根据fabricjs模糊算法，后台用java合成同样效果的模糊效果

1. 根据fabricjs模糊算法，后台用java合成同样效果的模糊效果，其中fabricjs模糊算法为：

  simpleBlur: function(options) {
        var resources = options.filterBackend.resources, canvas1, canvas2,
            width = options.imageData.width,
            height = options.imageData.height;

        if (!resources.blurLayer1) {
          resources.blurLayer1 = fabric.util.createCanvasElement();
          resources.blurLayer2 = fabric.util.createCanvasElement();
        }
        canvas1 = resources.blurLayer1;
        canvas2 = resources.blurLayer2;
        if (canvas1.width !== width || canvas1.height !== height) {
          canvas2.width = canvas1.width = width;
          canvas2.height = canvas1.height = height;
        }
        var ctx1 = canvas1.getContext('2d'),
            ctx2 = canvas2.getContext('2d'),
            nSamples = 15,
            random, percent, j, i,
            blur = this.blur * 0.06 * 0.5;

        // load first canvas
        ctx1.putImageData(options.imageData, 0, 0);
        ctx2.clearRect(0, 0, width, height);

        for (i = -nSamples; i <= nSamples; i++) {
          random = (Math.random() - 0.5) / 4;
          percent = i / nSamples;
          j = blur * percent * width + random;
          ctx2.globalAlpha = 1 - Math.abs(percent);
          ctx2.drawImage(canvas1, j, random);
          ctx1.drawImage(canvas2, 0, 0);
          ctx2.globalAlpha = 1;
          ctx2.clearRect(0, 0, canvas2.width, canvas2.height);
        }
        for (i = -nSamples; i <= nSamples; i++) {
          random = (Math.random() - 0.5) / 4;
          percent = i / nSamples;
          j = blur * percent * height + random;
          ctx2.globalAlpha = 1 - Math.abs(percent);
          ctx2.drawImage(canvas1, random, j);
          ctx1.drawImage(canvas2, 0, 0);
          ctx2.globalAlpha = 1;
          ctx2.clearRect(0, 0, canvas2.width, canvas2.height);
        }
        options.ctx.drawImage(canvas1, 0, 0);
        var newImageData = options.ctx.getImageData(0, 0, canvas1.width, canvas1.height);
        ctx1.globalAlpha = 1;
        ctx1.clearRect(0, 0, canvas1.width, canvas1.height);
        return newImageData;
      },

2. 这个图片模糊的算法是不同透明度图片叠加实现的，对应转化为java代码：

if(blur!=0){
			int nSamples = 15;
			float random = 0.0f;
			float percent = 0.0f;
			float j = 0.0f;
			blur = (float) (blur*0.06*0.5);
			BufferedImage image1 = new BufferedImage(width, height, BufferedImage.TYPE_INT_ARGB);		
			Graphics2D g1 = image1.createGraphics();			
			g1.drawImage(newImage,0,0,null); 
			//x上的处理
			for (int i = -nSamples; i <= nSamples; i++) {
		   	    BufferedImage image2 = new BufferedImage(width, height, BufferedImage.TYPE_INT_ARGB);
			     Graphics2D g2 = image2.createGraphics();	
			     random = (float) ((Math.random() - 0.5) / 4);
			     percent = (float) i / nSamples;
			     j = blur * percent * width + random;
			     float globalAlpha = 1 - Math.abs(percent);	
			     g2.setComposite(AlphaComposite.SrcOver.derive(globalAlpha));
			     g2.drawImage(image1,(int)j,(int)random,null);	
			     g1.drawImage(image2, 0, 0, null);
			     g2.dispose();			          
			      }
			//y上的处理
			for (int i = -nSamples; i <= nSamples; i++) {
			     BufferedImage image2 = new BufferedImage(width, height, BufferedImage.TYPE_INT_ARGB);
			     Graphics2D g2 = image2.createGraphics();
			     random = (float) ((Math.random() - 0.5) / 4);
			     percent = (float) i / nSamples;
			     j = blur * percent * height + random;
			     float globalAlpha = 1 - Math.abs(percent);
			     g2.setComposite(AlphaComposite.SrcOver.derive(globalAlpha));
			     g2.drawImage(image1,(int)random,(int)j,null);	    
				 g1.drawImage(image2, 0, 0, null);
				 g2.dispose();
			      }
			     g1.dispose();
		         newImage = image1;
		}

参考文章：http://www.jhlabs.com/ip/blurring.html