我有像这样获取的图像数据:

unsigned char* imageData = NULL;
GetImage(imageData);

imageData 以原始 BayerGR8 格式返回:也就是说:

G R G R G R G R ... 
B G B G B G B G ...
G R G R G R G R ...
      ...

其中每个像素占用 8 位。

正在抓取的图像为 2752x2200(像素)。

每当我设置位图然后使用该图像数据创建位图时，位图总是空白。这是我的位图设置:

    BITMAPFILEHEADER* bf = new BITMAPFILEHEADER;
    bf->bfType = 0x4d42;
    bf->bfSize = 6054400 + 54 + sizeof(BITMAPINFO);
    bf->bfOffBits = 54;

    BITMAPINFOHEADER* bih = new BITMAPINFOHEADER;
    bih->biSize = 40;
    bih->biWidth = 2752;
    bih->biHeight = 2200;
    bih->biPlanes = 1;
    bih->biBitCount = 8;
    bih->biCompression = 0;
    bih->biXPelsPerMeter = 2835;
    bih->biYPelsPerMeter = 2835;
    bih->biClrUsed = 0;
    bih->biClrImportant = 0;

到目前为止，我已经尝试了以下方法:

HDC hdc = ::GetDC(NULL);
HBITMAP hbit = CreateDIBitmap(hdc, globalinfoheader, CBM_INIT, imageData, pbmi, DIB_RGB_COLORS);

但正如我所说，位图最终变成了空白的浅灰色图像。我一直在查看关于位图的维基百科文章，我觉得它可能与 biClrUsed 值有关，但我不知道它应该是什么值并且到目前为止只是在玩数字。

我也试过将图像数据直接转换为 CImage(稍后我将其转换为 HBITMAP)，如下所示:

void ConvertImageBufferToCImage(unsigned char *pInBuffer, CImage *pOutImage)
{
    if ( NULL != *pOutImage )
    {
        unsigned char *pCursor = (unsigned char*)pOutImage->GetBits();
        int nHeight = 2200;
        int nWidth = 2752;
        int nStride = 0;

        if ( 0 < nStride)
        {
            for ( int y=0; y<nHeight; ++y )
            {
                for ( int x=0; x<nWidth; ++x )
                {
                    *pCursor = *pInBuffer;
                    ++pCursor;
                    ++pInBuffer;
                }
                // Consider stride
                pCursor += nStride;
            }
        }
        else
        {
            memcpy( pOutImage->GetBits(), pInBuffer, nWidth * nHeight );
        }
    }
}

我的问题是:如何获取我的原始 Bayer 图像数据并从中获取 rgb 颜色位图？

编辑:

明白了。这是我创建的让它工作的函数(插值方法):

/////////////////////////////////////////////////////////////
// ConvertBayer8ToBgr()
// Converts raw BayerGR8 pixels into
//     BGR pixels.
//
//  G | R | G | R              B G R | B G R | B G R | B G R
// --- --- --- ---            ------- ------- ------- -------
//  B | G | B | G        |\    B G R | B G R | B G R | B G R
// --- --- --- ---  -----  \  ------- ------- ------- -------
//  G | R | G | R   -----  /   B G R | B G R | B G R | B G R
// --- --- --- ---       |/   ------- ------- ------- -------
//  B | G | B | G              B G R | B G R | B G R | B G R
//
/////////////////////////////////////////////////////////////
void ConvertBayer8ToBGR(VmbUchar_t* bayerImgDat, VmbUchar_t* bgrOutputDat)
{
    VmbUchar_t* newimagedata_start = bgrOutputDat;

    int currentTempIndex = 0;
    int nearestBluesAvg = 0;
    int nearestRedsAvg = 0;
    int nearestGreensAvg = 0;

    for(int j = 0; j < 1100; j++)
    {
        for(int i = 0; i < 2752; i++) //G R G R G... 
        {
            if(currentTempIndex % 2 == 0 /* even, green */)
            {
                //avg blue
                if(j == 0) //if in the first row, only take next blue
                {
                    nearestBluesAvg = *(bayerImgDat+currentTempIndex+2752);
                }
                else
                {
                    nearestBluesAvg = (*(bayerImgDat + currentTempIndex + 2752) + *(bayerImgDat+currentTempIndex-2752)) / 2;
                }
                *bgrOutputDat = nearestBluesAvg; //b
                bgrOutputDat++;
                *bgrOutputDat = *(bayerImgDat + currentTempIndex); //g
                bgrOutputDat++;
                //avg red
                if(i == 0) //if in first column, only take next red 
                {
                    nearestRedsAvg = *(bayerImgDat+currentTempIndex+1);
                }
                else
                {
                    nearestRedsAvg = ( (*(bayerImgDat+currentTempIndex+1)) + (*(bayerImgDat+currentTempIndex-1)) ) / 2;
                }
                *bgrOutputDat = nearestRedsAvg; //r
                bgrOutputDat++;

