Imager - Perl extension for Generating 24 bit Images


  use Imager;

  $img = Imager->new();
    || print "failed: ",$img->{ERRSTR},"\n";
    || print "failed: ",$scaled->{ERRSTR},"\n";


Imager is a module for creating and altering images - It is not meant as a replacement or a competitor to ImageMagick or GD. Both are excellent packages and well supported.

Why a new module? Compiling PerlMagick can be complicated, and it lacks drawing functions. has those but only supports gif and png. I like studying graphics, so why not let others in a similar situation benefit? The basis for this module is code written to preprocess remote sensing data.

Note: Documentation is ordered in:


Basic concepts

Reading and writing images

Obtaining/setting attributes of images

Drawing Methods

Text rendering

Image resizing





Functional interface


Almost all functions take the parameters in the hash fashion. Example:


or just:


Basic concept

An Image object is created with $img = Imager->new() Should this fail for some reason an explanation can be found in $Imager::ERRSTR usually error messages are stored in $img->{ERRSTR}, but since no object is created this is the only way to give back errors. $Imager::ERRSTR is also used to report all errors not directly associated with an image object. Examples:

    $img=Imager->new(); # This is an empty image (size is 0 by 0)
    $img->open(file=>'lena.png',type=>'png'); # initializes from file

or if you want to start clean image:


The latter example creates a completely black image of width 400 and height 300 and 4 channels.

To create a color object call the function i_color_new, $color=i_color_new($r,$g,$b,$a). The parameters are all from 0 to 255 and are all converted to integers. Each is the red, green, blue, and alpha component of the color respectively. This object can then be passed to functions that require a color parameter.

Coordinates in Imager have the origin in the upper left corner. The horizontal coordinate increases to the left and he vertical downwards.

Reading and writing images

$img->read() has generally has two parameters, 'file' and 'type'. If the type of the file can be determined from the suffix of the file it can be omitted. Format dependant parameters are: For images of type 'raw' two extra parameters are needed 'xsize' and 'ysize', if the 'channel' parameter is omitted for type 'raw' it is assumed to be 3. gif and png images might have a palette are converted to truecolor bit when read. Alpha channel is preserved for png images irregardless of them being in RGB or gray colorspace. Similarly grayscale jpegs are one channel images after reading them. For jpeg images the iptc header information (stored in the APP13 header) is avaliable to some degree. You can get the raw header with $img->{IPTCRAW}, but you can also retrieve the most basic information with %hsh=$img->parseiptc() as always patches are welcome.

*Note that load() is now an alias for read but will be removed later*

$img->write has the same interface as open(). The earlier comments on read() for autodetecting filetypes apply. For jpegs quality can be adjusted via the 'jpegquality' parameter (0-100). The number of colorplanes in gifs are set with 'gifplanes' and should be between 1 (2 color) and 8 (256 colors). It is also possible to choose between two quantizing methods with the parameter 'gifquant'. If set to mc it uses the mediancut algorithm from either giflibrary. If set to lm it uses a local means algorithm. It is then possible to give some extra settings. lmdither is the dither deviation amount in pixels (manhattan distance). lmfixed can be an array ref who holds an array of i_color objects. Note that the local means algorithm needs much more cpu time but also gives considerable better results than the median cut algorithm.

Obtaining/setting attributes of images

To get the size of an image in pixels the $img->getwidth() and $img->getheight() are used.

To get the number of channels in an image $img->getchannels() is used. $img->getmask() and $img->setmask() are used to get/set the channel mask of the image.

  $img->setmask(mask=>1+2); # modify red and green only
  $img->setmask(mask=>8); # modify alpha only
  $img->setmask(mask=>$mask); # restore previous mask

The mask of an image describes which channels are updated when some operation is performed on an image. Naturally it is not possible to apply masks to operations like scaling that alter the dimensions of images.

It is possible to have Imager find the number of colors in an image by using $img->getcolorcount(). It requires memory proportionally to the number of colors in the image so it is possible to have it stop sooner if you only need to know if there are more than a certain number of colors in the image. If there are more colors than asked for the function return undef. Examples:

  if (!defined($img->getcolorcount(maxcolors=>512)) {
    print "Less than 512 colors in image\n";

Drawing Methods



It is possible to draw with graphics primitives onto images. Such primitives include boxes, arcs, circles and lines. A reference oriented list follows.

