说明：由于这里不能上传bitmap图像，我将用到的3个bitmap图片(红色字体，第2、3两图是24bit的位图)文件放在：http://download.csdn.net/source/1377107。请下载(忘记设定不需要资源分了，需要1个资源分)后，对照本文阅读。
Bitmaps should be saved in a file that uses the established bitmap file format and assigned a name with the three-character .bmp extension. The established bitmap file format consists of a BITMAPFILEHEADER structure followed by a BITMAPINFOHEADER, BITMAPV4HEADER, or BITMAPV5HEADER structure. An array of RGBQUAD structures (also called a color table) follows the bitmap information header structure. The color table is followed by a second array of indexes into the color table (the actual bitmap data).
The bitmap file format is shown in the following illustration.
Windows 95, Windows NT 4.0: Replace the BITMAPINFOHEADER structure with the BITMAPV4HEADER structure.
Windows 98/Me, Windows 2000/XP: Replace the BITMAPINFOHEADER structure with the BITMAPV5HEADER structure.
The members of the BITMAPFILEHEADER structure identify the file; specify the size of the file, in bytes; and specify the offset from the first byte in the header to the first byte of bitmap data. The members of the BITMAPINFOHEADER, BITMAPV4HEADER, or BITMAPV5HEADER structure specify the width and height of the bitmap, in pixels; the color format (count of color planes and color bits-per-pixel) of the display device on which the bitmap was created; whether the bitmap data was compressed before storage and the type of compression used; the number of bytes of bitmap data; the resolution of the display device on which the bitmap was created; and the number of colors represented in the data. The RGBQUAD structures specify the RGB intensity values for each of the colors in the device's palette.
The color-index array associates a color, in the form of an index to an RGBQUAD structure, with each pixel in a bitmap. Thus, the number of bits in the color-index array equals the number of pixels times the number of bits needed to index the RGBQUAD structures. For example, an 8x8 black-and-white bitmap has a color-index array of 8 * 8 * 1 = 64 bits, because one bit is needed to index two colors. The Redbrick.bmp, mentioned in About Bitmaps, is a 32x32 bitmap with 16 colors; its color-index array is 32 * 32 * 4 = 4096 bits because four bits index 16 colors.
To create a color-index array for a top-down bitmap, start at the top line in the bitmap. The index of the RGBQUAD for the color of the left-most pixel is the first n bits in the color-index array (where n is the number of bits needed to indicate all of the RGBQUAD structures). The color of the next pixel to the right is the next n bits in the array, and so forth. After you reach the right-most pixel in the line, continue with the left-most pixel in the line below. Continue until you finish with the entire bitmap. If it is a bottom-up bitmap, start at the bottom line of the bitmap instead of the top line, still going from left to right, and continue to the top line of the bitmap.
The following hexadecimal output shows the contents of the file Redbrick.bmp.
第一图
The following table shows the data bytes associated with the structures in a bitmap file.
Structure
Corresponding bytes
BITMAPFILEHEADER
0x00 0x0D
BITMAPINFOHEADER
0x0E 0x35
RGBQUAD array
0x36 0x75
Color-index array
0x76 0x275
第一图的存储结构：
32 * 32 pixels (8 times enlarged redbrick.bmp) :
Color table:
0 00 00 00 00
1 00 00 80 00
2 00 80 00 00
3 00 80 80 00
4 80 00 00 00
5 80 00 80 00
6 80 80 00 00
7 80 80 80 00
8 c0 c0 c0 00
9 00 00 ff 00
A 00 ff 00 00
B 00 ff ff 00
C ff 00 00 00
D ff 00 ff 00
E ff ff 00 00
F ff ff ff 00
The first three lines of Redbrick.bmp (bottom-up)
00 01 02 03 04 05 06 07 08 09 0A 0B 0C 0D 0E 0F
00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00
00 00 00 00 00 00 00 00 09 00 00 00 00 00 00 00
11 11 01 19 11 01 10 10 09 09 01 09 11 11 01 90
From the above data, we can see the first line of the picture should be black (index 0 means black from the color table), the second line is black except 18th pixels in line 2 is red (index 09 mean red from the color table). Obviously, we have the same result while look that enlarged redbrick.bmp above (bottom-up).
