点到线段的最短距离

Question

我需要一个基本函数来找到点和线段之间的最短距离。随意用您想要的任何语言编写解决方案；我可以把它翻译成我正在使用的（Javascript）。

编辑：我的线段由两个端点定义。所以我的线段AB由A (x1,y1)和B (x2,y2)这两个点定义。我试图找出这条线段和点C (x3,y3) 之间的距离。我的几何技能生疏了，所以我看到的例子很混乱，很抱歉。

Answer 1

上述功能不适用于垂直线。这是一个运行良好的功能！ 与点 p1、p2 连线。检查点是 p;

public float DistanceOfPointToLine2(PointF p1, PointF p2, PointF p)
{
  //          (y1-y2)x + (x2-x1)y + (x1y2-x2y1)
  //d(P,L) = --------------------------------
  //         sqrt( (x2-x1)pow2 + (y2-y1)pow2 )

  double ch = (p1.Y - p2.Y) * p.X + (p2.X - p1.X) * p.Y + (p1.X * p2.Y - p2.X * p1.Y);
  double del = Math.Sqrt(Math.Pow(p2.X - p1.X, 2) + Math.Pow(p2.Y - p1.Y, 2));
  double d = ch / del;
  return (float)d;
}

Answer 2

这与 C++ 答案相同，但移植到帕斯卡。 point 参数的顺序已更改以适合我的代码，但相同。

function Dot(const p1, p2: PointF): double;
begin
  Result := p1.x * p2.x + p1.y * p2.y;
end;
function SubPoint(const p1, p2: PointF): PointF;
begin
  result.x := p1.x - p2.x;
  result.y := p1.y - p2.y;
end;

function ShortestDistance2(const p,v,w : PointF) : double;
var
  l2,t : double;
  projection,tt: PointF;
begin
  // Return minimum distance between line segment vw and point p
  //l2 := length_squared(v, w);  // i.e. |w-v|^2 -  avoid a sqrt
  l2 := Distance(v,w);
  l2 := MPower(l2,2);
  if (l2 = 0.0) then begin
    result:= Distance(p, v);   // v == w case
    exit;
  end;
  // Consider the line extending the segment, parameterized as v + t (w - v).
  // We find projection of point p onto the line.
  // It falls where t = [(p-v) . (w-v)] / |w-v|^2
  t := Dot(SubPoint(p,v),SubPoint(w,v)) / l2;
  if (t < 0.0) then begin
    result := Distance(p, v);       // Beyond the 'v' end of the segment
    exit;
  end
  else if (t > 1.0) then begin
    result := Distance(p, w);  // Beyond the 'w' end of the segment
    exit;
  end;
  //projection := v + t * (w - v);  // Projection falls on the segment
  tt.x := v.x + t * (w.x - v.x);
  tt.y := v.y + t * (w.y - v.y);
  result := Distance(p, tt);
end;

Answer 3

%Matlab solution by Tim from Cody
function ans=distP2S(x0,y0,x1,y1,x2,y2)
% Point is x0,y0
z=complex(x0-x1,y0-y1);
complex(x2-x1,y2-y1);
abs(z-ans*min(1,max(0,real(z/ans))));

Answer 4

GLSL 版本：

// line (a -> b ) point p[enter image description here][1]
float distanceToLine(vec2 a, vec2 b, vec2 p) {
    float aside = dot((p - a),(b - a));
    if(aside< 0.0) return length(p-a);
    float bside = dot((p - b),(a - b));
    if(bside< 0.0) return length(p-b);
    vec2 pointOnLine = (bside*a + aside*b)/pow(length(a-b),2.0);
    return length(p - pointOnLine);
}

Answer 5

该算法基于找到指定线与包含指定点的正交线的交点，并计算其距离。如果是线段，我们必须检查交点是否在线段的点之间，如果不是，则最小距离在指定点和线段的端点之一之间。这是一个 C# 实现。

Double Distance(Point a, Point b)
{
    double xdiff = a.X - b.X, ydiff = a.Y - b.Y;
    return Math.Sqrt((long)xdiff * xdiff + (long)ydiff * ydiff);
}

