - 2 * a * b * cos(longdestrad - longoriginrad);
c = sqrt(abs(u + v));
return R * c;

Notice there is just one cosine and one square root. Vs 9 of them on the Haversine formula.

Here is the SQL Implementation to calculate the distance in km,

SELECT UserId, ( 3959 * acos( cos( radians( your latitude here ) ) * cos( radians(latitude) ) * 
cos( radians(longitude) - radians( your longitude here ) ) + sin( radians( your latitude here ) ) * 
sin( radians(latitude) ) ) ) AS distance FROM user HAVING
distance < 5  ORDER BY distance LIMIT 0 , 5;

For further details in the implementation by programming langugage, you can just go through the php script given here

Java implementation in according Haversine formula

double calculateDistance(double latPoint1, double lngPoint1, 
                         double latPoint2, double lngPoint2) {
    if(latPoint1 == latPoint2 && lngPoint1 == lngPoint2) {
        return 0d;
    }

    final double EARTH_RADIUS = 6371.0; //km value;

    //converting to radians
    latPoint1 = Math.toRadians(latPoint1);
    lngPoint1 = Math.toRadians(lngPoint1);
    latPoint2 = Math.toRadians(latPoint2);
    lngPoint2 = Math.toRadians(lngPoint2);

    double distance = Math.pow(Math.sin((latPoint2 - latPoint1) / 2.0), 2) 
            + Math.cos(latPoint1) * Math.cos(latPoint2)
            * Math.pow(Math.sin((lngPoint2 - lngPoint1) / 2.0), 2);
    distance = 2.0 * EARTH_RADIUS * Math.asin(Math.sqrt(distance));

    return distance; //km value
}

To calculate the distance between two points on a sphere you need to do the Great Circle calculation.

There are a number of C/C++ libraries to help with map projection at MapTools if you need to reproject your distances to a flat surface. To do this you will need the projection string of the various coordinate systems.

You may also find MapWindow a useful tool to visualise the points. Also as its open source its a useful guide to how to use the proj.dll library, which appears to be the core open source projection library.

For those looking for an Excel formula based on WGS-84 & GRS-80 standards:

=ACOS(COS(RADIANS(90-Lat1))*COS(RADIANS(90-Lat2))+SIN(RADIANS(90-Lat1))*SIN(RADIANS(90-Lat2))*COS(RADIANS(Long1-Long2)))*6371

Source

I made a custom function in R to calculate haversine distance(km) between two spatial points using functions available in R base package.

custom_hav_dist <- function(lat1, lon1, lat2, lon2) {
R <- 6371
Radian_factor <- 0.0174533
lat_1 <- (90-lat1)*Radian_factor
lat_2 <- (90-lat2)*Radian_factor
diff_long <-(lon1-lon2)*Radian_factor

distance_in_km <- 6371*acos((cos(lat_1)*cos(lat_2))+ 
                 (sin(lat_1)*sin(lat_2)*cos(diff_long)))
rm(lat1, lon1, lat2, lon2)
return(distance_in_km)
}

Sample output

custom_hav_dist(50.31,19.08,54.14,19.39)
[1] 426.3987

PS: To calculate distances in miles, substitute R in function (6371) with 3958.756 (and for nautical miles, use 3440.065).

function getDistanceFromLatLonInKm(lat1,lon1,lat2,lon2,units) {
  var R = 6371; // Radius of the earth in km
  var dLat = deg2rad(lat2-lat1);  // deg2rad below
  var dLon = deg2rad(lon2-lon1); 
  var a = 
    Math.sin(dLat/2) * Math.sin(dLat/2) +
    Math.cos(deg2rad(lat1)) * Math.cos(deg2rad(lat2)) * 
    Math.sin(dLon/2) * Math.sin(dLon/2)
    ; 
  var c = 2 * Math.atan2(Math.sqrt(a), Math.sqrt(1-a)); 
  var d = R * c; 
  var miles = d / 1.609344; 

if ( units == 'km' ) {  
return d; 
 } else {
return miles;
}}

Chuck's solution, valid for miles also.

Here's another converted to Ruby code:

include Math
#Note: from/to = [lat, long]

def get_distance_in_km(from, to)
  radians = lambda { |deg| deg * Math.PI / 180 }
  radius = 6371 # Radius of the earth in kilometer
  dLat = radians[to[0]-from[0]]
  dLon = radians[to[1]-from[1]]

  cosines_product = Math.sin(dLat/2) * Math.sin(dLat/2) + Math.cos(radians[from[0]]) * Math.cos(radians[to[1]]) * Math.sin(dLon/2) * Math.sin(dLon/2)

  c = 2 * Math.atan2(Math.sqrt(cosines_product), Math.sqrt(1-cosines_product)) 
  return radius * c # Distance in kilometer
end

One of the main challenges to calculating distances - especially large ones - is accounting for the curvature of the Earth. If only the Earth were flat, calculating the distance between two points would be as simple as for that of a straight line! The Haversine formula includes a constant (it's the R variable below) that represents the radius of the Earth. Depending on whether you are measuring in miles or kilometers, it would equal 3956 mi or 6367 km respectively.