第2图
Englarged MSNFaceTest.bmp
第2图前128字节的数据如下：
00000000h 0 1 2 3 4 5 6 7 8 9 A B C D E F
00000010h 42 4D 36 03 00 00 00 00 00 00 36 00 00 00 28 00
00000020h 00 00 10 00 00 00 10 00 00 00 01 00 18 00 00 00
00000030h 00 00 00 03 00 00 00 00 00 00 00 00 00 00 00 00
00000040h 00 00 00 00 00 00 22 19 15 23 1A 16 24 1B 17 24
00000050h 1C 15 22 1C 15 22 1C 15 1F 1C 14 25 24 1A 23 27
00000060h 1C 25 2B 20 28 32 26 28 37 29 27 38 2A 22 37 28
00000070h 1F 24 25 24 37 26 24 1E 19 21 1B 16 20 1A 15 21
00000080h 1B 14 22 1C 15 22 1C 15 1E 1C 14 1E 1F 15 23 29
BITMAPFILEHEADER
-----------------------
WORD bfType; 0x42 0x4D = BM
DWORD bfSize; 0x0336 = 822 bytes
WORD bfReserved1; 0x00
WORD bfReserved2; 0x00
DWORD bfOffBits; 0x36 = 54 bytes
BITMAPINFOHEADER
------------------------
DWORD biSize; 0x28 = 40 bytes
LONG biWidth; 0x10 = 16 pixels
LONG biHeight; 0x10 = 16 pixels
WORD biPlanes; 0x01 = 1
WORD biBitCount; 0x18 = 24 bits-per-pixels
DWORD biCompression; 0x00
DWORD biSizeImage; 0x300 = 768 bytes
LONG biXPelsPerMeter; 0x00 = 0 pixels-per-meter
LONG biYPelsPerMeter; 0x00 = 0 pixels-per-meter
DWORD biClrUsed; 0x00
DWORD biClrImportant; 0x00
RGBQUAD array (color table)
----------------------------
NO COLOR TABLE
BITMAP BITS (The first 2 pixels in the first line - bottom-up)
--------------------------------------------------------------
22 19 15 23 1A 16 (B-G-R)
第三图(MSNFaceTest01.bmp)
要注意的一点是在BMP位图中，位图的每行像素值要填充到一个四字节边界，即位图每行所占的存储长度为四字节的倍数，不足时将多余位用0填充
前面两图都刚好每一像素行都是4字节的倍数，因此没有碰到上面的情况。第一图是32 *32像素，每像素4-bit颜色值（在color table里面的index，color table的大小为16，即16种颜色），所以每行的字节数为32 *4 /8 = 16字节，显然为4的倍数，因此不需要补0；第二图是16*16像素，每像素为24 bits，因此每行的大小为 16 * 24 /8 = 48字节，显然也为4的倍数，故而也不存在补0的情况。下面给出一个需要补0的图形。
第3图
放大后的情形如下：
第3图的存储结构：
BITMAPFILEHEADER
-----------------------
WORD bfType; 0x42 0x4D = BM
DWORD bfSize; 0x0196 = 406 bytes
WORD bfReserved1; 0x00
WORD bfReserved2; 0x00
DWORD bfOffBits; 0x36 = 54 bytes
BITMAPINFOHEADER
------------------------
DWORD biSize; 0x28 = 40 bytes
LONG biWidth; 0x0E = 14 pixels
LONG biHeight; 0x08 = 8 pixels
WORD biPlanes; 0x01 = 1
WORD biBitCount; 0x18 = 24 bits-per-pixels
DWORD biCompression; 0x00
DWORD biSizeImage; 0x160 = 352 bytes
LONG biXPelsPerMeter; 0x00 = 0 pixels-per-meter
LONG biYPelsPerMeter; 0x00 = 0 pixels-per-meter
DWORD biClrUsed; 0x00
DWORD biClrImportant; 0x00
RGBQUAD array (color table)
----------------------------
NO COLOR TABLE
BITMAP BITS
图像的尺寸为14*8像素，每个像素的颜色值占24bit，那么由此可以得到图像数据的大小应该为14 * 8 * 24/8 = 336字节，但是从从上面我们可以看到图像数据（BITMAP BYTES）的大小为352字节。两种算法所得的结果为什么会不同呢？这就要考虑到图像每行都必须是4字节的倍数，如果不是则要补0，知道这行图像数据为4的倍数。
在上面我们以及知道图像的宽度为14像素，每个像素的颜色值占25bit即3字节，那么每行的字节数为14 * 3 = 42字节，不是4的倍数，那么就需要补2字节0，使每行为44字节，这样就变成了4的倍数，从而符合bmp的文件格式了。补0后的每行字节数为44，共有8行，因此图像数据的大小为44 * 8 = 352字节。这样就biSizeImage和一致了。
从上图中我们看到00000060h和00000061h处补了2字节0（第一行），在0000008ch和0000008dh处补了2字节0（第二行）
本文来自CSDN博客，转载请标明出处：http://blog.csdn.net/pathuang68/archive/2009/05/27/4219681.aspx