Reading an image 1

[ferumus]

Hello all,

I am new to much C++ syntax. I am an avid programmer in other languages. I am trying to develope an application that takes an image and modifies it automatically. But for some reason I can't find an efficient way to read a .pgm image (converting it to an array) and then modifying some of the pixels and then just displaying it.

I have this starting code I found:


#include <iostream.h>
#include <fstream.h>
#include <stdlib.h>

#include "image.h"

int readImage(char fname[], ImageType& image)
{
int i, j;
int N, M, Q;
unsigned char *charImage;
char header [100], *ptr;
ifstream ifp;

ifp.open(fname, ios::in);

if (!ifp) {
cout << "Can't read image: " << fname << endl;
exit(1);
}

// read header

ifp.getline(header,100,'\n');
if ( (header[0]!=80) || /* 'P' */
(header[1]!=53) ) { /* '5' */
cout << "Image " << fname << " is not PGM" << endl;
exit(1);
}

ifp.getline(header,100,'\n');
while(header[0]=='#')
ifp.getline(header,100,'\n');

M=strtol(header,&ptr,0);
N=atoi(ptr);

ifp.getline(header,100,'\n');
Q=strtol(header,&ptr,0);

charImage = (unsigned char *) new unsigned char [M*N];

ifp.read( reinterpret_cast<char *>(charImage), (M*N)*sizeof(unsigned char));

if (ifp.fail()) {
cout << "Image " << fname << " has wrong size" << endl;
exit(1);
}

ifp.close();

//
// Convert the unsigned characters to integers
//

int val;

for(i=0; i<N; i++)
for(j=0; j<M; j++) {
val = (int)charImage[i*M+j];
image.setPixelVal(i, j, val);
}

delete [] charImage;

return (1);

}



Any advice would be much appreciated.
Thanks,

fer

 

[smays]

I have looked at the PGM format specification at

http://netpbm.sourceforge.net/doc/pgm.html.

You mentioned you want to write a program that modifies a PGM image. Although untested, the code you have above should read in the PGM image into charImage[], which is not exactly how I would have done it, but it is a start.

Have you tried the code you supplied? If so, what was the result? The code does not supply a real definition of what "ImageType& image" is. Perhaps it is obvious to others what "ImageType& image" is, but it is not obvious to me. If you had trouble compiling the code, I would imaging "ImageType& image" would cause the compiler to choke.

What development tool are you using to write your program? C++ Builder? If you aren't using C++ Builder, is your development tool capable of showing images?

Steve.

 

[ferumus]

Thanks for the reply Steve.

You are right. That code gave me some build problems.
So I read the image something like:

unsigned char* read_pgm(char *file_name, int *width, int *height);
void write_pgm(char *file_name, char *comment, unsigned char *image,int width,int height);

strcpy(image_in,"pic.pgm");

in=read_pgm(image_in,width,height);

This worked in storing the image in an array of values from 0 to 255.

You mentioned showing images. I am using Dev C++ which I believe is not capable of this.
So I am writing the image to a file and then opening it.
With something like:


strcpy(image_out,"outPic.pgm");
write_pgm(image_out,"pic_mod1.c",out,width,height);
// here 'out' is the array of pixels after modification.


This works for basic modification now.

The issue I have now is converting the color image to a grayscale first. Then I can use the above to modify it. If anyone knows a way I would be grateful.


Thanks

 

[smays]

You will have to either stumble across some example code on the internet showing how to convert from color to gray or you will have to compare a color image and an otherwise identical grayscale image to develop your own algorithm to perform the conversion. Below is a link to an 'array' of identical images in both color and grayscale. You can compare the images and develop your own algorithm.

http://www.math.umbc.edu/~rouben/2003-09-math625/images.html

To compare the data I would recommend a Hex Editor. I have used XVI32 for years and I highly recommend it:

http://www.chmaas.handshake.de/delphi/freeware/xvi32/xvi32.htm

Steve.

 

[ferumus]

Thanks Steve.
Much appreciated.