java bmp 包_Java: NV21转BMP文件

NV21是Android camera的默认图像格式。这里分享下如何在Windows上用Java把NV21的数据转换成BMP文件。

什么是NV21格式

NV21是一种YUV (又称YCbCr) 图像格式. Y 表示亮度值(Luminance)， V 和 U 分别代表蓝色(Cb)和红色(Cr)色度(Chrominance)。 一个2×2的像素, 4个Y对应1个V和1个U。比如这样排布:

YYYYYYYY VUVU

NV21转RGB

先从byte数组中分离出Y,V,U的值(先V，后U)：

int total = width * height;

int[] rgb = new int[total];

int Y, Cb = 0, Cr = 0, index = 0;

int R, G, B;

for (int y = 0; y < height; y++) {

for (int x = 0; x < width; x++) {

Y = yuv[y * width + x];

if (Y < 0) Y += 255;

if ((x & 1) == 0) {

Cr = yuv[(y >> 1) * (width) + x + total];

Cb = yuv[(y >> 1) * (width) + x + total + 1];

if (Cb < 0) Cb += 127; else Cb -= 128;

if (Cr < 0) Cr += 127; else Cr -= 128;

}

}

}

通过公式转换：

R = Y + Cr + (Cr >> 2) + (Cr >> 3) + (Cr >> 5);

G = Y - (Cb >> 2) + (Cb >> 4) + (Cb >> 5) - (Cr >> 1) + (Cr >> 3) + (Cr >> 4) + (Cr >> 5);

B = Y + Cb + (Cb >> 1) + (Cb >> 2) + (Cb >> 6);

// Approximation

// R = (int) (Y + 1.40200 * Cr);

// G = (int) (Y - 0.34414 * Cb - 0.71414 * Cr);

// B = (int) (Y + 1.77200 * Cb);

if (R < 0) R = 0; else if (R > 255) R = 255;

if (G < 0) G = 0; else if (G > 255) G = 255;

if (B < 0) B = 0; else if (B > 255) B = 255;

rgb[index++] = 0xff000000 + (R << 16) + (G << 8) + B;

什么是BMP

BMP文件主要由3个部分组成：BMP头信息，DIB头信息，以及数据。参考一下维基上的例子：

Byte数组写入BMP文件

JDK的接口和Android SDK的接口是不同的。在Windows上可以用BufferedImage和ImageIO来保存数据。

BufferedImage bufferedImage = new BufferedImage(width, height, BufferedImage.TYPE_3BYTE_BGR);

bufferedImage.setRGB(0, 0, width, height, data, 0, width);

try {

ImageIO.write(bufferedImage, "JPG", new File(fileName));

} catch (IOException e) {

// TODO Auto-generated catch block

e.printStackTrace();

}

我们也可以根据BMP文件的结构自己创建一个类。可以参考下DIB结构的定义:

public class BMP {

// BMP Header

private byte[] id = { 0x42, 0x4D };

private int fileSize = 0;

private short spec1 = 0, spec2 = 0;

private int offset = 54;

// DIB Header

private int biSize = 40;

private int biWidth, biHeight;

private short biPlanes = 0, biBitCount = 32;

private int biCompression, biSizeImage, biXPelsPerMeter, biYPelsPerMeter, biClrUsed, biClrImportant;

// bitmap data

private int[] data;

public BMP(int width, int height, short pixelBits, int[] pixels) {

biWidth = width;

biHeight = height;

biBitCount = pixelBits;

data = pixels;

fileSize = width * height * pixelBits / 8 + offset;

}

public int getFileSize() {

return fileSize;

}

}

Java把整形值写入到文件中默认使用big-endian。如果用DataOutputStream来直接写文件，会出现问题。

DataOutputStream output;

try {

output = new DataOutputStream(new BufferedOutputStream(new FileOutputStream("big-endian.bmp")));

output.write(getHeader());

for (int i = 0; i < data.length; i++) {

output.writeInt(data[i]);

}

output.flush();

output.close();

} catch (FileNotFoundException e) {

// TODO Auto-generated catch block

e.printStackTrace();

} catch (IOException e) {

// TODO Auto-generated catch block

e.printStackTrace();

}

所以需要用ByteBuffer转换成little-endian再保存数据：

private byte[] getFile() {

ByteBuffer buffer = ByteBuffer.allocate(fileSize);

buffer.order(ByteOrder.LITTLE_ENDIAN);

buffer.put(id);

buffer.putInt(fileSize);

buffer.putShort(spec1);

buffer.putShort(spec2);

buffer.putInt(offset);

buffer.putInt(biSize);

buffer.putInt(biWidth);

buffer.putInt(biHeight);

buffer.putShort(biPlanes);

buffer.putShort(biBitCount);

buffer.putInt(biCompression);

buffer.putInt(biSizeImage);

buffer.putInt(biXPelsPerMeter);

buffer.putInt(biYPelsPerMeter);

buffer.putInt(biClrUsed);

buffer.putInt(biClrImportant);

for (int i = 0; i < data.length; i++) {

buffer.putInt(data[i]);

}

return buffer.array();

}

重新运行程序，图像正常：

源码