【发布时间】：2015-04-03 08:07:07
【问题描述】：
我正在尝试使用 C# 应用程序查看一个非常大的位图。
这里的主要问题是我无法将它直接加载到内存中，因此我尝试使用内存映射视图库来使用来自Reference One、Reference Two、Reference Three、Reference Four 和 @987654325 的指导来加载它@ .
到目前为止，我达到的内容如下：- 根据需要分部分读取位图（例如，我读取了前 200 行）。 从我读取的行创建另一个位图并显示它。
问题：- 重建的位图图像部分丢失了颜色信息，并被上下颠倒显示。
示例：- [注意我在这里使用小尺寸图像并尝试显示其中的一部分以用于测试目的]
真实形象：-
应该是输出（选择前 200 行并重建一个较小的位图并显示它）：- 如您所见，重建后的图像是无色且上下颠倒的。
现在是代码部分：- 负责整个过程的BMPMMF类
using System;
using System.Collections.Generic;
using System.Linq;
using System.Text;
using System.IO;
using System.IO.MemoryMappedFiles;
using System.Drawing;
using System.Runtime.InteropServices;
using System.Drawing.Imaging;
namespace BMPViewer
{
class BMPMMF
{
/// <summary>
/// It opens the image using memory mapped view and read the needed
/// parts, then call CreateBM to create a partially bitmap
/// </summary>
/// <param name="bmpFilename">Path to the physical bitmap</param>
/// <returns></returns>
public Bitmap readPartOfImage(string bmpFilename)
{
var headers = ReadHeaders(bmpFilename);
var mmf = MemoryMappedFile.CreateFromFile(bmpFilename, FileMode.Open);
int rowSize = headers.Item2.RowSize; // number of byes in a row
// Dictionary<ColorObject, int> rowColors = new Dictionary<ColorObject, int>();
int colorSize = Marshal.SizeOf(typeof(MyColor));
int width = rowSize / colorSize;//(headers.Item1.DataOffset+ rowSize) / colorSize;
int height = 200;
ColorObject cObj;
MyColor outObj;
ColorObject[][] rowColors = new ColorObject[height][];
// Read the view image and save row by row pixel
for (int j = 0; j < height; j++)
{
rowColors[j] = new ColorObject[width];
using (var view = mmf.CreateViewAccessor(headers.Item1.DataOffset + rowSize * j, rowSize, MemoryMappedFileAccess.Read))
{
for (long i = 0; i < rowSize; i += colorSize)
{
view.Read(i, out outObj);
cObj = new ColorObject(outObj);
rowColors[j][i / colorSize] = cObj;
}
}
}
return CreateBM( rowColors );
}
/// <summary>
/// Used to create a bitmap from provieded bytes
/// </summary>
/// <param name="rowColors">Contains bytes of bitmap</param>
/// <returns></returns>
private Bitmap CreateBM(ColorObject[][] rowColors )
{
int width = rowColors[0].Count();
int height = rowColors.Count();
//int width = rowColors.Values.Where(o => o == 0).Count();
Bitmap bitm = new Bitmap(width, height, PixelFormat.Format24bppRgb);
// new Bitmap(imgdat.GetUpperBound(1) + 1, imgdat.GetUpperBound(0) + 1, PixelFormat.Format24bppRgb);
BitmapData bitmapdat = bitm.LockBits(new Rectangle(0, 0, bitm.Width, bitm.Height), ImageLockMode.ReadWrite, bitm.PixelFormat);
int stride = bitmapdat.Stride;
byte[] bytes = new byte[stride * bitm.Height];
for (int r = 0; r < bitm.Height; r++)
{
for (int c = 0; c < bitm.Width; c++)
{
ColorObject color = rowColors[r][c];
bytes[(r * stride) + c * 3] = color.Blue;
bytes[(r * stride) + c * 3 + 1] = color.Green;
bytes[(r * stride) + c * 3 + 2] = color.Red;
}
}
System.IntPtr scan0 = bitmapdat.Scan0;
Marshal.Copy(bytes, 0, scan0, stride * bitm.Height);
bitm.UnlockBits(bitmapdat);
return bitm;
}