Box: $img->box(color=>$blue,xmin=>10,ymin=>30,xmax=>200,ymax=>300,filled=>1);

The Above example calls the box method for the image and the box covers the pixels with in the rectangle specified. If filled is ommited it is drawn as an outline. If any of the edges of the box are ommited it will snap to the outer edge of the image in that direction. Also if a color is omitted a color with (255,255,255,255) is used instead.

Arc: $img->arc(color=>$red, r=20, x=>200, y=>100, d1=>10, d2=>20 );

This creates a red arc with a 'center' at (200, 100) and spans 10 degrees and the slice has a radius of 20. SEE section on BUGS.

Circle: $img->circle(color=>$green, r=50, x=>200, y=>100 );

This creates a green circle with its center at (200, 100) and has a radius of 20.


Polyline is used to draw multilple lines between a series of points. The point set can either be specified as an arrayref to an array of array references (where each such array represents a point). The other way is to specify two array references.

Text rendering

To create a font object you can use:

  $t1font = Imager::Font->new(file=>'pathtofont.pfb');
  $ttfont = Imager::Font->new(file=>'pathtofont.ttf');

As is two types of font types are supported t1 postscript fonts and truetype fonts. You can see if they are supported in your binary with the %Imager::formats hash. It is possible to control other attributes the font such as default color, size and anti aliasing.

  $blue = Imager::Color(10,10,255,0);
  $t1font = Imager::Font->new(file=>'pathtofont.pfb',

To draw text on images the string method of the images is used. A font must be passed to the method.


This would put a 40 pixel high text in the top left corner of an image. If you measure the actuall pixels it varies since the fonts usually do not use their full height. It seems that the color and size can be specified twice. When a font is created only the actual font specified matters. It his however convenient to store default values in a font, such as color and size. If parameters are passed to the string function they are used instead of the defaults stored in the font.

If string() is called with the channel parameter then the color isn't used and the font is drawn in only one channel. This can be quite handy to create overlays. See the examples for tips about this.

Sometimes it is necessary to know how much space a string takes before rendering it. The bounding_box() method of a font can be used for that. examples:


The first example gets the so called glyph metrics. First is the

Image resizing

To scale an image so porportions are maintained use the $img->scale() method. if you give either a xpixels or ypixels parameter they will determine the width or height respectively. If both are given the one resulting in a larger image is used. example: $img is 700 pixels wide and 500 pixels tall.

  $img->scale(xpixels=>400); # 400x285
  $img->scale(ypixels=>400); # 560x400

  $img->scale(xpixels=>400,ypixels=>400); # 560x400
  $img->scale(xpixels=>400,ypixels=>400,type=>min); # 400x285

  $img->scale(scalefactor=>0.25); 175x125 $img->scale(); # 350x250

if you want to create low quality previews of images you can pass qtype=>'preview' to scale and it will use nearest neighbor sampling instead of filtering. It is much faster but also generates worse looking images - especially if the original has a lot of sharp variations and the scaled image is by more than 3-5 times smaller than the original.

If you need to scale images per axis it is best to do it simply by calling scaleX and scaleY. You can pass either 'scalefactor' or 'pixels' to both functions.

Another way to resize an image size is to crop it. The parameters to crop are the edges of the area that you want in the returned image. If a parameter is omited a default is used instead.

  $img->crop(left=>50, right=>100, top=>10, bottom=>100); 
  $img->crop(left=>50, top=>10, width=>50, height=>90);
  $img->crop(left=>50, right=>100); # top 


A special image method is the filter method. An example is:


This will call the autolevels filter. Here is a list of the filters that are always avaliable in Imager. This list can be obtained by running the filterlist.perl script that comes with the module source.

  Filter          Arguments
  autolevels      lsat(0.1) usat(0.1) skew(0)
  noise           amount(3) subtype(0)
  contrast        intensity

The default values are in parenthesis. All parameters must have some value but if a parameter has a default value it may be omitted when calling the filter function.


