操作图像的内存有效方式?

Posted

技术标签:

【中文标题】操作图像的内存有效方式?【英文标题】:Memory efficient way to manipulate images? 【发布时间】:2016-08-23 17:05:12 【问题描述】:

我有一些代码可以获取两张图片并将它们的中间部分合并为一张图片。该代码有效,但使用了大量内存,因此在某些内存不足的设备上失败。

有没有办法让这段代码更节省内存?

代码

private void convertImages()
    ImageView imageView = (ImageView) findViewById(R.id.imageView);

    File leftFile = new File(Environment.getExternalStorageDirectory(), "images/left.jpg");
    File rightFile = new File(Environment.getExternalStorageDirectory(), "images/right.jpg");

    Bitmap left = BitmapFactory.decodeFile(leftFile.getAbsolutePath());
    Bitmap right = BitmapFactory.decodeFile(rightFile.getAbsolutePath());

    Rect srcRect = new Rect( (int)(left.getWidth()*0.25), 0, (int)(left.getWidth()*0.75), left.getHeight() );
    Rect dstRectLeft = new Rect( 0, 0, (int)(srcRect.width()/2), srcRect.height() );
    Rect dstRectRight = new Rect( (int)(srcRect.width()/2), 0, srcRect.width(), srcRect.height() );

    Bitmap outBitmap = Bitmap.createBitmap(srcRect.width(), srcRect.height(), Bitmap.Config.ARGB_8888);
    Canvas outCanvas = new Canvas(outBitmap);
    outCanvas.drawBitmap(left, srcRect, dstRectLeft, null);
    outCanvas.drawBitmap(right, srcRect, dstRectRight, null);

    imageView.setImageBitmap( outBitmap );

【问题讨论】:

这两个源图像似乎保证具有相同的尺寸。因此,您可以加载第一个图像来构建矩形,然后释放图像(通过将其设置为 null 然后调用 gc)。然后再次加载图像,但以一对一的方式(即在每次使用后释放图像)。然后将每个图像裁剪到新的尺寸,在每个步骤中将它们中的每一个写入一个文件。最后,加载裁剪后的图像以创建最终图像。 您可以通过以下操作简单地删除内存中的一张图像:1)向左加载 2)向左绘制以输出 3)向右加载,4)向右绘制以输出。 【参考方案1】:

您可以尝试压缩。试试这个有用的图像压缩器类,为了您的方便,复制并粘贴到这个线程中:

android: Compressing images creates black borders on left and top margin

压缩器 java 会将您的图片压缩到其大小的 1/10 左右,因此对于 3mb 的图片,它将变为 300kb。希望这会为您节省一些记忆。

import android.content.Context;
import android.graphics.Bitmap;
import android.graphics.BitmapFactory;
import android.graphics.Canvas;
import android.graphics.Matrix;
import android.graphics.Paint;
import android.media.ExifInterface;

import java.io.FileNotFoundException;
import java.io.FileOutputStream;
import java.io.IOException;

import timber.log.Timber;

//http://voidcanvas.com/whatsapp-like-image-compression-in-android/
public class ImageCompressor 

    public ImageCompressor() 

    public static String compressImage(String imagePath, Context context) 
        Bitmap scaledBitmap = null;
        String filename = "compressed_" +imagePath.substring(imagePath.lastIndexOf("/")+1);

        BitmapFactory.Options options = new BitmapFactory.Options();

//      by setting this field as true, the actual bitmap pixels are not loaded in the memory. Just the bounds are loaded. If
//      you try the use the bitmap here, you will get null.
        options.inJustDecodeBounds = true;
        Timber.e( "imagePath "+imagePath);
        Timber.e("filename "+filename);
        Bitmap bmp = BitmapFactory.decodeFile(imagePath, options);
        if (options == null) 
            Timber.e("zero bitmap");
        
        int actualHeight = options.outHeight;
        int actualWidth = options.outWidth;
        float imgRatio = actualWidth / actualHeight;

        float maxHeight = actualHeight * 10/20;
        float maxWidth = actualWidth * 10/20;
        float maxRatio = maxWidth / maxHeight;

//      width and height values are set maintaining the aspect ratio of the image

        if (actualHeight > maxHeight || actualWidth > maxWidth) 
            if (imgRatio < maxRatio) 

                imgRatio = maxHeight / actualHeight;
                actualWidth = (int) (imgRatio * actualWidth);
                actualHeight = (int) maxHeight;

             else if (imgRatio > maxRatio) 

                imgRatio = maxWidth / actualWidth;
                actualHeight = (int) (imgRatio * actualHeight);
                actualWidth = (int) maxWidth;

             else 

                actualHeight = (int) maxHeight;
                actualWidth = (int) maxWidth;