The basic formula is:

dlon = lon2 - lon1
dlat = lat2 - lat1
a = (sin(dlat/2))^2 + cos(lat1) * cos(lat2) * (sin(dlon/2))^2
c = 2 * atan2( sqrt(a), sqrt(1-a) )
distance = R * c (where R is the radius of the Earth)
 
R = 6367 km OR 3956 mi

     lat1, lon1: The Latitude and Longitude of point 1 (in decimal degrees)
     lat2, lon2: The Latitude and Longitude of point 2 (in decimal degrees)
     unit: The unit of measurement in which to calculate the results where:
     'M' is statute miles (default)
     'K' is kilometers
     'N' is nautical miles

Sample

function distance(lat1, lon1, lat2, lon2, unit) {
    try {
        var radlat1 = Math.PI * lat1 / 180
        var radlat2 = Math.PI * lat2 / 180
        var theta = lon1 - lon2
        var radtheta = Math.PI * theta / 180
        var dist = Math.sin(radlat1) * Math.sin(radlat2) + Math.cos(radlat1) * Math.cos(radlat2) * Math.cos(radtheta);
        dist = Math.acos(dist)
        dist = dist * 180 / Math.PI
        dist = dist * 60 * 1.1515
        if (unit == "K") {
            dist = dist * 1.609344
        }
        if (unit == "N") {
            dist = dist * 0.8684
        }
        return dist
    } catch (err) {
        console.log(err);
    }
}

As this is the most popular discussion of the topic I'll add my experience from late 2019-early 2020 here. To add to the existing answers - my focus was to find an accurate AND fast (i.e. vectorized) solution.

Let's start with what is mostly used by answers here - the Haversine approach. It is trivial to vectorize, see example in python below:

def haversine(lat1, lon1, lat2, lon2):
    """
    Calculate the great circle distance between two points
    on the earth (specified in decimal degrees)

    All args must be of equal length.
    Distances are in meters.
    
    Ref:
    https://stackoverflow.com/questions/29545704/fast-haversine-approximation-python-pandas
    https://ipython.readthedocs.io/en/stable/interactive/magics.html
    """
    Radius = 6.371e6
    lon1, lat1, lon2, lat2 = map(np.radians, [lon1, lat1, lon2, lat2])

    dlon = lon2 - lon1
    dlat = lat2 - lat1

    a = np.sin(dlat/2.0)**2 + np.cos(lat1) * np.cos(lat2) * np.sin(dlon/2.0)**2

    c = 2 * np.arcsin(np.sqrt(a))
    s12 = Radius * c
    
    # initial azimuth in degrees
    y = np.sin(lon2-lon1) * np.cos(lat2)
    x = np.cos(lat1)*np.sin(lat2) - np.sin(lat1)*np.cos(lat2)*np.cos(dlon)
    azi1 = np.arctan2(y, x)*180./math.pi

    return {'s12':s12, 'azi1': azi1}

Accuracy-wise, it is least accurate. Wikipedia states 0.5% of relative deviation on average without any sources. My experiments show less of a deviation. Below is the comparison ran on 100,000 random points vs my library, which should be accurate to millimeter levels:

np.random.seed(42)
lats1 = np.random.uniform(-90,90,100000)
lons1 = np.random.uniform(-180,180,100000)
lats2 = np.random.uniform(-90,90,100000)
lons2 = np.random.uniform(-180,180,100000)
r1 = inverse(lats1, lons1, lats2, lons2)
r2 = haversine(lats1, lons1, lats2, lons2)
print("Max absolute error: {:4.2f}m".format(np.max(r1['s12']-r2['s12'])))
print("Mean absolute error: {:4.2f}m".format(np.mean(r1['s12']-r2['s12'])))
print("Max relative error: {:4.2f}%".format(np.max((r2['s12']/r1['s12']-1)*100)))
print("Mean relative error: {:4.2f}%".format(np.mean((r2['s12']/r1['s12']-1)*100)))

Output:

Max absolute error: 26671.47m
Mean absolute error: -2499.84m
Max relative error: 0.55%
Mean relative error: -0.02%

So on average 2.5km deviation on 100,000 random pairs of coordinates, which may be good for majority of cases.