Another special image method is transform. It can be used to generate warps and rotations and such features. It can be given the operations in postfix notation or the module Affix::Infix2Postfix can be used. Look in the test case t/t55trans.t for an example.

Later versions of Imager also support a transform2() class method which allows you perform a more general set of operations, rather than just specifying a spatial transformation as with the transform() method, you can also perform colour transformations, image synthesis and image combinations.

transform2() takes an reference to an options hash, and a list of images to operate one (this list may be empty):

  my %opts;
  my @imgs;
  my $img = Imager::transform2(\%opts, @imgs)
      or die "transform2 failed: $Imager::ERRSTR";

The options hash may define a transformation function, and optionally:

The tranformation function is specified using either the expr or rpnexpr member of the options.

Infix expressions

You can supply infix expressions to transform 2 with the expr keyword.

$opts{expr} = 'return getp1(w-x, h-y)'

The 'expression' supplied follows this general grammar:

   ( identifier '=' expr ';' )* 'return' expr

This allows you to simplify your expressions using variables.

A more complex example might be:

$opts{expr} = 'pix = getp1(x,y); return if(value(pix)>0.8,pix*0.8,pix)'

Currently to use infix expressions you must have the Parse::RecDescent module installed (available from CPAN). There is also what might be a significant delay the first time you run the infix expression parser due to the compilation of the expression grammar.

Postfix expressions

You can supply postfix or reverse-polish notation expressions to transform2() through the rpnexpr keyword.

The parser for rpnexpr emulates a stack machine, so operators will expect to see their parameters on top of the stack. A stack machine isn't actually used during the image transformation itself.

You can store the value at the top of the stack in a variable called foo using !foo and retrieve that value again using @foo. The !foo notation will pop the value from the stack.

An example equivalent to the infix expression above:

 $opts{rpnexpr} = 'x y getp1 !pix @pix value 0.8 gt @pix 0.8 * @pix ifp'

transform2() has a fairly rich range of operators.

+, *, -, /, %, **

multiplication, addition, subtraction, division, remainder and exponentiation. Multiplication, addition and subtraction can be used on colour values too - though you need to be careful - adding 2 white values together and multiplying by 0.5 will give you grey, not white.

Division by zero (or a small number) just results in a large number. Modulo zero (or a small number) results in zero.

sin(N), cos(N), atan2(y,x)

Some basic trig functions. They work in radians, so you can't just use the hue values.

distance(x1, y1, x2, y2)

Find the distance between two points. This is handy (along with atan2()) for producing circular effects.


Find the square root. I haven't had much use for this since adding the distance() function.


Find the absolute value.

getp1(x,y), getp2(x,y), getp3(x, y)

Get the pixel at position (x,y) from the first, second or third image respectively. I may add a getpn() function at some point, but this prevents static checking of the instructions against the number of images actually passed in.

value(c), hue(c), sat(c), hsv(h,s,v)

Separates a colour value into it's value (brightness), hue (colour) and saturation elements. Use hsv() to put them back together (after suitable manipulation).

red(c), green(c), blue(c), rgb(r,g,b)

Separates a colour value into it's red, green and blue colours. Use rgb(r,b,g) to put it back together.


Convert a value to an integer. Uses a C int cast, so it may break on large values.

if(cond,ntrue,nfalse), if(cond,ctrue,cfalse)

A simple (and inefficient) if function.


Relational operators (typically used with if()). Since we're working with floating point values the equalities are 'near equalities' - an epsilon value is used.

&amp;&amp;, ||, not(n)

Basic logical operators.

A few examples:

rpnexpr=>'x 25 % 15 * y 35 % 10 * getp1 !pat x y getp1 !pix @pix sat 0.7 gt @pat @pix ifp'

tiles a smaller version of the input image over itself where the colour has a saturation over 0.7.

rpnexpr=>'x 25 % 15 * y 35 % 10 * getp1 !pat y 360 / !rat x y getp1 1 @rat - pmult @pat @rat pmult padd'

tiles the input image over itself so that at the top of the image the full-size image is at full strength and at the bottom the tiling is most visible.