/// <summary>
/// Returns a tuple that contains necessary information about bitmap header
/// </summary>
/// <param name="filename"></param>
/// <returns></returns>
private Tuple<BmpHeader, DibHeader> ReadHeaders(string filename)
{
var bmpHeader = new BmpHeader();
var dibHeader = new DibHeader();
using (var fs = new FileStream(filename, FileMode.Open, FileAccess.Read))
{
using (var br = new BinaryReader(fs))
{
bmpHeader.MagicNumber = br.ReadInt16();
bmpHeader.Filesize = br.ReadInt32();
bmpHeader.Reserved1 = br.ReadInt16();
bmpHeader.Reserved2 = br.ReadInt16();
bmpHeader.DataOffset = br.ReadInt32();
dibHeader.HeaderSize = br.ReadInt32();
if (dibHeader.HeaderSize != 40)
{
throw new ApplicationException("Only Windows V3 format supported.");
}
dibHeader.Width = br.ReadInt32();
dibHeader.Height = br.ReadInt32();
dibHeader.ColorPlanes = br.ReadInt16();
dibHeader.Bpp = br.ReadInt16();
dibHeader.CompressionMethod = br.ReadInt32();
dibHeader.ImageDataSize = br.ReadInt32();
dibHeader.HorizontalResolution = br.ReadInt32();
dibHeader.VerticalResolution = br.ReadInt32();
dibHeader.NumberOfColors = br.ReadInt32();
dibHeader.NumberImportantColors = br.ReadInt32();
}
}
return Tuple.Create(bmpHeader, dibHeader);
}
}
public struct MyColor
{
public byte Red;
public byte Green;
public byte Blue;
//public byte Alpha;
}
public class ColorObject
{
public ColorObject(MyColor c)
{
this.Red = c.Red;
this.Green = c.Green;
this.Blue = c.Blue;
// this.Alpha = c.Alpha;
}
public byte Red;
public byte Green;
public byte Blue;
// public byte Alpha;
}
public class BmpHeader
{
public short MagicNumber { get; set; }
public int Filesize { get; set; }
public short Reserved1 { get; set; }
public short Reserved2 { get; set; }
public int DataOffset { get; set; }
}
public class DibHeader
{
public int HeaderSize { get; set; }
public int Width { get; set; }
public int Height { get; set; }
public short ColorPlanes { get; set; }
public short Bpp { get; set; }
public int CompressionMethod { get; set; }
public int ImageDataSize { get; set; }
public int HorizontalResolution { get; set; }
public int VerticalResolution { get; set; }
public int NumberOfColors { get; set; }
public int NumberImportantColors { get; set; }
public int RowSize
{
get
{
return 4 * ((Bpp * Width) / 32);
}
}
}
}
这是如何使用它：-
Bitmap bmpImage = bmp.readPartOfImage(filePath); // path to bitmap
pictBoxBMP.Image = bmpImage; // set the picture box image to the new Partially created bitmap
我寻求的解决方案：- 只是正确显示部分创建的位图，我猜位图重建或使用内存映射视图读取位图时存在问题。
Update#1 : 应用@TaW solution 后，我得到了显示的颜色，即使它们与原始颜色有点不同，但可以接受。
【问题讨论】：
-
根据您忽略调色板的图像，在每个扫描行和行顺序上填充字节...查看位图格式规范以更好地理解所有字段（维基百科有很好的起始文章 - en.wikipedia.org/wiki/BMP_file_format )
-
@AlexeiLevenkov 我已经检查过了，我读取的图像格式是 Format24bppRgb（3 个字节，一个为红色，一个为蓝色，一个为绿色）。我重建的图像是 Format24bppRgb。
-
看起来您可能没有考虑行的 stride。大多数图像格式以某种对齐方式开始每一行，并填充前一行。
-
@CoryNelson 解释更多
-
尽管您说“我已经检查过了”，但我还是建议您更仔细地阅读它 - 即正如@CoryNelson 建议的那样，您没有考虑“像素格式由DIB Header 或Extra 位掩码定义。像素数组中的每一行被填充为4 个字节的倍数”（来自维基）
标签： c# image bitmap memory-mapped-files