How to determine if a point is in a 2D triangle? [closed]

Answer 1

If you are looking for speed, here is a procedure that might help you.

Sort the triangle vertices on their ordinates. This takes at worst three comparisons. Let Y0, Y1, Y2 be the three sorted values. By drawing three horizontals through them you partition the plane into two half planes and two slabs. Let Y be the ordinate of the query point.

if Y < Y1
    if Y <= Y0 -> the point lies in the upper half plane, outside the triangle; you are done
    else Y > Y0 -> the point lies in the upper slab
else
    if Y >= Y2 -> the point lies in the lower half plane, outside the triangle; you are done
    else Y < Y2 -> the point lies in the lower slab

Costs two more comparisons. As you see, quick rejection is achieved for points outside of the "bounding slab".

Optionally, you can supply a test on the abscissas for quick rejection on the left and on the right (X <= X0' or X >= X2'). This will implement a quick bounding box test at the same time, but you'll need to sort on the abscissas too.

Eventually you will need to compute the sign of the given point with respect to the two sides of the triangle that delimit the relevant slab (upper or lower). The test has the form:

((X - Xi) * (Y - Yj) > (X - Xi) * (Y - Yj)) == ((X - Xi) * (Y - Yk) > (X - Xi) * (Y - Yk))

The complete discussion of i, j, k combinations (there are six of them, based on the outcome of the sort) is out of the scope of this answer and "left as an exercise to the reader"; for efficiency, they should be hard-coded.

If you think that this solution is complex, observe that it mainly involves simple comparisons (some of which can be precomputed), plus 6 subtractions and 4 multiplies in case the bounding box test fails. The latter cost is hard to beat as in the worst case you cannot avoid comparing the test point against two sides (no method in other answers has a lower cost, some make it worse, like 15 subtractions and 6 multiplies, sometimes divisions).

UPDATE: Faster with a shear transform

As explained just above, you can quickly locate the point inside one of the four horizontal bands delimited by the three vertex ordinates, using two comparisons.

You can optionally perform one or two extra X tests to check insideness to the bounding box (dotted lines).

Then consider the "shear" transform given by X'= X - m Y, Y' = Y, where m is the slope DX/DY for the highest edge. This transform will make this side of the triangle vertical. And since you know on what side of the middle horizontal you are, it suffices to test the sign with respect to a single side of the triangle.

Assuming you precomputed the slope m, as well as the X' for the sheared triangle vertices and the coefficients of the equations of the sides as X = m Y + p, you will need in the worst case

• two ordinate comparisons for vertical classification;
• optionally one or two abscissa comparisons for bounding box rejection;
• computation of X' = X - m Y;
• one or two comparisons with the abscissas of the sheared triangle;
• one sign test X >< m' Y + p' against the relevant side of the sheared triangle.

Answer 2

Here is an efficient Python implementation:

def PointInsideTriangle2(pt,tri):
    '''checks if point pt(2) is inside triangle tri(3x2). @Developer'''
    a = 1/(-tri[1,1]*tri[2,0]+tri[0,1]*(-tri[1,0]+tri[2,0])+ \
        tri[0,0]*(tri[1,1]-tri[2,1])+tri[1,0]*tri[2,1])
    s = a*(tri[2,0]*tri[0,1]-tri[0,0]*tri[2,1]+(tri[2,1]-tri[0,1])*pt[0]+ \
        (tri[0,0]-tri[2,0])*pt[1])
    if s<0: return False
    else: t = a*(tri[0,0]*tri[1,1]-tri[1,0]*tri[0,1]+(tri[0,1]-tri[1,1])*pt[0]+ \
              (tri[1,0]-tri[0,0])*pt[1])
    return ((t>0) and (1-s-t>0))

and an example output:

Answer 3

C# version of the barycentric method posted by andreasdr and Perro Azul. Note that the area calculation can be avoided if s and t have opposite signs. I verified correct behavior with a pretty thorough unit test.

public static bool PointInTriangle(Point p, Point p0, Point p1, Point p2)
{
    var s = p0.Y * p2.X - p0.X * p2.Y + (p2.Y - p0.Y) * p.X + (p0.X - p2.X) * p.Y;
    var t = p0.X * p1.Y - p0.Y * p1.X + (p0.Y - p1.Y) * p.X + (p1.X - p0.X) * p.Y;

    if ((s < 0) != (t < 0))
        return false;

    var A = -p1.Y * p2.X + p0.Y * (p2.X - p1.X) + p0.X * (p1.Y - p2.Y) + p1.X * p2.Y;

    return A < 0 ?
            (s <= 0 && s + t >= A) :
            (s >= 0 && s + t <= A);
}