rpnexpr=>'x y getp1 !pix @pix value 0.96 gt @pix sat 0.1 lt and 128 128 255 rgb @pix ifp'

replace pixels that are white or almost white with a palish blue

rpnexpr=>'x 35 % 10 * y 45 % 8 * getp1 !pat x y getp1 !pix @pix sat 0.2 lt @pix value 0.9 gt and @pix @pat @pix value 2 / 0.5 + pmult ifp'

Tiles the input image overitself where the image isn't white or almost white.

rpnexpr=>'x y 160 180 distance !d y 180 - x 160 - atan2 !a @d 10 / @a + 3.1416 2 * % !a2 @a2 180 * 3.1416 / 1 @a2 sin 1 + 2 / hsv'

Produces a spiral.

rpnexpr=>'x y 160 180 distance !d y 180 - x 160 - atan2 !a @d 10 / @a + 3.1416 2 * % !a2 @a 180 * 3.1416 / 1 @a2 sin 1 + 2 / hsv'

A spiral built on top of a colour wheel.

For details on expression parsing see Imager::Expr. For details on the virtual machine used to transform the images, see Imager::regmach.pod.

Blending Images To put an image or a part of an image directly into another it is best to call the <CODE>paste()</CODE> method on the image you want to add to.


That will take paste $srcimage into $img with the upper left corner at (30,50). If no values are given for left or top they will default to 0.

A more complicated way of blending images is where one image is put 'over' the other with a certain amount of opaqueness. The method that does this is rubthrough.


That will take the image $srcimage and overlay it with the upper left corner at (30,50). The $srcimage must be a 4 channel image. The last channel is used as an alpha channel.


It is possible to add filters to the module without recompiling the module itself. This is done by using DSOs (Dynamic shared object) avaliable on most systems. This way you can maintain our own filters and not have to get me to add it, or worse patch every new version of the Module. Modules can be loaded AND UNLOADED at runtime. This means that you can have a server/daemon thingy that can do something like:

  load_plugin("dynfilt/")  || die "unable to load plugin\n";
  unload_plugin("dynfilt/") || die "unable to load plugin\n";
    || die "error in write()\n";

Someone decides that the filter is not working as it should - dyntest.c modified and recompiled.

  load_plugin("dynfilt/") || die "unable to load plugin\n";

An example plugin comes with the module - Please send feedback to if you test this.

Note: This seems to test ok on the following systems: Linux, Solaris, HPUX, OpenBSD, FreeBSD, TRU64/OSF1, AIX. If you test this on other systems please let me know.



An image object is a wrapper around the raw handle to an image. It is stored in the IMG value of the object hash. When Imager-&gt;new() is called the IMG member is set to undef by default but if you give it arguments like $img=Imager-&gt;new(xsize=&gt;100,ysize=&gt;100) then it will return a 3 channel image of size 100 x 100..

Functional interface



Use only if you cannot do what you need to do with the OO interface. This is mostly intended for people who want to develop the OO interface or the XS part of the module.

  $bool   = i_has_format($str);
  $colref = i_color_set($colref,$r,$g,$b,$a);
  $imref  = i_img_empty($imref,$x,$y);
  $imref  = i_img_empty_ch($imref,$x,$y,$channels);
  @iminfo = i_img_info($imref);
  $ch_msk = i_img_getmask($imref);

   $imref = i_scaleaxis($imref,$scale_factor,$axis);


   @bbox  = i_t1_bbox($fontnum,$pts,$str,$strlen);
   $imref = i_readjpeg($imref,$fd);
   $CPTI) = i_readjpeg($imref,$fd);
   $imref = i_readpng($imref,$fd);
   $imref = i_readgif($imref,$fd);
 $colour_list) = i_readgif($imref,$fd);

   $imref = i_readppm($imref,$fd);


This documentation is all very messy!

box, arc, circle do not support antialiasing yet. arc, is only filled as of yet. Some routines do not return $self where they should. This affects code like this, $img-&gt;box()-&gt;arc() where an object is expected.

When saving Gif images the program does NOT try to shave of extra colors if it is possible. If you specify 128 colors and there are only 2 colors used - it will have a 128 colortable anyway.

There are some undocumented functions lying around - you can look at the *sigh* ugly list of EXPORT symbols at the top of and try to find out what a call does if you feel adventureus.


Arnar M. Hrafnkelsson,


perl(1), Affix::Infix2Postfix(3).