Javascript canvas 的像素操作,模糊滤镜不起作用
Javascript pixel manipulation of canvas, Blur filter not working
我目前正在 canvas 开发一个简单的绘图板网站。
http://webbox.cs.du.edu/~ansbashe/My_Art/myart.html
我想用这个 javascript 代码(在 webworker 中)模糊 canvas。问题是返回的数据用原始图片填充 canvas 而不是我期望的模糊图像。我可以做些什么来获得我想要的结果吗?
var filter = [ [1, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 1, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 1, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 1, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 1, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 1, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 1, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 1, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 1]];
var factor = 1/9;
var bias = 0;
var w = 890;
var h = 400;
function filts() {
filt = 0;
for(var x = 0; x < w; x++)
{
for(var y = 0; y < h; y++)
{
var r = 0;
var g = 0;
var b = 0;
var imgX = ((x - Math.floor(filter.length/2)+filterX + w) % w);
var imgY = ((y - Math.floor(filter.length/2)+filterY + h)%h);
for(var filterX = 0; filterX < filter.length; filterX++) {
for(var filterY =0; filterY < filter.length; filterY++) {
r += imgData.data[imgX*4 + imgY * w * 4] * filter[filterY][filterX];
g += imgData.data[imgX*4 + imgY*w*4 + 1] * filter[filterY][filterX];
b += imgData.data[imgX*4 + imgY*w*4 + 2] * filter[filterY][filterX];
}
}
var index = (x + (y * w)) * 4;
retData.data[imgX*4 + imgY*w*4] = Math.min(Math.max(Math.floor(factor *r + bias),0),255);
retData.data[imgX*4 + imgY*w*4 + 1] = Math.min(Math.max(Math.floor(factor *g + bias),0),255);
retData.data[imgX*4 + imgY*w*4 + 2] = Math.min(Math.max(Math.floor(factor *b + bias),0),255);
retData.data[imgX*4 + imgY*w*4 + 3] = imgData.data[imgX*4 + imgY*w*4 + 3];
}
}
postMessage({img: retData});
}
感谢您的宝贵时间。
您需要偏移您从过滤器循环中读取的坐标,否则您每次都会从相同的像素读取(imgX,imgY):
for(var filterX = 0; filterX < filter.length; filterX++) {
var filterImgX = (imgX + filterX) % w;
for(var filterY =0; filterY < filter.length; filterY++) {
var filterImgY = (imgY + filterY) % h;
r += imgData.data[filterImgX*4 + filterImgY * w * 4] * filter[filterY][filterX];
g += imgData.data[filterImgX*4 + filterImgY*w*4 + 1] * filter[filterY][filterX];
b += imgData.data[filterImgX*4 + filterImgY*w*4 + 2] * filter[filterY][filterX];
}
}
我目前正在 canvas 开发一个简单的绘图板网站。
http://webbox.cs.du.edu/~ansbashe/My_Art/myart.html
我想用这个 javascript 代码(在 webworker 中)模糊 canvas。问题是返回的数据用原始图片填充 canvas 而不是我期望的模糊图像。我可以做些什么来获得我想要的结果吗?
var filter = [ [1, 0, 0, 0, 0, 0, 0, 0, 0],
[0, 1, 0, 0, 0, 0, 0, 0, 0],
[0, 0, 1, 0, 0, 0, 0, 0, 0],
[0, 0, 0, 1, 0, 0, 0, 0, 0],
[0, 0, 0, 0, 1, 0, 0, 0, 0],
[0, 0, 0, 0, 0, 1, 0, 0, 0],
[0, 0, 0, 0, 0, 0, 1, 0, 0],
[0, 0, 0, 0, 0, 0, 0, 1, 0],
[0, 0, 0, 0, 0, 0, 0, 0, 1]];
var factor = 1/9;
var bias = 0;
var w = 890;
var h = 400;
function filts() {
filt = 0;
for(var x = 0; x < w; x++)
{
for(var y = 0; y < h; y++)
{
var r = 0;
var g = 0;
var b = 0;
var imgX = ((x - Math.floor(filter.length/2)+filterX + w) % w);
var imgY = ((y - Math.floor(filter.length/2)+filterY + h)%h);
for(var filterX = 0; filterX < filter.length; filterX++) {
for(var filterY =0; filterY < filter.length; filterY++) {
r += imgData.data[imgX*4 + imgY * w * 4] * filter[filterY][filterX];
g += imgData.data[imgX*4 + imgY*w*4 + 1] * filter[filterY][filterX];
b += imgData.data[imgX*4 + imgY*w*4 + 2] * filter[filterY][filterX];
}
}
var index = (x + (y * w)) * 4;
retData.data[imgX*4 + imgY*w*4] = Math.min(Math.max(Math.floor(factor *r + bias),0),255);
retData.data[imgX*4 + imgY*w*4 + 1] = Math.min(Math.max(Math.floor(factor *g + bias),0),255);
retData.data[imgX*4 + imgY*w*4 + 2] = Math.min(Math.max(Math.floor(factor *b + bias),0),255);
retData.data[imgX*4 + imgY*w*4 + 3] = imgData.data[imgX*4 + imgY*w*4 + 3];
}
}
postMessage({img: retData});
}
感谢您的宝贵时间。
您需要偏移您从过滤器循环中读取的坐标,否则您每次都会从相同的像素读取(imgX,imgY):
for(var filterX = 0; filterX < filter.length; filterX++) {
var filterImgX = (imgX + filterX) % w;
for(var filterY =0; filterY < filter.length; filterY++) {
var filterImgY = (imgY + filterY) % h;
r += imgData.data[filterImgX*4 + filterImgY * w * 4] * filter[filterY][filterX];
g += imgData.data[filterImgX*4 + filterImgY*w*4 + 1] * filter[filterY][filterX];
b += imgData.data[filterImgX*4 + filterImgY*w*4 + 2] * filter[filterY][filterX];
}
}