Load BMP data (for heightmaps etc)

Overview

BMP is a simple image format developed by Microsoft. The colordepth wary between 2 (1-bit), 16 (4-bit), 256 (8-bit), 16-bit and 24-bit. The image is usually not compressed, but it might be compressed with the RLE algorithm. This tutorial will cover how to read pixels from a BMP image into an 2-dimensional array.

File format

File header

Offset	Bytes	Description	C/C++ datatype
0x00 - 0x01	2	This bytes must contain the ASCII character 'B' and 'M'	Char
0x02 - 0x05	4	Size of the file in bytes	Integer
0x06 - 0x07	4	Reserved, must be zero
0x08 - 0x09	4	Reserved, must be zero
0x0A - 0x0D	4	Offset to pixeldata	Integer

Image header

Offset	Bytes	Description	C/C++ datatype
0x0E - 0x11	4	Size of the image header, must be 40 (0x28)	Integer
0x12 - 0x15	4	Image width in pixels	Integer
0x16 - 0x19	4	Image height in pixels	Integer
0x1A - 0x1B	4	Nr of planes, must be one	Integer
0x1C - 0x1D	4	Bits per pixel, must be 1,2,4,8,16,24 or 32	Integer
0x1E - 0x21	4	Compression 0 = No compression 1 = RLE compression for 8bit bitmaps 2 = RLE compression for 4bit bitmaps Integer Following fields may be ignored most of the time. They is used to describe and store the palette. ||
0x22 - 0x25	4	Size of image data, may be zero if no compression is used	Integer
0x26 - 0x29	4	Horizontal resolution of the image	Integer
0x2A - 0x2D	4	Vertical resolution of the image	Integer
0x2E - 0x31	4	Number of color indexes actually used, if set to zero all colors is used	Integer
0x32 - 0x35	4	Nr of colors indexes required to render the image, if set to zero all color is used	Integer
0x36 -	4	Optional: This part may contain the palette used by the image. The format is RGBx (x is 1 byte reserved, must be set to zero) ||

Reading the header

The first step to read our image is to read its header to find out its size,bits-per-pixel and encoding. We will begin by making a simple struct to hold our data and a couple of enumerations for eventual errors. Lets start with struct.

Code:

1. typedef struct Bitmap {
2.         int offset;
3.         int width;
4.         int height;
5.         int bpp;
6.         int encoding;
7.         int size;
8.         bool loaded;
9.         unsigned char* pData;
10. };

For now this is all we need to store, but we will add more later. I don't think any of these need any further explanation.

Code:

1. enum BITMAP_RESULTS {
2.         BMP_OK,
3.         BMP_NO_FILE,
4.         BMP_INVALID_HEADER,
5.         BMP_OUT_OF_MEMORY,
6.         BMP_UNSUPPORTED,
7.         BMP_LOADED,
8. };

The first enum simply indicates everythings good soo far. BMP_NO_FILE means the file could not be opened, most likely the file doesn't exist but it might be permission error. BMP_INVALID_HEADER tells us the file isn't a valid BMP file, either it doesn't start with BM or the length is wrong. BMP_OUT_OF_MEMORY is self explanary. BMP_UNSUPPORTED is going to be used when the bpp has weired modes and/or encodings. The last one, BMP_LOADED, is used when there is already a bitmap loaded.

In order to read the header we first need to load the file into memory:

Code:

1. #include <fstream>
2.  
3. using namespace std;
4.  
5. int loadBMP(Bitmap* img,char* filename){
6.         ifstream::pos_type size
7.         ifstream file(filename, ios::in|ios::binary|ios::ate);
8.         if(file.is_open()){
9.                 size = file.tellg();
10.                 memblock = new char [size];
11.                 file.seekg (0, ios::beg);
12.                 file.read (memblock, size);
13.                 file.close();
14.         }
15. }

Loading 8-bit monocrome image

TBD

Loading 24-bit truecolor image

TBD

Summary

TBD