                currentTempIndex++;
            }
            else /* odd, red*/
            {
                //avg blue
                if(i == 1099) //if in last column, take just left-down blue pixel
                {
                    nearestBluesAvg = *(bayerImgDat+currentTempIndex-1+2752);
                }
                else // else take both left-down and right-down
                {
                    nearestBluesAvg = (*(bayerImgDat+currentTempIndex+1+2752) + *(bayerImgDat+currentTempIndex-1+2752)) / 2;
                }
                *bgrOutputDat = nearestBluesAvg; //b
                bgrOutputDat++;
                //avg green
                nearestGreensAvg = (*(bayerImgDat+currentTempIndex-1) + *(bayerImgDat+currentTempIndex+2752)) / 2;
                *bgrOutputDat = nearestGreensAvg;  //g
                bgrOutputDat++;
                *bgrOutputDat = *(bayerImgDat + currentTempIndex); //r
                bgrOutputDat++;

                currentTempIndex++;
            }
        }
        for(int i = 0; i < 2752; i++)//B G B G B G B....
        {
            if(currentTempIndex % 2 == 0 /* even, blue */)
            {

                *bgrOutputDat = *(bayerImgDat + currentTempIndex); //b
                bgrOutputDat++;
                //avg green
                nearestGreensAvg = (*(bayerImgDat + currentTempIndex + 1) + *(bayerImgDat + currentTempIndex -2752)) / 2;
                *bgrOutputDat = nearestGreensAvg; //g
                bgrOutputDat++;
                //avg red
                if(i == 0) //if first column, take only right-up pixel
                {
                    nearestRedsAvg = *(bayerImgDat+currentTempIndex+1-2752);
                }
                else //else take both left-up and right-up pixels
                {
                    nearestRedsAvg = (*(bayerImgDat+currentTempIndex-1-2752) + *(bayerImgDat+currentTempIndex+1-2752)) / 2;
                }
                *bgrOutputDat = nearestRedsAvg; //r
                bgrOutputDat++;

                currentTempIndex++;

            }
            else /* odd, green*/
            {
                //avg blue
                if(i == 2751) //if in last column, only take previous blue (next blue doesnt exist)
                {
                    nearestBluesAvg = *(bayerImgDat + currentTempIndex - 1);
                }
                else //else take both next and previous
                {
                    nearestBluesAvg = (*(bayerImgDat+currentTempIndex+1) + *(bayerImgDat+currentTempIndex-1)) / 2;
                }
                *bgrOutputDat = nearestBluesAvg; //b
                bgrOutputDat++;
                *bgrOutputDat = *(bayerImgDat + currentTempIndex); //g
                bgrOutputDat++;
                //avg red
                if(j == 1099) //if in last row, only take previous red (next red doesn't exist)
                {
                    nearestRedsAvg = *(bayerImgDat+currentTempIndex-2752);
                }
                else //else take both
                {
                    nearestRedsAvg = (*(bayerImgDat+currentTempIndex+2752) + *(bayerImgDat+currentTempIndex-2752)) / 2;
                }
                *bgrOutputDat = nearestRedsAvg; //r
                bgrOutputDat++;

                currentTempIndex++;
            }
        }
    }


    bgrOutputDat = newimagedata_start;
}

最佳答案

您必须对拜耳图像的组件进行插值。这是一个有用的链接:

http://www.siliconimaging.com/RGB%20Bayer.htm

往下看插值部分。

就目标图像而言，只需创建标准 RGB 24 位或 32 位位图，并在从拜耳图像中插入组件时设置位。

您的问题似乎一半是关于颜色转换，一半是关于为什么您的位图代码不起作用。在 Windows 中创建 RGB 位图的示例不胜枚举，因此我不再赘述。

关于c++ - 如何从原始拜耳图像数据(字节数组)中获取位图图像？，我们在Stack Overflow上找到一个类似的问题： https://stackoverflow.com/questions/16531122/