CODEWITHSUNDEEP
cbmp
name
name

Reputation: 419

Don't understand how this code is scaling a bmp image

The following code was given to me by my instructor. I just don't understand how this is scaling a bmp image. I know the basics about bmp images (the info on wikipedia). I know that this method is supposed to multiply the rows and cols of the new image by whatever scale is. I tried to run the code by hand but it confused me even more. Any help will be much appreciated. Thanks! 

int enlarge(PIXEL* original, int rows, int cols, int scale, 
        PIXEL** new, int* newrows, int* newcols) 
{
    //scaling the new rows & cols
    *newcols = cols * scale;
    *newrows = rows * scale;

    //memory allocated for enlaged bmp 
    *new = (PIXEL*)malloc(*newrows * *newcols * sizeof(PIXEL));

    int row, col, sx, sy;


    //transverse through every row 
    for (row = 0; row < rows; row++ )
    //transvere through every col  
    for (col = 0; col < cols; col++ ){
        //im unsure what this is for 
        PIXEL* o = original + (row * cols) + col;
    for(sy = 0; sy < scale; sy++ )
    for(sx = 0; sx < scale; sx++ )
          { 
              //im unsure what this is for 
              PIXEL* n = *new + (scale * row) * *newcols + (scale * col) + (sy * *newcols) + sx;
              *n = *o;
          }
    }
    return 0; 
}

Here is the struct for PIXEL. 

typedef struct {
  unsigned char r;
  unsigned char g;
  unsigned char b;
} PIXEL;

There is additional code but I do not think that is needed for this question.

Upvotes: 0

Views: 454

Answers (1)

Matteo Italia
Matteo Italia

Reputation: 126907

    PIXEL* o = original + (row * cols) + col;

Here he is retrieving a pointer to the source pixel in the original image; it's just trivial pointer arithmetic, based on the fact that the rows in the bitmap are consecutive in memory. In general, in a C-style matrix width-wide the address of the element (x, y) is beginning + (y * width) + x.

Then, he loops over a square scale x scale wide in the target image.

for(sy = 0; sy < scale; sy++ )
for(sx = 0; sx < scale; sx++ )
      { 
          //im unsure what this is for 
          PIXEL* n = *new + (scale * row) * *newcols + (scale * col) + (sy * *newcols) + sx;

The n pointer points to the target pixel in the destination image; if you match the formula above from the source image and rearrange a bit the terms, you'll see he is accessing the new image, at position

(scale * col + sx, scale * row + sy)

(remember that the new image is *newcols wide).

          *n = *o;

Here he's just copying the source pixel to the target pixel.

In practice, he's "expanding" each source pixel into a scale x scale square in the target image.

Upvotes: 1

Related Questions