Boolean IsBetween(double x, double a, double b)
{
    return ((a <= b && x >= a && x <= b) || (a > b && x <= a && x >= b));
}

Double GetDistance(Point pt, Point pt1, Point pt2, out Point intersection)
{
    Double a, x, y, R;

    if (pt1.X != pt2.X) {
        a = (double)(pt2.Y - pt1.Y) / (pt2.X - pt1.X);
        x = (a * (pt.Y - pt1.Y) + a * a * pt1.X + pt.X) / (a * a + 1);
        y = a * x + pt1.Y - a * pt1.X; }
    else { x = pt1.X;  y = pt.Y; }

    if (IsBetween(x, pt1.X, pt2.X) && IsBetween(y, pt1.Y, pt2.Y)) {
        intersection = new Point((int)x, (int)y);
        R = Distance(intersection, pt); }
    else {
        double d1 = Distance(pt, pt1), d2 = Distance(pt, pt2);
        if (d1 < d2) { intersection = pt1; R = d1; }
        else { intersection = pt2; R = d2; }}

    return R;
}

Answer 6

这是一个基于向量数学的；此解决方案也适用于更高维度并且还报告交点（在线段上）。

def dist(x1,y1,x2,y2,px,py):
    a = np.array([[x1,y1]]).T
    b = np.array([[x2,y2]]).T
    x = np.array([[px,py]]).T
    tp = (np.dot(x.T, b) - np.dot(a.T, b)) / np.dot(b.T, b)
    tp = tp[0][0]
    tmp = x - (a + tp*b)
    d = np.sqrt(np.dot(tmp.T,tmp)[0][0])
    return d, a+tp*b

x1,y1=2.,2.
x2,y2=5.,5.
px,py=4.,1.

d, inters = dist(x1,y1, x2,y2, px,py)
print (d)
print (inters)

结果是

2.1213203435596424
[[2.5]
 [2.5]]

这里解释了数学

https://brilliant.org/wiki/distance-between-point-and-line/

Answer 7

飞镖和颤振的解决方案：

import 'dart:math' as math;
 class Utils {
   static double shortestDistance(Point p1, Point p2, Point p3){
      double px = p2.x - p1.x;
      double py = p2.y - p1.y;
      double temp = (px*px) + (py*py);
      double u = ((p3.x - p1.x)*px + (p3.y - p1.y)* py) /temp;
      if(u>1){
        u=1;
      }
      else if(u<0){
        u=0;
      }
      double x = p1.x + u*px;
      double y = p1.y + u*py;
      double dx = x - p3.x;
      double dy = y - p3.y;
      double dist = math.sqrt(dx*dx+dy*dy);
      return dist;
   }
}

class Point {
  double x;
  double y;
  Point(this.x, this.y);
}

Answer 8

二维坐标数组的 Python Numpy 实现：

import numpy as np


def dist2d(p1, p2, coords):
    ''' Distance from points to a finite line btwn p1 -> p2 '''
    assert coords.ndim == 2 and coords.shape[1] == 2, 'coords is not 2 dim'
    dp = p2 - p1
    st = dp[0]**2 + dp[1]**2
    u = ((coords[:, 0] - p1[0]) * dp[0] + (coords[:, 1] - p1[1]) * dp[1]) / st

    u[u > 1.] = 1.
    u[u < 0.] = 0.

    dx = (p1[0] + u * dp[0]) - coords[:, 0]
    dy = (p1[1] + u * dp[1]) - coords[:, 1]

    return np.sqrt(dx**2 + dy**2)


# Usage:
p1 = np.array([0., 0.])
p2 = np.array([0., 10.])

# List of coordinates
coords = np.array(
    [[0., 0.],
     [5., 5.],
     [10., 10.],
     [20., 20.]
     ])

d = dist2d(p1, p2, coords)