Next option is Vincenty's formulae which is accurate up to millimeters, depending on convergence criteria and can be vectorized as well. It does have the issue with convergence near antipodal points. You can make it converge at those points by relaxing convergence criteria, but accuracy drops to 0.25% and more. Outside of antipodal points Vincenty will provide results close to Geographiclib within relative error of less than 1.e-6 on average.

Geographiclib, mentioned here, is really the current golden standard. It has several implementations and fairly fast, especially if you are using C++ version.

Now, if you are planning to use Python for anything above 10k points I'd suggest to consider my vectorized implementation. I created a geovectorslib library with vectorized Vincenty routine for my own needs, which uses Geographiclib as fallback for near antipodal points. Below is the comparison vs Geographiclib for 100k points. As you can see it provides up to 20x improvement for inverse and 100x for direct methods for 100k points and the gap will grow with number of points. Accuracy-wise it will be within 1.e-5 rtol of Georgraphiclib.

Direct method for 100,000 points
94.9 ms ± 25 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)
9.79 s ± 1.4 s per loop (mean ± std. dev. of 7 runs, 1 loop each)

Inverse method for 100,000 points
1.5 s ± 504 ms per loop (mean ± std. dev. of 7 runs, 1 loop each)
24.2 s ± 3.91 s per loop (mean ± std. dev. of 7 runs, 1 loop each)

If you want the driving distance/route (posting it here because this is the first result for the distance between two points on google but for most people the driving distance is more useful), you can use Google Maps Distance Matrix Service:

getDrivingDistanceBetweenTwoLatLong(origin, destination) {

 return new Observable(subscriber => {
  let service = new google.maps.DistanceMatrixService();
  service.getDistanceMatrix(
    {
      origins: [new google.maps.LatLng(origin.lat, origin.long)],
      destinations: [new google.maps.LatLng(destination.lat, destination.long)],
      travelMode: 'DRIVING'
    }, (response, status) => {
      if (status !== google.maps.DistanceMatrixStatus.OK) {
        console.log('Error:', status);
        subscriber.error({error: status, status: status});
      } else {
        console.log(response);
        try {
          let valueInMeters = response.rows[0].elements[0].distance.value;
          let valueInKms = valueInMeters / 1000;
          subscriber.next(valueInKms);
          subscriber.complete();
        }
       catch(error) {
        subscriber.error({error: error, status: status});
       }
      }
    });
});
}

You could use a module like geolib too:

How to install:

$ npm install geolib

How to use:

import { getDistance } from 'geolib'

const distance = getDistance(
    { latitude: 51.5103, longitude: 7.49347 },
    { latitude: "51° 31' N", longitude: "7° 28' E" }
)

console.log(distance)

Documentation: https://www.npmjs.com/package/geolib

FSharp version, using miles:

let radialDistanceHaversine location1 location2 : float = 
                let degreeToRadian degrees = degrees * System.Math.PI / 180.0
                let earthRadius = 3959.0
                let deltaLat = location2.Latitude - location1.Latitude |> degreeToRadian
                let deltaLong = location2.Longitude - location1.Longitude |> degreeToRadian
                let a =
                    (deltaLat / 2.0 |> sin) ** 2.0
                    + (location1.Latitude |> degreeToRadian |> cos)
                    * (location2.Latitude |> degreeToRadian |> cos)
                    * (deltaLong / 2.0 |> sin) ** 2.0
                atan2 (a |> sqrt) (1.0 - a |> sqrt)
                * 2.0
                * earthRadius

Dart lang:

import 'dart:math' show cos, sqrt, asin;

double calculateDistance(LatLng l1, LatLng l2) {
  const p = 0.017453292519943295;
  final a = 0.5 -
      cos((l2.latitude - l1.latitude) * p) / 2 +
      cos(l1.latitude * p) *
          cos(l2.latitude * p) *
          (1 - cos((l2.longitude - l1.longitude) * p)) /
          2;
  return 12742 * asin(sqrt(a));
}

You can calculate it by using Haversine formula which is:

a = sin²(Δφ/2) + cos φ1 ⋅ cos φ2 ⋅ sin²(Δλ/2)
c = 2 ⋅ atan2( √a, √(1−a) )
d = R ⋅ c

An example to calculate distance between two points is given below

Suppose i have to calculate distance between New Delhi to London, so how can i use this formula :

New delhi co-ordinates= 28.7041° N, 77.1025° E
London co-ordinates= 51.5074° N, 0.1278° W

var R = 6371e3; // metres
var φ1 = 28.7041.toRadians();
var φ2 = 51.5074.toRadians();
var Δφ = (51.5074-28.7041).toRadians();
var Δλ = (0.1278-77.1025).toRadians();

var a = Math.sin(Δφ/2) * Math.sin(Δφ/2) +
        Math.cos(φ1) * Math.cos(φ2) *
        Math.sin(Δλ/2) * Math.sin(Δλ/2);
var c = 2 * Math.atan2(Math.sqrt(a), Math.sqrt(1-a));

var d = R * c; // metres
d = d/1000; // km

The functions needed for an accurate calculation of distance between lat-long points are complex, and the pitfalls are many. I would not recomend haversine or other spherical solutions due to the big inaccuracies (the earth is not a perfect sphere). The vincenty formula is better, but will in some cases throw errors, even when coded correctly.