            
        

//      setting inSampleSize value allows to load a scaled down version of the original image

        options.inSampleSize = calculateInSampleSize(options, actualWidth, actualHeight);

//      inJustDecodeBounds set to false to load the actual bitmap
        options.inJustDecodeBounds = false;

//      this options allow android to claim the bitmap memory if it runs low on memory
        options.inPurgeable = true;
        options.inInputShareable = true;
        options.inTempStorage = new byte[16 * 1024];

        try 
//          load the bitmap from its path
            bmp = BitmapFactory.decodeFile(imagePath, options);
         catch (OutOfMemoryError exception) 
            exception.printStackTrace();

        
        try 
            scaledBitmap = Bitmap.createBitmap(actualWidth, actualHeight, Bitmap.Config.ARGB_8888);
         catch (OutOfMemoryError exception) 
            exception.printStackTrace();
        

        float ratioX = actualWidth / (float) options.outWidth;
        float ratioY = actualHeight / (float) options.outHeight;
        float middleX = actualWidth / 2.0f;
        float middleY = actualHeight / 2.0f;

        Matrix scaleMatrix = new Matrix();
        scaleMatrix.setScale(ratioX, ratioY, middleX, middleY);

        Canvas canvas = new Canvas(scaledBitmap);
        canvas.setMatrix(scaleMatrix);
        canvas.drawBitmap(bmp, middleX - bmp.getWidth() / 2, middleY - bmp.getHeight() / 2, new Paint(Paint.FILTER_BITMAP_FLAG));

//      check the rotation of the image and display it properly
        ExifInterface exif;
        try 
            exif = new ExifInterface(imagePath);

            int orientation = exif.getAttributeInt(
                    ExifInterface.TAG_ORIENTATION, 0);
            Timber.e("Exif: " + orientation);
            Matrix matrix = new Matrix();
            if (orientation == 6) 
                matrix.postRotate(90);
                Timber.e( "Exif: " + orientation);
             else if (orientation == 3) 
                matrix.postRotate(180);
                Timber.e( "Exif: " + orientation);
             else if (orientation == 8) 
                matrix.postRotate(270);
                Timber.e( "Exif: " + orientation);
            
            scaledBitmap = Bitmap.createBitmap(scaledBitmap, 0, 0,
                    scaledBitmap.getWidth(), scaledBitmap.getHeight(), matrix,
                    true);
         catch (IOException e) 
            e.printStackTrace();
         catch (NullPointerException e) 
            e.printStackTrace();
        

        FileOutputStream out = null;

        try 

            out = context.openFileOutput(filename, Context.MODE_PRIVATE);
            scaledBitmap.compress(Bitmap.CompressFormat.JPEG, 90, out);

            out.close();

         catch (FileNotFoundException e) 
            e.printStackTrace();
         catch (IOException e) 
            e.printStackTrace();
        
            return filename;
        

    public static int calculateInSampleSize(BitmapFactory.Options options, int reqWidth, int reqHeight) 
        final int height = options.outHeight;
        final int width = options.outWidth;
        int inSampleSize = 1;
        if (height > reqHeight || width > reqWidth) 
            final int heightRatio = Math.round((float) height/ (float) reqHeight);
            final int widthRatio = Math.round((float) width / (float) reqWidth);
            inSampleSize = heightRatio < widthRatio ? heightRatio : widthRatio;      
        final float totalPixels = width * height;
        final float totalReqPixelsCap = reqWidth * reqHeight * 2;
        while (totalPixels / (inSampleSize * inSampleSize) > totalReqPixelsCap) 
            inSampleSize++;
        
        return inSampleSize;
    



【讨论】:

似乎有效,但它也会将图像大小调整为宽度/高度的一半? 是的。但是如果你把代码中的 10/20 去掉,它不会调整一半。【参考方案2】:

您可以使用 Bitmap.createScaledBitmap(src, dstWidth, dstHeight, filter);缩小位图的方法,使其尺寸更小

【讨论】:

以上是关于操作图像的内存有效方式?的主要内容,如果未能解决你的问题,请参考以下文章

如何以编程方式有效地平移整个图像?

如何在 QQuickPaintedItem 中以有效的方式绘制顺序图像

通过网络连续传输图像的最有效方式

有效地矢量化图像块处理?

如何以内存有效的方式将大量文件添加到 zip 中?

如何以内存有效的方式在python中将矩阵一分为二?