# Single coordinate
coord = np.array([25., 25.])
d = dist2d(p1, p2, coord[np.newaxis, :])

Answer 9

Matlab 直接 Grumdrig 实现

function ans=distP2S(px,py,vx,vy,wx,wy)
% [px py vx vy wx wy]
  t=( (px-vx)*(wx-vx)+(py-vy)*(wy-vy) )/idist(vx,wx,vy,wy)^2;
  [idist(px,vx,py,vy) idist(px,vx+t*(wx-vx),py,vy+t*(wy-vy)) idist(px,wx,py,wy) ];
  ans(1+(t>0)+(t>1)); % <0 0<=t<=1 t>1     
 end

function d=idist(a,b,c,d)
 d=abs(a-b+1i*(c-d));
end

Answer 10

如果它是一条无限线，而不是线段，最简单的方法是这个（在 ruby​​ 中），其中 mx + b 是线，(x1, y1) 是已知点

(y1 - mx1 - b).abs / Math.sqrt(m**2 + 1)

Answer 11

刚遇到这个，我想我会添加一个 Lua 实现。它假定点以表格 {x=xVal, y=yVal} 的形式给出，而线或线段由包含两个点的表格给出（参见下面的示例）：

function distance( P1, P2 )
    return math.sqrt((P1.x-P2.x)^2 + (P1.y-P2.y)^2)
end

-- Returns false if the point lies beyond the reaches of the segment
function distPointToSegment( line, P )
    if line[1].x == line[2].x and line[1].y == line[2].y then
        print("Error: Not a line!")
        return false
    end

    local d = distance( line[1], line[2] )

    local t = ((P.x - line[1].x)*(line[2].x - line[1].x) + (P.y - line[1].y)*(line[2].y - line[1].y))/(d^2)

    local projection = {}
    projection.x = line[1].x + t*(line[2].x-line[1].x)
    projection.y = line[1].y + t*(line[2].y-line[1].y)

    if t >= 0 and t <= 1 then   -- within line segment?
        return distance( projection, {x=P.x, y=P.y} )
    else
        return false
    end
end

-- Returns value even if point is further down the line (outside segment)
function distPointToLine( line, P )
    if line[1].x == line[2].x and line[1].y == line[2].y then
        print("Error: Not a line!")
        return false
    end

    local d = distance( line[1], line[2] )

    local t = ((P.x - line[1].x)*(line[2].x - line[1].x) + (P.y - line[1].y)*(line[2].y - line[1].y))/(d^2)

    local projection = {}
    projection.x = line[1].x + t*(line[2].x-line[1].x)
    projection.y = line[1].y + t*(line[2].y-line[1].y)

    return distance( projection, {x=P.x, y=P.y} )
end

示例用法：

local P1 = {x = 0, y = 0}
local P2 = {x = 10, y = 10}
local line = { P1, P2 }
local P3 = {x = 7, y = 15}
print(distPointToLine( line, P3 ))  -- prints 5.6568542494924
print(distPointToSegment( line, P3 )) -- prints false

Answer 12

基于 formula 的 JavaScript 中更简洁的解决方案：

distToSegment: function (point, linePointA, linePointB){

    var x0 = point.X;
    var y0 = point.Y;

    var x1 = linePointA.X;
    var y1 = linePointA.Y;

    var x2 = linePointB.X;
    var y2 = linePointB.Y;

    var Dx = (x2 - x1);
    var Dy = (y2 - y1);

    var numerator = Math.abs(Dy*x0 - Dx*y0 - x1*y2 + x2*y1);
    var denominator = Math.sqrt(Dx*Dx + Dy*Dy);
    if (denominator == 0) {
        return this.dist2(point, linePointA);
    }

    return numerator/denominator;

}

Answer 13

想在 GLSL 中执行此操作，但如果可能，最好避免所有这些条件。使用clamp() 避免了两种端点情况：

// find closest point to P on line segment AB:
vec3 closest_point_on_line_segment(in vec3 P, in vec3 A, in vec3 B) {
    vec3 AP = P - A, AB = B - A;
    float l = dot(AB, AB);
    if (l <= 0.0000001) return A;    // A and B are practically the same
    return AP - AB*clamp(dot(AP, AB)/l, 0.0, 1.0);  // do the projection
}

如果您可以确定 A 和 B 永远不会彼此非常接近，则可以将其简化为删除 if()。事实上，即使 A 和 B 相同，我的 GPU 仍然使用这个无条件版本给出正确的结果（但这是使用 GLSL 除以零的 pre-OpenGL 4.1 未定义） ：