_[edit]
^{accepted suggested modification by @Pierre; see comments}

Answer 4

I agree with Andreas Brinck, barycentric coordinates are very convenient for this task. Note that there is no need to solve an equation system every time: just evaluate the analytical solution. Using Andreas' notation, the solution is:

s = 1/(2*Area)*(p0y*p2x - p0x*p2y + (p2y - p0y)*px + (p0x - p2x)*py);
t = 1/(2*Area)*(p0x*p1y - p0y*p1x + (p0y - p1y)*px + (p1x - p0x)*py);

where Area is the (signed) area of the triangle:

Area = 0.5 *(-p1y*p2x + p0y*(-p1x + p2x) + p0x*(p1y - p2y) + p1x*p2y);

Just evaluate s, t and 1-s-t. The point p is inside the triangle if and only if they are all positive.

EDIT: Note that the above expression for the area assumes that the triangle node numbering is counter-clockwise. If the numbering is clockwise, this expression will return a negative area (but with correct magnitude). The test itself (s>0 && t>0 && 1-s-t>0) doesn't depend on the direction of the numbering, however, since the expressions above that are multiplied by 1/(2*Area) also change sign if the triangle node orientation changes.

EDIT 2: For an even better computational efficiency, see coproc's comment below (which makes the point that if the orientation of the triangle nodes (clockwise or counter-clockwise) is known beforehand, the division by 2*Area in the expressions for s and t can be avoided). See also Perro Azul's jsfiddle-code in the comments under Andreas Brinck's answer.

Answer 5

Since there's no JS answer,
Clockwise & Counter-Clockwise solution:

function triangleContains(ax, ay, bx, by, cx, cy, x, y) {

    let det = (bx - ax) * (cy - ay) - (by - ay) * (cx - ax)

    return  det * ((bx - ax) * (y - ay) - (by - ay) * (x - ax)) > 0 &&
            det * ((cx - bx) * (y - by) - (cy - by) * (x - bx)) > 0 &&
            det * ((ax - cx) * (y - cy) - (ay - cy) * (x - cx)) > 0 

}

EDIT: there was a typo for det computation (cy - ay instead of cx - ax), this is fixed.

https://jsfiddle.net/jniac/rctb3gfL/

function triangleContains(ax, ay, bx, by, cx, cy, x, y) {

    let det = (bx - ax) * (cy - ay) - (by - ay) * (cx - ax)
	
    return  det * ((bx - ax) * (y - ay) - (by - ay) * (x - ax)) > 0 &&
            det * ((cx - bx) * (y - by) - (cy - by) * (x - bx)) > 0 &&
            det * ((ax - cx) * (y - cy) - (ay - cy) * (x - cx)) > 0 

}






let width = 500, height = 500

// clockwise
let triangle1 = {

	A : { x: 10, y: -10 },
	C : { x: 20, y: 100 },
	B : { x: -90, y: 10 },
	
	color: '#f00',

}

// counter clockwise
let triangle2 = {

	A : { x: 20, y: -60 },
	B : { x: 90, y: 20 },
	C : { x: 20, y: 60 },

	color: '#00f',
	
}


let scale = 2
let mouse = { x: 0, y: 0 }






// DRAW >

let wrapper = document.querySelector('div.wrapper')

wrapper.onmousemove = ({ layerX:x, layerY:y }) => {
	
	x -= width / 2
	y -= height / 2
	x /= scale
	y /= scale
	
	mouse.x = x
	mouse.y = y
	
	drawInteractive()

}

function drawArrow(ctx, A, B) {

	let v = normalize(sub(B, A), 3)
	let I = center(A, B)
	
	let p
	
	p = add(I, rotate(v, 90), v)
	ctx.moveTo(p.x, p.y)
	ctx.lineTo(I.x, I .y)
	p = add(I, rotate(v, -90), v)
	ctx.lineTo(p.x, p.y)

}

function drawTriangle(ctx, { A, B, C, color }) {

	ctx.beginPath()
	ctx.moveTo(A.x, A.y)
	ctx.lineTo(B.x, B.y)
	ctx.lineTo(C.x, C.y)
	ctx.closePath()
	
	ctx.fillStyle = color + '6'
	ctx.strokeStyle = color
	ctx.fill()
	
	drawArrow(ctx, A, B)
	drawArrow(ctx, B, C)
	drawArrow(ctx, C, A)
	
	ctx.stroke()

