layout: post
title: Android 馬賽克(Mosaics)效果
date: 2015-12-13
categories: blog
tags: [技術(shù),Android]
category: Android
description: 使用兩種不同的方案實現(xiàn)Android平臺下的馬賽克效果

本文首發(fā)于KuTear,轉(zhuǎn)載請注明
代碼地址
GitHub

前幾天看見開源項目[ExplosionField]效果好贊,看了下代碼,他的實現(xiàn)大致就是在原界面之上覆蓋一成自定義的View,獲取到點擊的那個View的內(nèi)容(Bitmap),然后在覆蓋的那個自定義View的特定位置畫出來,之后就是對這個Bitmap做一些列拆分,變化重繪的過程.在這里根據(jù)他對bitmap的拆分,感覺用來實現(xiàn)Bitmap的效果也是不錯的,就試著做一做.

在這里介紹使用兩種方式實現(xiàn)馬賽克效果.開始之前先看看效果

Screenshot from 2015-12-13 19-09-50.png

感覺還可以

1. 直接繪制

   public static Bitmap getMosaicsBitmap(Bitmap bmp, double precent) {
   long start = System.currentTimeMillis();
   int bmpW = bmp.getWidth();
   int bmpH = bmp.getHeight();
   Bitmap resultBmp = Bitmap.createBitmap(bmpW, bmpH, Bitmap.Config.ARGB_8888);
   Canvas canvas = new Canvas(resultBmp);
   Paint paint = new Paint();
   double unit;
   if (precent == 0) {
       unit = bmpW;
   } else {
       unit = 1 / precent;
   }
   double resultBmpW = bmpW / unit;
   double resultBmpH = bmpH / unit;
   for (int i = 0; i < resultBmpH; i++) {
       for (int j = 0; j < resultBmpW; j++) {
           int pickPointX = (int) (unit * (j + 0.5));
           int pickPointY = (int) (unit * (i + 0.5));
           int color;
           if (pickPointX >= bmpW || pickPointY >= bmpH) {
               color = bmp.getPixel(bmpW / 2, bmpH / 2);
           } else {
               color = bmp.getPixel(pickPointX, pickPointY);
           }
           paint.setColor(color);
           canvas.drawRect((int) (unit * j), (int) (unit * i), (int) (unit * (j + 1)), (int) (unit * (i + 1)), paint);
       }
   }
   canvas.setBitmap(null);
   long end = System.currentTimeMillis();
   Log.v(TAG, "DrawTime:" + (end - start));
   return resultBmp;
   } 
   

2. 修改像素點

    public static Bitmap getMosaicsBitmaps(Bitmap bmp, double precent) {
    long start = System.currentTimeMillis();
    int bmpW = bmp.getWidth();
    int bmpH = bmp.getHeight();
    int[] pixels = new int[bmpH * bmpW];
    bmp.getPixels(pixels, 0, bmpW, 0, 0, bmpW, bmpH);
    int raw = (int) (bmpW * precent);
    int unit;
    if (raw == 0) {
        unit = bmpW;
    } else {
        unit = bmpW / raw; //原來的unit*unit像素點合成一個，使用原左上角的值
    }
    if (unit >= bmpW || unit >= bmpH) {
        return getMosaicsBitmap(bmp, precent);
    }
    for (int i = 0; i < bmpH; ) {
        for (int j = 0; j < bmpW; ) {
            int leftTopPoint = i * bmpW + j;
            for (int k = 0; k < unit; k++) {
                for (int m = 0; m < unit; m++) {
                    int point = (i + k) * bmpW + (j + m);
                    if (point < pixels.length) {
                        pixels[point] = pixels[leftTopPoint];
                    }
                }
            }
            j += unit;
        }
        i += unit;
    }
    long end = System.currentTimeMillis();
    Log.v(TAG, "DrawTime:" + (end - start));
    return Bitmap.createBitmap(pixels, bmpW, bmpH, Bitmap.Config.ARGB_8888);
}

從效率上來看,第二中方式效率會高10倍,只要是因為第一種方式繪制的次數(shù)太多了,而繪制是比較費(fèi)時間的.這里,特別提示,不要在大量的循環(huán)語句內(nèi)部使用Log.v(...),這是一個很耗時間的操作.