// find closest point to P on line segment AB:
vec3 closest_point_on_line_segment(in vec3 P, in vec3 A, in vec3 B) {
    vec3 AP = P - A, AB = B - A;
    return AP - AB*clamp(dot(AP, AB)/dot(AB, AB), 0.0, 1.0);
}

计算距离很简单——GLSL 提供了一个 distance() 函数，您可以在这个最近点和 P 上使用它。

灵感来自Iñigo Quilez's code for a capsule distance function

Answer 14

与this answer 相同，但在 Visual Basic 中除外。使其可用作 Microsoft Excel 和 VBA/宏中的用户定义函数。

函数返回点(x,y)到(x1,y1)和(x2,y2)定义的线段的最近距离

Function DistanceToSegment(x As Double, y As Double, x1 As Double, y1 As Double, x2 As Double, y2 As Double)

  Dim A As Double
  A = x - x1
  Dim B As Double
  B = y - y1
  Dim C  As Double
  C = x2 - x1
  Dim D As Double
  D = y2 - y1

  Dim dot As Double
  dot = A * C + B * D
  Dim len_sq As Double
  len_sq = C * C + D * D
  Dim param As Double
  param = -1

  If (len_sq <> 0) Then
      param = dot / len_sq
  End If

  Dim xx As Double
  Dim yy As Double

  If (param < 0) Then
    xx = x1
    yy = y1
  ElseIf (param > 1) Then
    xx = x2
    yy = y2
  Else
    xx = x1 + param * C
    yy = y1 + param * D
  End If

  Dim dx As Double
  dx = x - xx
  Dim dy As Double
  dy = y - yy

  DistanceToSegment = Math.Sqr(dx * dx + dy * dy)

End Function

Answer 15

卢亚： 查找线段（不是整条线）和点之间的最小距离

function solveLinearEquation(A1,B1,C1,A2,B2,C2)
--it is the implitaion of a method of solving linear equations in x and y
  local f1 = B1*C2 -B2*C1
  local f2 = A2*C1-A1*C2
  local f3 = A1*B2 -A2*B1
  return {x= f1/f3, y= f2/f3}
end


function pointLiesOnLine(x,y,x1,y1,x2,y2)
  local dx1 = x-x1
  local  dy1 = y-y1
  local dx2 = x-x2
  local  dy2 = y-y2
  local crossProduct = dy1*dx2 -dx1*dy2

if crossProduct ~= 0  then  return  false
else
  if ((x1>=x) and (x>=x2)) or ((x2>=x) and (x>=x1)) then
    if ((y1>=y) and (y>=y2)) or ((y2>=y) and (y>=y1)) then
      return true
    else return false end
  else  return false end
end
end


function dist(x1,y1,x2,y2)
  local dx = x1-x2
  local dy = y1-y2
  return math.sqrt(dx*dx + dy* dy)
 end


function findMinDistBetnPointAndLine(x1,y1,x2,y2,x3,y3)
-- finds the min  distance between (x3,y3) and line (x1,y2)--(x2,y2)
   local A2,B2,C2,A1,B1,C1
   local dx = y2-y1
   local dy = x2-x1
   if dx == 0 then A2=1 B2=0 C2=-x3 A1=0 B1=1 C1=-y1 
   elseif dy == 0 then A2=0 B2=1 C2=-y3 A1=1 B1=0 C1=-x1
   else
      local m1 = dy/dx
      local m2 = -1/m1
      A2=m2 B2=-1 C2=y3-m2*x3 A1=m1 B1=-1 C1=y1-m1*x1
   end
 local intsecPoint= solveLinearEquation(A1,B1,C1,A2,B2,C2)
if pointLiesOnLine(intsecPoint.x, intsecPoint.y,x1,y1,x2,y2) then
   return dist(intsecPoint.x, intsecPoint.y, x3,y3)
 else
   return math.min(dist(x3,y3,x1,y1),dist(x3,y3,x2,y2))
end
end

Answer 16

此答案基于accepted answer 的 JavaScript 解决方案。 它主要只是格式更好，函数名称更长，当然函数语法更短，因为它在 ES6 + CoffeeScript 中。

