这是一个非常好的教程-

http://www.html5canvastutorials.com/advanced/html5-canvas-mouse-coordinates/

 <canvas id="myCanvas" width="578" height="200"></canvas>
<script>
  function writeMessage(canvas, message) {
    var context = canvas.getContext('2d');
    context.clearRect(0, 0, canvas.width, canvas.height);
    context.font = '18pt Calibri';
    context.fillStyle = 'black';
    context.fillText(message, 10, 25);
  }
  function getMousePos(canvas, evt) {
    var rect = canvas.getBoundingClientRect();
    return {
      x: evt.clientX - rect.left,
      y: evt.clientY - rect.top
    };
  }
  var canvas = document.getElementById('myCanvas');
  var context = canvas.getContext('2d');

  canvas.addEventListener('mousemove', function(evt) {
    var mousePos = getMousePos(canvas, evt);
    var message = 'Mouse position: ' + mousePos.x + ',' + mousePos.y;
    writeMessage(canvas, message);
  }, false);

希望这能有所帮助!

下面是对Ryan Artecona关于画布宽度可变(%)的答案的一个小修改:

 HTMLCanvasElement.prototype.relMouseCoords = function (event) {
    var totalOffsetX = 0;
    var totalOffsetY = 0;
    var canvasX = 0;
    var canvasY = 0;
    var currentElement = this;

    do {
        totalOffsetX += currentElement.offsetLeft;
        totalOffsetY += currentElement.offsetTop;
    }
    while (currentElement = currentElement.offsetParent)

    canvasX = event.pageX - totalOffsetX;
    canvasY = event.pageY - totalOffsetY;

    // Fix for variable canvas width
    canvasX = Math.round( canvasX * (this.width / this.offsetWidth) );
    canvasY = Math.round( canvasY * (this.height / this.offsetHeight) );

    return {x:canvasX, y:canvasY}
}

我不确定所有这些遍历父元素和做各种奇怪事情的答案的意义是什么。

HTMLElement。getBoundingClientRect方法被设计用来处理任何元素的实际屏幕位置。这包括滚动，所以不需要像scrollTop这样的东西:

(来自MDN)已经完成的视口区域(或 的时候，任何其他可滚动的元素)都被考虑在内 边界矩形

正常的图片

最简单的方法已经贴在这里了。只要不涉及宽泛的CSS规则，这是正确的。

处理拉伸的画布/图像

当图像像素宽度与它的CSS宽度不匹配时，你需要对像素值应用一些比率:

/* Returns pixel coordinates according to the pixel that's under the mouse cursor**/
HTMLCanvasElement.prototype.relativeCoords = function(event) {
  var x,y;
  //This is the current screen rectangle of canvas
  var rect = this.getBoundingClientRect();
  var top = rect.top;
  var bottom = rect.bottom;
  var left = rect.left;
  var right = rect.right;
  //Recalculate mouse offsets to relative offsets
  x = event.clientX - left;
  y = event.clientY - top;
  //Also recalculate offsets of canvas is stretched
  var width = right - left;
  //I use this to reduce number of calculations for images that have normal size 
  if(this.width!=width) {
    var height = bottom - top;
    //changes coordinates by ratio
    x = x*(this.width/width);
    y = y*(this.height/height);
  } 
  //Return as an array
  return [x,y];
}

只要画布没有边界，它就适用于拉伸图像(jsFiddle)。

处理CSS边框

如果画布的边框很厚，事情就会变得有点复杂。你需要从边界矩形中减去边框。这可以使用. getcomputedstyle来完成。这个答案描述了这个过程。

然后函数增大一点:

/* Returns pixel coordinates according to the pixel that's under the mouse cursor**/
HTMLCanvasElement.prototype.relativeCoords = function(event) {
  var x,y;
  //This is the current screen rectangle of canvas
  var rect = this.getBoundingClientRect();
  var top = rect.top;
  var bottom = rect.bottom;
  var left = rect.left;
  var right = rect.right;
  //Subtract border size
  // Get computed style
  var styling=getComputedStyle(this,null);
  // Turn the border widths in integers
  var topBorder=parseInt(styling.getPropertyValue('border-top-width'),10);
  var rightBorder=parseInt(styling.getPropertyValue('border-right-width'),10);
  var bottomBorder=parseInt(styling.getPropertyValue('border-bottom-width'),10);
  var leftBorder=parseInt(styling.getPropertyValue('border-left-width'),10);
  //Subtract border from rectangle
  left+=leftBorder;
  right-=rightBorder;
  top+=topBorder;
  bottom-=bottomBorder;
  //Proceed as usual
  ...
}

我想不出有什么会混淆最后这个函数。在JsFiddle见。

笔记

如果您不喜欢修改本机原型，只需更改函数并使用(canvas, event)调用它(并将任何此替换为canvas)。

Edit 2018:这个答案很老了，它使用检查不再需要的旧浏览器，因为clientX和clienti属性在所有当前的浏览器中都有效。您可能想要查看Patriques Answer，以获得一个更简单、更近期的解决方案。

最初的回答: 正如我当时发现的一篇文章所描述的那样，但现在已经不存在了:

var x;
var y;
if (e.pageX || e.pageY) { 
  x = e.pageX;
  y = e.pageY;
}
else { 
  x = e.clientX + document.body.scrollLeft + document.documentElement.scrollLeft; 
  y = e.clientY + document.body.scrollTop + document.documentElement.scrollTop; 
} 
x -= gCanvasElement.offsetLeft;
y -= gCanvasElement.offsetTop;

对我来说效果很好。

根据最新的Quirksmode，所有主流浏览器都支持clientX和clientY方法。 所以，它开始了-在滚动条页面上的滚动div中工作的良好工作代码:

function getCursorPosition(canvas, event) {
var x, y;

canoffset = $(canvas).offset();
x = event.clientX + document.body.scrollLeft + document.documentElement.scrollLeft - Math.floor(canoffset.left);
y = event.clientY + document.body.scrollTop + document.documentElement.scrollTop - Math.floor(canoffset.top) + 1;

return [x,y];
}

这也需要jQuery $(canvas).offset()。

更新(5/5/16):应该使用patriques的答案，因为它既简单又可靠。

Since the canvas isn't always styled relative to the entire page, the canvas.offsetLeft/Top doesn't always return what you need. It will return the number of pixels it is offset relative to its offsetParent element, which can be something like a div element containing the canvas with a position: relative style applied. To account for this you need to loop through the chain of offsetParents, beginning with the canvas element itself. This code works perfectly for me, tested in Firefox and Safari but should work for all.

function relMouseCoords(event){
    var totalOffsetX = 0;
    var totalOffsetY = 0;
    var canvasX = 0;
    var canvasY = 0;
    var currentElement = this;

    do{
        totalOffsetX += currentElement.offsetLeft - currentElement.scrollLeft;
        totalOffsetY += currentElement.offsetTop - currentElement.scrollTop;
    }
    while(currentElement = currentElement.offsetParent)

    canvasX = event.pageX - totalOffsetX;
    canvasY = event.pageY - totalOffsetY;

    return {x:canvasX, y:canvasY}
}
HTMLCanvasElement.prototype.relMouseCoords = relMouseCoords;

最后一行可以方便地获取相对于canvas元素的鼠标坐标。要得到有用的坐标只需要

coords = canvas.relMouseCoords(event);
canvasX = coords.x;
canvasY = coords.y;