DCMTK:单张jpg、bmp图像转dcm文件

网上关于相关系列的文章非常的零散，特别不方便查找，于是查看img2dcm的源码，并将其提取出来作为项目需要，特记录于此。

前提

在dcm文件中，bmp格式图片和jpg格式图片保存方式完全不同，无压缩的bmp图片是直接将像素信息以tag:value的形式存储在dcm文件中，而有压缩的jpg图片则是先将图片的像素点信息插入进DcmPixelSequence中，然后将DcmPixelSequence以tag：value的形式存储在dcm文件中。
具体含义，便用代码分析，边解释。

在具体分析之前，先将主体代码贴下来，如下：

DcmFileFormat *fileFormat = new DcmFileFormat();
DcmDataset *dataSet = fileFormat->getDataset();
I2DImgSource *imgSource = 0;
if(imgName.endsWith("jpg", Qt::CaseInsensitive)) {
    imgSource = new I2DJpegSource();
}
else if(imgName.endsWith("bmp", Qt::CaseInsensitive)) {
    imgSource = new I2DBmpSource();
}
std::string temp = imgName.toStdString();//注：这句和下一句不可和为一句，否则，转换为char*会失败
const char *value = temp.c_str();
imgSource->setImageFile(OFString(value));
Uint16 samplesPerPixel, rows, cols, bitsAlloc, bitsStored, highBit, pixelRepr, planConf;
Uint16 pixAspectH =1; Uint16 pixAspectV = 1;
OFString photoMetrInt;
E_TransferSyntax outputTS = EXS_Unknown;
char* pixData = 0;
Uint32 length;

OFCondition cond = imgSource->readPixelData(rows, cols,
  samplesPerPixel, photoMetrInt, bitsAlloc, bitsStored, highBit, pixelRepr,
  planConf, pixAspectH, pixAspectV, pixData, length, outputTS);
if(cond.bad()) {
    qDebug()<<"read image data failed:"<<cond.text();
    return false;
}
DcmXfer transport(outputTS);
if (transport.isEncapsulated()) {
    jpg格式处理
}
else {
    bmp格式处理
}

其中，基类I2DImgSource就要根据输入文件格式不同，分别予以不同的初始化。然后插入像素信息。

BMP格式

直接将像素信息添加到dataSet中即可：

dataSet->putAndInsertUint8Array(DCM_PixelData, OFreinterpret_cast(Uint8*, pixData), length);

JPG格式

详细解释，如图所示：

由此可见，在插入像素信息时，先插入4个像素的tag，然后插入像素长度，然后插入像素值。所以，具体代码信息为：

DcmPixelSequence* pixelSequence = new DcmPixelSequence(DCM_PixelSequenceTag);
if (pixelSequence == NULL) {
   qDebug()<<"allocate pixel sequence failed"<<endl;
   return false;
}
// insert empty offset table into sequence
DcmPixelItem *offsetTable = new DcmPixelItem(DCM_PixelItemTag);
if (offsetTable == NULL) {
   delete pixelSequence;
   pixelSequence = 0;
   qDebug()<<"allocate pixel item tag failed"<<endl;
   return false;
}
//插入像素tag
OFCondition status = pixelSequence->insert(offsetTable);
if(status.bad()) {
   delete offsetTable;
   offsetTable = 0;
   delete pixelSequence;
   pixelSequence = 0;
   qDebug()<<"insert table failed:"<<status.text()<<endl;
   return false;
}
// store compressed frame into pixel sequence
DcmOffsetList dummyList;
status = pixelSequence->storeCompressedFrame(dummyList, OFreinterpret_cast(Uint8*,pixData), length, 0);
if (status.bad()) {
   delete pixelSequence;
   pixelSequence = 0;
   qDebug()<<"store frame failed:"<<status.text()<<endl;
   return false;
}

// insert pixel data attribute incorporating pixel sequence into dataset
DcmPixelData *pixelData = new DcmPixelData(DCM_PixelData);
if (pixelData == NULL) {
   delete pixelSequence;
   pixelSequence = 0;
   qDebug()<<"allocate DcmPixelData failed"<<endl;
   return false;
}
/* tell pixel data element that this is the original presentation of the pixel data
* pixel data and how it compressed
*/
pixelData->putOriginalRepresentation(outputTS, NULL, pixelSequence);
status = dset->insert(pixelData, OFTrue);
if (status.bad()) {
   delete pixelData;
   pixelData = 0; // also deletes contained pixel sequence
   qDebug()<<"put origin failed:"<<status.text()<<endl;
   return false;
}
status = dset->chooseRepresentation(outputTS, NULL);
return true;

最后才将相关信息保存到dcm中。注意，此时在保存成dcm文件时，E_TransferSyntax信息一定要和outputTS的信息保持一致，否则，有可能出现保存出来的dcm文件像素信息丢失。

具体的源代码，见CSDN链接：
http://download.csdn.net/detail/syz9011/9261663

注意，该链接内文件中，并未给出PatientData和ExaminationData源代码，里面只是常见的get和set，请各位自行去掉。或者，直接将里面用到该类的get的地方，替换为任意值即可。