JavaScript 版本 (ES6)

distanceSquared = (v, w)=> Math.pow(v.x - w.x, 2) + Math.pow(v.y - w.y, 2);
distance = (v, w)=> Math.sqrt(distanceSquared(v, w));

distanceToLineSegmentSquared = (p, v, w)=> {
    l2 = distanceSquared(v, w);
    if (l2 === 0) {
        return distanceSquared(p, v);
    }
    t = ((p.x - v.x) * (w.x - v.x) + (p.y - v.y) * (w.y - v.y)) / l2;
    t = Math.max(0, Math.min(1, t));
    return distanceSquared(p, {
        x: v.x + t * (w.x - v.x),
        y: v.y + t * (w.y - v.y)
    });
}
distanceToLineSegment = (p, v, w)=> {
    return Math.sqrt(distanceToLineSegmentSquared(p, v));
}

CoffeeScript 版本

distanceSquared = (v, w)-> (v.x - w.x) ** 2 + (v.y - w.y) ** 2
distance = (v, w)-> Math.sqrt(distanceSquared(v, w))

distanceToLineSegmentSquared = (p, v, w)->
    l2 = distanceSquared(v, w)
    return distanceSquared(p, v) if l2 is 0
    t = ((p.x - v.x) * (w.x - v.x) + (p.y - v.y) * (w.y - v.y)) / l2
    t = Math.max(0, Math.min(1, t))
    distanceSquared(p, {
        x: v.x + t * (w.x - v.x)
        y: v.y + t * (w.y - v.y)
    })

distanceToLineSegment = (p, v, w)->
    Math.sqrt(distanceToLineSegmentSquared(p, v, w))

Answer 17

在R中

     #distance beetween segment ab and point c in 2D space
getDistance_ort_2 <- function(a, b, c){
  #go to complex numbers
  A<-c(a[1]+1i*a[2],b[1]+1i*b[2])
  q=c[1]+1i*c[2]
  
  #function to get coefficients of line (ab)
  getAlphaBeta <- function(A)
  { a<-Re(A[2])-Re(A[1])
    b<-Im(A[2])-Im(A[1])
    ab<-as.numeric()
    ab[1] <- -Re(A[1])*b/a+Im(A[1])
    ab[2] <-b/a
    if(Im(A[1])==Im(A[2])) ab<- c(Im(A[1]),0)
    if(Re(A[1])==Re(A[2])) ab <- NA
    return(ab)
  }
  
  #function to get coefficients of line ortogonal to line (ab) which goes through point q
  getAlphaBeta_ort<-function(A,q)
  { ab <- getAlphaBeta(A) 
  coef<-c(Re(q)/ab[2]+Im(q),-1/ab[2])
  if(Re(A[1])==Re(A[2])) coef<-c(Im(q),0)
  return(coef)
  }
  
  #function to get coordinates of interception point 
  #between line (ab) and its ortogonal which goes through point q
  getIntersection_ort <- function(A, q){
    A.ab <- getAlphaBeta(A)
    q.ab <- getAlphaBeta_ort(A,q)
    if (!is.na(A.ab[1])&A.ab[2]==0) {
      x<-Re(q)
      y<-Im(A[1])}
    if (is.na(A.ab[1])) {
      x<-Re(A[1])
      y<-Im(q)
    } 
    if (!is.na(A.ab[1])&A.ab[2]!=0) {
      x <- (q.ab[1] - A.ab[1])/(A.ab[2] - q.ab[2])
      y <- q.ab[1] + q.ab[2]*x}
    xy <- x + 1i*y  
    return(xy)
  }
  
  intersect<-getIntersection_ort(A,q)
  if ((Mod(A[1]-intersect)+Mod(A[2]-intersect))>Mod(A[1]-A[2])) {dist<-min(Mod(A[1]-q),Mod(A[2]-q))
  } else dist<-Mod(q-intersect)
  return(dist)
}



 

Answer 18

在javascript中使用几何：

var a = { x:20, y:20};//start segment    
var b = { x:40, y:30};//end segment   
var c = { x:37, y:14};//point   

