我最近搬到了一个新的开发盒,从Windows XP 32位到Windows 7 64位.这两台机器都运行.Net Framework 4.0版,用于使用Visual Studio 2010进行开发.
升级到Windows 7 64位后,我的代码将多页Tiff图像分割成单独的图像现在已被破坏(之前在XP 32 Bit上工作正常,除了MS填充顺序错误).在调试之后,.Net Framework似乎正确地读取了位图元数据,但是,堆栈中的某些组件错误地持久化了一些Tiff标签(273,278和279).我已经尝试了几种方法来完成拆分,包括GDI和FreeImage库,但都在.Net中失败.我能够使用Image Magick和另一个带有效标签值的第三方工具成功拆分Tiff.
具体来说,Tiff标签273,278(应匹配257但不匹配),279具有不正确的值.
这是一个已知的Microsoft问题吗?有解决方法吗?我做错了吗?非常沮丧,因为这在XP 32上正常工作并且操作系统不是部署选项.
// Copy bytes into our memory
using (MemoryStream ms = new MemoryStream())
{
using (BinaryWriter bw = new BinaryWriter(ms))
{
// Get the multi page tiff image's size, and allocate a place to put it.
int size = RdmInstance.ImageSize;
byte[] imgContents = new byte[size + 1];
// Create the pointer and assign the Rdm image address to it
IntPtr pointerToImage = new IntPtr(RdmInstance.ImageData);
// Copy the bytes from unmanaged memory to managed memory
Marshal.Copy(pointerToImage, imgContents, 0, size);
// Now, write that contents into our memory stream
bw.Write(imgContents);
// Check for multiple tiff pages, split them out and write them out correctly for the Fed
TiffBitmapDecoder decoder = new TiffBitmapDecoder(ms, BitmapCreateOptions.PreservePixelFormat, BitmapCacheOption.Default);
if (decoder.Frames.Count > 0)
{
// check for multi page tiff
for (int i = 0; i < decoder.Frames.Count; i++)
{
log.InfoFormat("Loading Multi Page Tiff Frame [{0}]... to bitmap", i);
// First is front, second is back
// TODO - it would be better to get this out of tiff tag RDM sets with the page info
string fileName = (i == 0) ? frontFileName : backFileName;
BitmapSource bmSrc = decoder.Frames[i];
TiffBitmapEncoder encoder = new TiffBitmapEncoder();
encoder.Compression = TiffCompressOption.Ccitt4;
encoder.Frames.Add(BitmapFrame.Create(bmSrc));
log.InfoFormat("Saving Multi Page Tiff Frame [{0}]... to file {1}.", i, fileName);
using (var fs = new FileStream(fileName, FileMode.Create))
{
encoder.Save(fs);
}
/*
* jknipp - 6/4/2010
* Microsoft has a bug in their TiffBitmapEncoder where
* they incorrectly set tag 266 (Fill Order) to 0, where the TIFF
* spec says it should be 1 or 2. We fix this here.
* Reopen the stupid file and fix the fill order
*/
using (var file = new FileStream(fileName, FileMode.Open))
{
TiffBitmapDecoder output = new TiffBitmapDecoder(file, BitmapCreateOptions.PreservePixelFormat, BitmapCacheOption.Default);
InPlaceBitmapMetadataWriter metadata = output.Frames[0].CreateInPlaceBitmapMetadataWriter();
var fillOrder = metadata.GetQuery("/ifd/{ushort=266}");
log.DebugFormat("Read Fill Order Metadata tag as {0}", fillOrder);
// If .Net added a bogus fill order, correct it
if (fillOrder != null && (ushort)fillOrder == 0)
{
log.InfoFormat("Correcting FILL ORDER in file {0}", fileName);
metadata.SetQuery("/ifd/{ushort=266}", (ushort)1);
// Try to save new metadata
if (metadata.TrySave())
{
fillOrder = metadata.GetQuery("/ifd/{ushort=266}");
log.Info("Fill order correction successful!");
log.DebugFormat("Read New Fill Order Metadata tag as {0}", fillOrder);
}
}
}
}
}
}
}
解决方法:
您可以尝试使用LibTiff.Net库.它是免费的开源(BSD许可证).
该库附带了tiffcp实用程序,可以从您的代码中使用它来分割/合并TIFF图像.或者您可以使用该实用程序的源代码作为示例.
LibTiff.Net还包含使用应用程序中的tiffcp合并和分割TIFF图像的示例.
免责声明:我是图书馆的维护者之一.