android BitmapFactory的OutOfMemoryError: bitmap ...

来源：互联网发布：php 正则验证身份证号编辑：程序博客网时间：2024/06/07 11:23

原文地址：http://my.oschina.net/jeffzhao/blog/80900

网上有很多解决android加载bitmap内存溢出的方法，搜了一圈做下整理总结。项目里需求是拍摄多图之后上传，部分手机会内存溢出。

常用一种解决方法：即将载入的图片缩小，这种方式以牺牲图片的质量为代价。在BitmapFactory中有一个内部类BitmapFactory.Options，其中当options.inSampleSize值>1时，根据文档：

If set to a value > 1, requests the decoder to subsample the original image, returning a smaller image to save memory. (1 -> decodes full size; 2 -> decodes 1/4th size; 4 -> decode 1/16th size). Because you rarely need to show and have full size bitmap images on your phone. For manipulations smaller sizes are usually enough.
options.inSampleSize是以2的指数的倒数被进行放缩

现在问题是怎么确定inSampleSize的值？每张图片的放缩大小的比例应该是不一样的！这样的话就要运行时动态确定。在BitmapFactory.Options中提供了另一个成员inJustDecodeBounds。
设置inJustDecodeBounds为true后，decodeFile并不分配空间，但可计算出原始图片的长度和宽度，即opts.width和opts.height。有了这两个参数，再通过一定的算法，即可得到一个恰当的inSampleSize。Android提供了一种动态计算的方法，见computeSampleSize().

01publicstatic int computeSampleSize(BitmapFactory.Options options,
02        intminSideLength, int maxNumOfPixels) {
03    intinitialSize = computeInitialSampleSize(options, minSideLength,
04            maxNumOfPixels);
05  
06    introundedSize;
07    if(initialSize <= 8) {
08        roundedSize =1;
09        while(roundedSize < initialSize) {
10            roundedSize <<=1;
11        }
12    }else {
13        roundedSize = (initialSize +7) / 8* 8;
14    }
15  
16    returnroundedSize;
17}
18  
19privatestatic int computeInitialSampleSize(BitmapFactory.Options options,
20        intminSideLength, int maxNumOfPixels) {
21    doublew = options.outWidth;
22    doubleh = options.outHeight;
23  
24    intlowerBound = (maxNumOfPixels == -1) ?1 :
25            (int) Math.ceil(Math.sqrt(w * h / maxNumOfPixels));
26    intupperBound = (minSideLength == -1) ?128 :
27            (int) Math.min(Math.floor(w / minSideLength),
28            Math.floor(h / minSideLength));
29  
30    if(upperBound < lowerBound) {
31        returnlowerBound;
32    }
33  
34    if((maxNumOfPixels == -1) &&
35            (minSideLength == -1)) {
36        return1;
37    }else if (minSideLength == -1) {
38        returnlowerBound;
39    }else {
40        returnupperBound;
41    }
42}

以上只做为参考，我们只要用这函数即可，opts.inSampleSize = computeSampleSize(opts, -1,128*128);

要点：
1、用decodeFileDescriptor()来生成bimap比decodeFile()省内存

1FileInputStream is = =new FileInputStream(path);
2bmp = BitmapFactory.decodeFileDescriptor(is.getFD(),null, opts);

替换

1Bitmap bmp = BitmapFactory.decodeFile(imageFile, opts);
2   imageView.setImageBitmap(bmp);

原因：
查看BitmapFactory的源码，对比一下两者的实现，可以发现decodeFile()最终是以流的方式生成bitmap

decodeFile源码：

01publicstatic Bitmap decodeFile(String pathName, Options opts) {
02    Bitmap bm =null;
03    InputStream stream =null;
04    try{
05        stream =new FileInputStream(pathName);
06        bm = decodeStream(stream,null, opts);
07    }catch (Exception e) {
08        /*  do nothing.
09            If the exception happened on open, bm will be null.
10        */
11    }finally {
12        if(stream != null) {
13            try{
14                stream.close();
15            }catch (IOException e) {
16                // do nothing here
17            }
18        }
19    }
20    returnbm;
21}

decodeFileDescriptor的源码，可以找到native本地方法decodeFileDescriptor，通过底层生成bitmap

decodeFileDescriptor源码：

01   publicstatic Bitmap decodeFileDescriptor(FileDescriptor fd, Rect outPadding, Options opts) {
02       if(nativeIsSeekable(fd)) {
03           Bitmap bm = nativeDecodeFileDescriptor(fd, outPadding, opts);
04           if(bm == null && opts != null && opts.inBitmap !=null) {
05               thrownew IllegalArgumentException("Problem decoding into existing bitmap");
06           }
07           returnfinishDecode(bm, outPadding, opts);
08       }else {
09           FileInputStream fis =new FileInputStream(fd);
10           try{
11               returndecodeStream(fis, outPadding, opts);
12           }finally {
13               try{
14                   fis.close();
15               }catch (Throwable t) {/* ignore */}
16           }
17       }
18   }
19
20privatestatic native Bitmap nativeDecodeFileDescriptor(FileDescriptor fd,Rect padding, Options opts);

2、当在android设备中载入较大图片资源时，可以创建一些临时空间，将载入的资源载入到临时空间中。

1opts.inTempStorage =new byte[16* 1024];

完整代码：

01publicstatic OutputStream decodeBitmap(String path) {

02

03 BitmapFactory.Options opts =new BitmapFactory.Options();

04 opts.inJustDecodeBounds =true;// 设置成了true,不占用内存，只获取bitmap宽高

05 BitmapFactory.decodeFile(path, opts);

06 opts.inSampleSize = computeSampleSize(opts, -1,1024 * 800);

07

08 opts.inJustDecodeBounds =false;// 这里一定要将其设置回false，因为之前我们将其设置成了true

09 opts.inPurgeable =true;

10 opts.inInputShareable =true;

11 opts.inDither =false;

12 opts.inPurgeable =true;

13 opts.inTempStorage =new byte[16* 1024];

14 FileInputStream is =null;

15 Bitmap bmp =null;

16 InputStream ins =null;

17 ByteArrayOutputStream baos =null;

18 try{

19 is =new FileInputStream(path);

20 bmp = BitmapFactory.decodeFileDescriptor(is.getFD(),null, opts); double scale = getScaling(opts.outWidth * opts.outHeight,1024 * 600);

21 Bitmap bmp2 = Bitmap.createScaledBitmap(bmp,

22 (int) (opts.outWidth * scale),

23 (int) (opts.outHeight * scale),true);

24 bmp.recycle();

25 baos =new ByteArrayOutputStream();

26 bmp2.compress(Bitmap.CompressFormat.JPEG,100, baos);

27 bmp2.recycle();

28 returnbaos;

29 }catch (FileNotFoundException e) {

30 e.printStackTrace();

31 }catch (IOException e) {

32 e.printStackTrace();

33 }finally {

34 try{

35 is.close();

36 ins.close();

37 baos.close();

38 }catch (IOException e) {

39 e.printStackTrace();

40 }

41 System.gc();

42 }

43 returnbaos;

44 }

45

46privatestatic double getScaling(int src, int des) {

47/**

48 * 目标尺寸÷原尺寸 sqrt开方，得出宽高百分比

49 */

50 doublescale = Math.sqrt((double) des / (double) src);

51 returnscale;

52}

0 0