// magnitude from a to c    
var ac = Math.sqrt( Math.pow( ( a.x - c.x ), 2 ) + Math.pow( ( a.y - c.y ), 2) );    
// magnitude from b to c   
var bc = Math.sqrt( Math.pow( ( b.x - c.x ), 2 ) + Math.pow( ( b.y - c.y ), 2 ) );    
// magnitude from a to b (base)     
var ab = Math.sqrt( Math.pow( ( a.x - b.x ), 2 ) + Math.pow( ( a.y - b.y ), 2 ) );    
 // perimeter of triangle     
var p = ac + bc + ab;    
 // area of the triangle    
var area = Math.sqrt( p/2 * ( p/2 - ac) * ( p/2 - bc ) * ( p/2 - ab ) );    
 // height of the triangle = distance    
var h = ( area * 2 ) / ab;    
console.log ("height: " + h);

Answer 19

用于 3D 线段和点的 Eigen C++ 版本

// Return minimum distance between line segment: head--->tail and point
double MinimumDistance(Eigen::Vector3d head, Eigen::Vector3d tail,Eigen::Vector3d point)
{
    double l2 = std::pow((head - tail).norm(),2);
    if(l2 ==0.0) return (head - point).norm();// head == tail case

    // Consider the line extending the segment, parameterized as head + t (tail - point).
    // We find projection of point onto the line.
    // It falls where t = [(point-head) . (tail-head)] / |tail-head|^2
    // We clamp t from [0,1] to handle points outside the segment head--->tail.

    double t = max(0,min(1,(point-head).dot(tail-head)/l2));
    Eigen::Vector3d projection = head + t*(tail-head);

    return (point - projection).norm();
}

Answer 20

根据上面 Joshua 的回答，这是一个自包含的 Delphi / Pascal 版本的函数。对 VCL 屏幕图形使用 TPoint，但应该很容易根据需要进行调整。

function DistancePtToSegment( pt, pt1, pt2: TPoint): double;
var
   a, b, c, d: double;
   len_sq: double;
   param: double;
   xx, yy: double;
   dx, dy: double;
begin
   a := pt.x - pt1.x;
   b := pt.y - pt1.y;
   c := pt2.x - pt1.x;
   d := pt2.y - pt1.y;

   len_sq := (c * c) + (d * d);
   param := -1;

   if (len_sq <> 0) then
   begin
      param := ((a * c) + (b * d)) / len_sq;
   end;

   if param < 0 then
   begin
      xx := pt1.x;
      yy := pt1.y;
   end
   else if param > 1 then
   begin
      xx := pt2.x;
      yy := pt2.y;
   end
   else begin
      xx := pt1.x + param * c;
      yy := pt1.y + param * d;
   end;

   dx := pt.x - xx;
   dy := pt.y - yy;
   result := sqrt( (dx * dx) + (dy * dy))
end;

Answer 21

Swift 实现http://paulbourke.net/geometry/pointlineplane/source.c

    static func magnitude(p1: CGPoint, p2: CGPoint) -> CGFloat {
        let vector = CGPoint(x: p2.x - p1.x, y: p2.y - p1.y)
        return sqrt(pow(vector.x, 2) + pow(vector.y, 2))
    }

    /// http://paulbourke.net/geometry/pointlineplane/
    /// http://paulbourke.net/geometry/pointlineplane/source.c
    static func pointDistanceToLine(point: CGPoint, lineStart: CGPoint, lineEnd: CGPoint) -> CGFloat? {

        let lineMag = magnitude(p1: lineEnd, p2: lineStart)
        let u = (((point.x - lineStart.x) * (lineEnd.x - lineStart.x)) +
                ((point.y - lineStart.y) * (lineEnd.y - lineStart.y))) /
                (lineMag * lineMag)

        if u < 0 || u > 1 {
            // closest point does not fall within the line segment
            return nil
        }

        let intersectionX = lineStart.x + u * (lineEnd.x - lineStart.x)
        let intersectionY = lineStart.y + u * (lineEnd.y - lineStart.y)

        return magnitude(p1: point, p2: CGPoint(x: intersectionX, y: intersectionY))
    }

