使用pydicom合并大视频DICOM触发struct.error错误求助

我用Python3的pydicom库合并多个含视频的DICOM文件，每个输入文件用pydicom.dcmread读取后，pixel_array是形状为(number_of_frames, rows, columns, 3)的NumPy数组，仅帧数不同。我把这些数组用numpy拼接成大数组，写入Dataset后调用save_as保存：

• 拼接后视频较小时，代码正常运行；
• 当视频尺寸达到(6000,500,500,3)这类规模时，触发错误：

Traceback (most recent call last):
  File "...\Lib\site-packages\pydicom\tag.py", line 28, in tag_in_exception
    yield
  File "...\Lib\site-packages\pydicom\filewriter.py", line 662, in write_dataset
    write_data_element(fp, dataset.get_item(tag), dataset_encoding)
  File "...\Lib\site-packages\pydicom\filewriter.py", line 620, in write_data_element
    fp.write_UL(0xFFFFFFFF if is_undefined_length else value_length)
  File "...\Lib\site-packages\pydicom\filebase.py", line 119, in write_leUL
    self.write(pack(b"<L", val))
               ^^^^^^^^^^^^^^^^
struct.error: argument out of range

The above exception was the direct cause of the following exception:

Traceback (most recent call last):
  File "...\dicom_concatenate.py", line 56, in work
    first_dicom.save_as(absolute_save_path)
  File "...\Lib\site-packages\pydicom\dataset.py", line 2061, in save_as
    pydicom.dcmwrite(filename, self, write_like_original)
  File "...\Lib\site-packages\pydicom\filewriter.py", line 1153, in dcmwrite
    _write_dataset(fp, dataset, write_like_original)
  File "...\Lib\site-packages\pydicom\filewriter.py", line 889, in _write_dataset
    write_dataset(fp, get_item(dataset, slice(0x00010000, None)))
  File "...\Lib\site-packages\pydicom\filewriter.py", line 661, in write_dataset
    with tag_in_exception(tag):
  File "...\contextlib.py", line 155, in __exit__
    self.gen.throw(typ, value, traceback)
  File "...\Lib\site-packages\pydicom\tag.py", line 32, in tag_in_exception
    raise type(exc)(msg) from exc
struct.error: With tag (7fe0, 0010) got exception: argument out of range
Traceback (most recent call last):
  File "...\Lib\site-packages\pydicom\tag.py", line 28, in tag_in_exception
    yield
  File "...\Lib\site-packages\pydicom\filewriter.py", line 662, in write_dataset
    write_data_element(fp, dataset.get_item(tag), dataset_encoding)
  File "...\Lib\site-packages\pydicom\filewriter.py", line 620, in write_data_element
    fp.write_UL(0xFFFFFFFF if is_undefined_length else value_length)
  File "...\Lib\site-packages\pydicom\filebase.py", line 119, in write_leUL
    self.write(pack(b"<L", val))
               ^^^^^^^^^^^^^^^^
struct.error: argument out of range

这个错误的核心是DICOM的像素数据长度超出了无符号长整型（UL）的最大值：

• 标签(7fe0, 0010)是DICOM中的像素数据（Pixel Data）字段；
• 无符号长整型（UL）的最大值是0xFFFFFFFF（即4294967295字节，约4GB）；
• 拼接后的像素数据总大小为6000*500*500*3 = 4500000000字节，已经超过4GB上限，导致pydicom在尝试用struct.pack写入UL类型的长度值时，因数值超出范围触发struct.error。

方法1：使用未定义长度存储像素数据 #

pydicom支持对大尺寸数据使用未定义长度，写入时会自动分片段存储，避开UL的长度限制。修改保存逻辑：

# 直接使用dcmwrite并指定参数
pydicom.dcmwrite(absolute_save_path, first_dicom, write_like_original=False, transfer_syntax=pydicom.uid.ImplicitVRLittleEndian)

或手动设置未定义长度：

# 写入拼接后的像素数据
first_dicom.PixelData = concatenated_array.tobytes()
# 标记像素数据使用未定义长度
first_dicom.is_undefined_length('PixelData') = True
# 保存时禁用原格式写入
first_dicom.save_as(absolute_save_path, write_like_original=False)

方法2：拆分DICOM文件 #

如果必须使用固定长度存储，可将大视频拆分为多个符合4GB限制的DICOM文件，比如按每4000帧拆分（4000500500*3=3000000000字节，约2.79GB，小于4GB）。

方法3：使用压缩传输语法 #

若原始DICOM为未压缩格式，换成压缩传输语法（如JPEG 2000无损压缩）可大幅减小像素数据体积，从而避免超出长度限制：

from pydicom.uid import JPEG2000Lossless

# 保存时指定压缩传输语法
pydicom.dcmwrite(absolute_save_path, first_dicom, write_like_original=False, transfer_syntax=JPEG2000Lossless)

注意：使用压缩语法需确保pydicom的依赖库（如pylibjpeg及对应编解码器）已安装。

内容的提问来源于stack exchange，提问作者温泽海