Class Hierarchy
Fl_Image
|
+----Fl_Bitmap,
Fl_Pixmap,
Fl_RGB_Image,
Fl_Shared_Image,
Fl_Tiled_Image
Include Files
#include <FL/Fl_Image.H>
Description
Fl_Image is the base class used for caching and
drawing all kinds of images in FLTK. This class keeps track of
common image data such as the pixels, colormap, width, height,
and depth. Virtual methods are used to provide type-specific
image handling.
Since the Fl_Image class does not support image
drawing by itself, calling the draw() method results in
a box with an X in it being drawn instead.
Methods
The constructor creates an empty image with the specified
width, height, and depth. The width and height are in pixels.
The depth is 0 for bitmaps, 1 for pixmap (colormap) images, and
1 to 4 for color images.
The destructor is a virtual method that frees all memory used
by the image.
The color_average() method averages the colors in
the image with the FLTK color value c. The i
argument specifies the amount of the original image to combine
with the color, so a value of 1.0 results in no color blend, and
a value of 0.0 results in a constant image of the specified
color. The original image data is not altered by this
method.
The copy() method creates a copy of the specified
image. If the width and height are provided, the image is
resized to the specified size. The image should be deleted (or in
the case of Fl_Shared_Image, released) when you are done
with it.
The count() method returns the number of data values
associated with the image. The value will be 0 for images with
no associated data, 1 for bitmap and color images, and greater
than 2 for pixmap images.
The first form of the d() method returns the current
image depth. The return value will be 0 for bitmaps, 1 for
pixmaps, and 1 to 4 for color images.
The second form is a protected method that sets the current
image depth.
The first form of the data() method returns a
pointer to the current image data array. Use the
count() method to find the size of the data array.
The second form is a protected method that sets the current
array pointer and count of pointers in the array.
The desaturate() method converts an image to
grayscale. If the image contains an alpha channel (depth = 4),
the alpha channel is preserved. This method does not alter
the original image data.
The draw() methods draw the image. The first form
specifies the upper-lefthand corner of the image. The second
form specifies a bounding box for the image, with the origin
(upper-lefthand corner) of the image offset by the cx
and cy arguments.
The protected method draw_empty() draws a box with
an X in it. It can be used to draw any image that lacks image
data.
The first form of the h() method returns the current
image height in pixels.
The second form is a protected method that sets the current
image height.
The inactive() method calls
color_average(FL_BACKGROUND_COLOR, 0.33f) to produce
an image that appears grayed out. This method does not
alter the original image data.
The label() methods are an obsolete way to set the
image attribute of a widget or menu item. Use the
image() or deimage() methods of the
Fl_Widget and Fl_Menu_Item classes
instead.
The first form of the ld() method returns the current
line data size in bytes. Line data is extra data that is included
after each line of color image data and is normally not present.
The second form is a protected method that sets the current
line data size in bytes.
If the image has been cached for display, delete the cache
data. This allows you to change the data used for the image and
then redraw it without recreating an image object.
The first form of the w() method returns the current
image width in pixels.
The second form is a protected method that sets the current
image width.
From Anonymous, 18:11 Dec 10, 2008 (score=3)
The data() and count() explanations leave a lot to be desired. How is the "image data array" formatted and how do we get the pixels out? With RGB images it's easy to figure out after a quick look at Erco's cheat page, but what about indexed (paletted) images?
It turns out indexed images are stored by fltk in a format very similar to XPM. The members of the data array are as follows:
[0] (char *) : null-terminated string containing image information.
[1] (unsigned char *) : palette data (4 * palette length: index, r, g, b for each entry)
[2...H+2] (char *) : null-terminated string of characters representing palette indices
The exception to this rule is XPM images, which are stored with the normal (not compressed) XPM palette, so data[1] is an array of null-terminated strings.
In arrays 2 through H(eight)+2, a space always represents transparency.
This code sample should help explain how to get pixel values out of indexed images. Uncomment some of the printf statements to get a better idea of what's going on.
// Convert fltk shared image to PNG.