Answer 22

Lua 解决方案

-- distance from point (px, py) to line segment (x1, y1, x2, y2)
function distPointToLine(px,py,x1,y1,x2,y2) -- point, start and end of the segment
    local dx,dy = x2-x1,y2-y1
    local length = math.sqrt(dx*dx+dy*dy)
    dx,dy = dx/length,dy/length -- normalization
    local p = dx*(px-x1)+dy*(py-y1)
    if p < 0 then
        dx,dy = px-x1,py-y1
        return math.sqrt(dx*dx+dy*dy), x1, y1 -- distance, nearest point
    elseif p > length then
        dx,dy = px-x2,py-y2
        return math.sqrt(dx*dx+dy*dy), x2, y2 -- distance, nearest point
    end
    return math.abs(dy*(px-x1)-dx*(py-y1)), x1+dx*p, y1+dy*p -- distance, nearest point
end

对于折线（多于两段的线）：

-- if the (poly-)line has several segments, just iterate through all of them:
function nearest_sector_in_line (x, y, line)
    local x1, y1, x2, y2, min_dist
    local ax,ay = line[1], line[2]
    for j = 3, #line-1, 2 do
        local bx,by = line[j], line[j+1]
        local dist = distPointToLine(x,y,ax,ay,bx,by)
        if not min_dist or dist < min_dist then
            min_dist = dist
            x1, y1, x2, y2 = ax,ay,bx,by
        end
        ax, ay = bx, by
    end
    return x1, y1, x2, y2
end

例子：

-- call it:
local x1, y1, x2, y2 = nearest_sector_in_line (7, 4, {0,0, 10,0, 10,10, 0,10})

Answer 23

这是 HSQLDB 的 SQL 实现：

CREATE FUNCTION dist_to_segment(px double, py double, vx double, vy double, wx double, wy double)
  RETURNS double
BEGIN atomic
   declare l2 double;
   declare t double;
   declare nx double;
   declare ny double;
   set l2 =(vx - wx)*(vx - wx) + (vy - wy)*(vy - wy);
   IF l2 = 0 THEN
     RETURN sqrt((vx - px)*(vx - px) + (vy - py)*(vy - py));
   ELSE
     set t = ((px - vx) * (wx - vx) + (py - vy) * (wy - vy)) / l2;
     set t = GREATEST(0, LEAST(1, t));
     set nx=vx + t * (wx - vx);
     set ny=vy + t * (wy - vy);
     RETURN sqrt((nx - px)*(nx - px) + (ny - py)*(ny - py));
   END IF;
END;

Postgres 的实现：

CREATE FUNCTION dist_to_segment(px numeric, py numeric, vx numeric, vy numeric, wx numeric, wy numeric)
  RETURNS numeric
AS $$
   declare l2 numeric;
   declare t numeric;
   declare nx numeric;
   declare ny numeric;
BEGIN 
   l2 := (vx - wx)*(vx - wx) + (vy - wy)*(vy - wy);
   IF l2 = 0 THEN
     RETURN sqrt((vx - px)*(vx - px) + (vy - py)*(vy - py));
   ELSE
     t := ((px - vx) * (wx - vx) + (py - vy) * (wy - vy)) / l2;
     t := GREATEST(0, LEAST(1, t));
     nx := vx + t * (wx - vx);
     ny := vy + t * (wy - vy);
     RETURN sqrt((nx - px)*(nx - px) + (ny - py)*(ny - py));
   END IF;
END;
$$ LANGUAGE plpgsql;

Answer 24

我需要一个 Godot (GDscript) 实现，所以我根据grumdrig's 接受的答案写了一个：

func minimum_distance(v: Vector2, w: Vector2, p: Vector2):
    # Return minimum distance between line segment vw and point p
    var l2: float = (v - w).length_squared()  # i.e. |w-v|^2 -  avoid a sqrt
    if l2 == 0.0:
        return p.distance_to(v) # v == w case

    # Consider the line extending the segment, parameterized as v + t (w - v).
    # We find projection of point p onto the line.
    # It falls where t = [(p-v) . (w-v)] / |w-v|^2
    # We clamp t from [0,1] to handle points outside the segment vw.
    var t: float = max(0, min(1, (p - v).dot(w - v) / l2))
    var projection: Vector2 = v + t * (w - v)  # Projection falls on the segment
    
    return p.distance_to(projection)