Description
EE231000 Introduction to Programming
Today’s digital images are composed of individual picture elements, call pixels. With a large
number of pixels, the image can be clear. But, with a large number of pixels the image file would
need a larger disk space for storage. Thus, most of the popular image file standards involve some
compression scheme to reduce the storage overhead. In this lab, we’ll concentrate on one of the
simplest color image format, portable pixmap format (PPM).
An image can be represented by a two-dimension array of pixels, the x-direction is the width
and y-direction is the height of the image. In PPM, each pixel is represented by 3 bytes for red,
green and blue colors. Thus, the total number of color levels is 2
24
, ∼16M. The format of a PPM
file is
P6
W H
255
R0G0B0R1G1B1 · · · RN GN BN
Where W and H are two integers and 255 is the intensity level for each color components (R, G,
and B). Since each color component R, G, B takes one byte, the intensity level is 255. After the
header, the remaining file consists of 3N, N=W×H, bytes, each byte representing the intensity of
one of the color components. This file can be read as following: the first line read using a string
of two characters, the second line read using “%d %d” for two integers, the third line read using
a string of 3 characters, then the remaining file read using 3N characters. Note that the new-line
character after the 3rd line should be handled carefully to ensure the image are read in correctly.
Also, the pixels are arranged in column-major fashion, instead of row-major matrix storage use
by C compiler.
In this lab, you will need to read in an image file, an EE Department logo and the NTHU
logo. You will need to covert the original color image to a blue-and-white image, add the EE
Department logo to the lower right corner of the image and the NTHU logo to the center of the
image. The EE department logo retains its colors, but the NTHU logo modifies the image to a
purple tone.
To convert a color image to a black-and-white image, the following method has been popular.
Let Rc, Gc and Bc be the intensities of a color pixel and Rg, Gg, Bg be the intensities of the
gray-scale image, then
Rg = Gg = Bg = Rc × 0.2126 + Gc × 0.7152 + Bc × 0.0722.
Note that when R = G = B at a pixel, this pixel has a gray color. And the levels of gray scale
is 256. When R = G = B =255, the color is white and when R = G = B = 0, the pixel has the
color black.
1
However, in this lab we will convert the image to a blue-and-while image, which can be
done simply making the blue component equals to the maximum intensity at all time. Thus,
when the gray level of the pixel is 255, we see a white pixel as it should be. And when the gray
level is 0, we will see a blue pixel instead. In this way, a blue-and-white image is obtained.
When placing EE Department logo, the pixels of the original image are simple replaced by
the logo pixels when the logo pixel is not white. But in placing the NTHU logo, we simply set
the red component of the pixel to be 255 when the NTHU logo pixel is not white. This will make
the logo purple since both blue and red color components are always on.
Five image files are available for you to test your program: pic1.ppm, pic3.ppm, pic7.ppm,
pic8.ppm, and pic9.ppm. The EE Department and the NTHU logos are also given: EElogo.ppm
and NTHU.ppm.
Your program needs to read an image file, EE Department log and NTHU logo files and
produce an output file. Thus, the execution of your program should be invoked by the following
command line
$ ./a.out ∼ee231002/lab13/pic1.ppm ∼ee231002/lab13/EElogo.ppm \
∼ee231002/lab13/NTHU.ppm pic1 out1.ppm
Where pic1.ppm and pic1_out1.ppm can be replaced by other image input and output files.
In order to read from a file, a FILE variable needs to be declared, the file opened and assigned
to the file variable, then fscanf can be used to read and fprintf for write. After all data
have been read and written, those file variables should be closed. The following example shows
reading a integer from a file data.in and write to data.out.
FILE *fin,*fout;
int k;
fin=fopen(“data.in”,”r”);
fout=fopen(“data.out”,”w”);
fscanf(fin,”%d”,&k);
fprintf(fout,”%d\n”,k);
fclose(fout);
fclose(fin);
Of course, your can use fopen(argv[1],”r”); to open a file specified by the command line
argument.
Notes.
1. Create a directory lab13 and use it as the working directory.
2. Name your program source file as lab13.c.
3. The first few lines of your program should be comments as the following.
/* EE231002 Lab13. Image Processing
2
ID, Name
Date:
*/
4. After you finish verifying your program, you can submit your source code by
$ ∼ee231002/bin/submit lab13 lab13.c
If you see a ”submitted successfully” message, then you are done. In case you want to
check which file and at what time you submitted your labs, you can type in the following
command:
$ ∼ee231002/bin/subrec
It will show the last few submission records.
5. You should try to write the program as efficient as possible. The format of your program
should be compact and easy to understand. These are part of the grading criteria.
6. An example of data structure to store the PPM image is shown below. You can but not
required to use this data structure.
typedef struct sPIXEL { // a single pixel
unsigned char r,g,b; // three color components
} PIXEL;
typedef struct sIMG { // an image of PPM style
char header[3]; // header, either P3 or P6
int W,H; // width and height of the image
int level; // intensity level of each color component
PIXEL **PX; // two-dimensional array for all the pixels
} IMG;
3