#ifndef _IMAGE_H
#define _IMAGE_H 1
#include "map"
#include "string"
#include <FL/Fl_Shared_Image.H>
#include "png.hpp"
class ImageHandler
{
public:
// save image as a png
static bool save(Fl_Shared_Image *image, const char *filename)
{
if (image->count() > 1)
return save_indexed(image, filename);
else
return save_inline(image, filename);
}
protected:
// save indexed images (gif... no xpm support yet)
static bool save_indexed(Fl_Shared_Image *image, const char *filename) // indexed images
{
png::image< png::rgba_pixel > png(image->w(), image->h());
const char * const *data = image->data();
// printf("Converting indexed image: %s\n",data[0]);
// fltk compressed palette length is stored as negative number - make sure it exists
std::string id(data[0]); // string containing image data
if (id.find_first_of('-') == std::string::npos)
{
fprintf(stderr,"XPM to PNG not currently supported.\n");
return false;
}
// extract palette size from image data string
int palsize = atoi(id.substr( id.find_first_of('-') +1, id.size() - id.find_last_of(' ') ).c_str());
// printf("Palette size: %i\n",palsize);
// for (int i = 2; i<image->count(); i++) printf("%s\n", data[i]);
// map fltk palette index characters to their position in the palette data array
uchar *pd = (uchar *)data[1]; // palette data array
std::map<uchar, int> palette;
for (uchar c=0; c<palsize*4; c+=4)
{
palette[pd[c]] = c;
//printf("%c: %02x%02x%02x \n", pd[c],pd[c+1],pd[c+2],pd[c+3]);
}
for ( int y=0; y<image->h(); y++ ) // Y loop
{
for ( int x=0; x<image->w(); x++ ) // X loop
{
const char *buf = image->data()[y+2];
char i = *(buf+x); // character we are on representing a palette index
char r,g,b,a;
r=g=b=0; // color defaults to black
a=255; // alpha defaults to fully visible
// set alpha to 0 for transparent pixels
if (i == ' ')
{
a = 0;
continue;
}
// set the pixel in the pngpp image
r = pd[palette[i]+1];
g = pd[palette[i]+2];
b = pd[palette[i]+3];
png.set_pixel(x,y,png::rgba_pixel(r,g,b,a));
} // end Y loop
} // end X loop
png.write(filename); // write the pngpp image to disk
// printf("depth: %i / width: %i / height: %i / count: %i \n", image->d(), image->w(), image->h(), image->count());
return true;
} // end save_indexed
// save non-indexed images (bmp, jpg, pnm) (see Erco's cheat page)
static bool save_inline(Fl_Shared_Image *image, const char *filename) // non-indexed images
{
// printf("Converting color image\n");
png::image< png::rgba_pixel > png(image->w(), image->h());
const char *buf = image->data()[0];
for ( int y=0; y<image->h(); y++ ) // Y loop
{
for ( int x=0; x<image->w(); x++ ) // X loop
{
long index = (y * image->w() * image->d()) + (x * image->d()); // X/Y -> buf index
char r,g,b,a;
r=g=b=0;
a=0xFF;
switch ( image->d() ) // check image depth
{
case 4: // 24bit + alpha
{
a = *(buf+index+3);
}
case 3: // 24bit
{
r = *(buf+index+0);
g = *(buf+index+1);
b = *(buf+index+2);
break;
case 2: // 8bit + alpha
{
a = *(buf+index+1);
}
case 1: // 8bit
{
r = g = b = *(buf+index);
break;
}
}
// fl reports depth 0 for xbm... add support here
default: // ??
{
fprintf(stderr, "Not supported: %d channels\n", image->d());
return false;
}
}
png.set_pixel(x,y,png::rgba_pixel(r,g,b,a));
} // end Y loop
} // end X loop
png.write(filename);
return true;
} // end save_inline
}; // end class
#endif
[ Reply ]From gebhart, 11:40 Jan 04, 2006 (score=3)
Is there any way to get the coordinates of a mouse-click when the user clicks on an image, without writing your own class?
I am developing a program with a combination of FLTK and C (no classes) and have an Fl_RGB_Image associated with a Fl_Box in my program. I would like to know the coordinates of a mouse-click when the user clicks on the image. Unfortunately, I am not developing my own classes, the callback of the Fl_Box is not invoked when I click on the image, and Fl_RGB_Image does not have a callback function tied to it.
Any ideas?
[ Reply ] |