Instead of coding the functions yourself I suggest using geopy which have implemented the very accurate geographiclib for distance calculations (paper from author).

#pip install geopy
from geopy.distance import geodesic
NY = [40.71278,-74.00594]
Beijing = [39.90421,116.40739]
print("WGS84: ",geodesic(NY, Beijing).km) #WGS84 is Standard
print("Intl24: ",geodesic(NY, Beijing, ellipsoid='Intl 1924').km) #geopy includes different ellipsoids
print("Custom ellipsoid: ",geodesic(NY, Beijing, ellipsoid=(6377., 6356., 1 / 297.)).km) #custom ellipsoid

#supported ellipsoids:
#model             major (km)   minor (km)     flattening
#'WGS-84':        (6378.137,    6356.7523142,  1 / 298.257223563)
#'GRS-80':        (6378.137,    6356.7523141,  1 / 298.257222101)
#'Airy (1830)':   (6377.563396, 6356.256909,   1 / 299.3249646)
#'Intl 1924':     (6378.388,    6356.911946,   1 / 297.0)
#'Clarke (1880)': (6378.249145, 6356.51486955, 1 / 293.465)
#'GRS-67':        (6378.1600,   6356.774719,   1 / 298.25)

The only drawback with this library is that it doesn't support vectorized calculations. For vectorized calculations you can use the new geovectorslib.

#pip install geovectorslib
from geovectorslib import inverse
print(inverse(lats1,lons1,lats2,lons2)['s12'])

lats and lons are numpy arrays. Geovectorslib is very accurate and extremly fast! I haven't found a solution for changing ellipsoids though. The WGS84 ellipsoid is used as standard, which is the best choice for most uses.

Here is the Erlang implementation

lat_lng({Lat1, Lon1}=_Point1, {Lat2, Lon2}=_Point2) ->
  P = math:pi() / 180,
  R = 6371, % Radius of Earth in KM
  A = 0.5 - math:cos((Lat2 - Lat1) * P) / 2 +
    math:cos(Lat1 * P) * math:cos(Lat2 * P) * (1 - math:cos((Lon2 - Lon1) * P))/2,
  R * 2 * math:asin(math:sqrt(A)).

If you are using python; pip install geopy

from geopy.distance import geodesic


origin = (30.172705, 31.526725)  # (latitude, longitude) don't confuse
destination = (30.288281, 31.732326)

print(geodesic(origin, destination).meters)  # 23576.805481751613
print(geodesic(origin, destination).kilometers)  # 23.576805481751613
print(geodesic(origin, destination).miles)  # 14.64994773134371

function distance($lat1, $lon1, $lat2, $lon2) { 
    $pi80 = M_PI / 180; 
    $lat1 *= $pi80; $lon1 *= $pi80; $lat2 *= $pi80; $lon2 *= $pi80; 
    $dlat = $lat2 - $lat1; 
    $dlon = $lon2 - $lon1; 
    $a = sin($dlat / 2) * sin($dlat / 2) + cos($lat1) * cos($lat2) * sin($dlon / 2) * sin($dlon / 2);  
    $km = 6372.797 * 2 * atan2(sqrt($a), sqrt(1 - $a)); 
    return $km; 
}

Here's a Scala implementation:

  def calculateHaversineDistance(lat1: Double, lon1: Double, lat2: Double, lon2: Double): Double = {
    val long2 = lon2 * math.Pi / 180
    val lat2 = lat2 * math.Pi / 180
    val long1 = lon1 * math.Pi / 180
    val lat1 = lat1 * math.Pi / 180

    val dlon = long2 - long1
    val dlat = lat2 - lat1
    val a = math.pow(math.sin(dlat / 2), 2) + math.cos(lat1) * math.cos(lat2) * math.pow(math.sin(dlon / 2), 2)
    val c = 2 * math.atan2(Math.sqrt(a), math.sqrt(1 - a))
    val haversineDistance = 3961 * c // 3961 = radius of earth in miles
    haversineDistance
  }