}

function contains({ A, B, C }, P) {

	return triangleContains(A.x, A.y, B.x, B.y, C.x, C.y, P.x, P.y)

}

function resetCanvas(canvas) {

	canvas.width = width
	canvas.height = height
	
	let ctx = canvas.getContext('2d')

	ctx.resetTransform()
	ctx.clearRect(0, 0, width, height)
	ctx.setTransform(scale, 0, 0, scale, width/2, height/2)
	
}

function drawDots() {

	let canvas = document.querySelector('canvas#dots')
	let ctx = canvas.getContext('2d')

	resetCanvas(canvas)
	
	let count = 1000

	for (let i = 0; i < count; i++) {

		let x = width * (Math.random() - .5)
		let y = width * (Math.random() - .5)
		
		ctx.beginPath()
		ctx.ellipse(x, y, 1, 1, 0, 0, 2 * Math.PI)
		
		if (contains(triangle1, { x, y })) {
		
			ctx.fillStyle = '#f00'
		
		} else if (contains(triangle2, { x, y })) {
		
			ctx.fillStyle = '#00f'
		
		} else {
		
			ctx.fillStyle = '#0003'
		
		}

		
		ctx.fill()
		
	}
	
}

function drawInteractive() {

	let canvas = document.querySelector('canvas#interactive')
	let ctx = canvas.getContext('2d')

	resetCanvas(canvas)
	
	ctx.beginPath()
	ctx.moveTo(0, -height/2)
	ctx.lineTo(0, height/2)
	ctx.moveTo(-width/2, 0)
	ctx.lineTo(width/2, 0)
	ctx.strokeStyle = '#0003'
	ctx.stroke()
	
	drawTriangle(ctx, triangle1)
	drawTriangle(ctx, triangle2)
	
	ctx.beginPath()
	ctx.ellipse(mouse.x, mouse.y, 4, 4, 0, 0, 2 * Math.PI)
	
	if (contains(triangle1, mouse)) {
	
		ctx.fillStyle = triangle1.color + 'a'
		ctx.fill()
		
	} else if (contains(triangle2, mouse)) {
	
		ctx.fillStyle = triangle2.color + 'a'
		ctx.fill()
		
	} else {
	
		ctx.strokeStyle = 'black'
		ctx.stroke()
		
	}
	
}

drawDots()
drawInteractive()










// trigo

function add(...points) {
	
	let x = 0, y = 0
	
	for (let point of points) {
	
		x += point.x
		y += point.y
	
	}
	
	return { x, y }

}

function center(...points) {
	
	let x = 0, y = 0
	
	for (let point of points) {
	
		x += point.x
		y += point.y
	
	}
	
	x /= points.length
	y /= points.length
	
	return { x, y }

}

function sub(A, B) {

	let x = A.x - B.x
	let y = A.y - B.y
	
	return { x, y }

}

function normalize({ x, y }, length = 10) {

	let r = length / Math.sqrt(x * x + y * y)
	
	x *= r
	y *= r
	
	return { x, y }

}

function rotate({ x, y }, angle = 90) {

	let length = Math.sqrt(x * x + y * y)
	
	angle *= Math.PI / 180
	angle += Math.atan2(y, x)
	
	x = length * Math.cos(angle)
	y = length * Math.sin(angle)
	
	return { x, y }

}

* {
	margin: 0;
}

html {
	font-family: monospace;
}

body {
	padding: 32px;
}

span.red {
	color: #f00;
}

span.blue {
	color: #00f;
}

canvas {
	position: absolute;
	border: solid 1px #ffffd;
}

<p><span class="red">red triangle</span> is clockwise</p>
<p><span class="blue">blue triangle</span> is couter clockwise</p>
<br>
<div class="wrapper">
	<canvas id="dots"></canvas>
	<canvas id="interactive"></canvas>
</div>

I'm using here the same method as described above: a point is inside ABC if it is respectively on the "same" side of each line AB, BC, CA.

Answer 6

One of the easiest ways to check if the area formed by the vertices of triangle (x1,y1),(x2,y2),(x3,y3) is positive or not.

Area can by calculated by the formula:

1/2 [x1(y2–y3) + x2(y3–y1) + x3(y1–y2)]

or python code can be written as:

def triangleornot(p1,p2,p3):
    return (1/ 2) [p1[0](p2[1]–p3[1]) + p2[0] (p3[1]–p1[1]) + p3[0] (p1[0]–p2[0])]