20. Importing Images
While you can always make your own images, you may sometimes want to
import and use existing images from other sources. In this chapter, I'll
show how to import images from scanners and Kodak PhotoCD discs. We'll
begin with recipes for taking screen shots.
20.1 Taking Screen Shots
A screen shot is a picture of all or part of the display
screen. The following recipes show you how to take screen shots in X and
in the console.
20.1.1 Taking a Screen Shot in X
@sf{Debian}: `imagemagick'
@sf{WWW}: ftp://ftp.wizards.dupont.com/pub/ImageMagick/
Use
import, part of the ImageMagick suite, to take a screen shot
in X.
import can capture the entire screen, a single window, or
an arbitrary rectangular area, taking as an argument the name
of the file to save to. As with other ImageMagick tools, the image
format of the output file depends on the file extension you specify:
`.eps' for EPS,
`.tiff' for TIFF,
`.jpeg' for JPEG, and
so on. (For a complete list, see
Converting Images between Formats).
After you give the command, the mouse pointer changes to a set of
cross-hairs. You then use the mouse to specify which window to take the
shot of, as follows:
-
Left-click on a window to capture it.
-
Left-click on the root window to capture the entire screen.
-
Left-click and drag the mouse across an area of the screen to form a
rectangular selection outline; release the mouse button to capture the
selected area.
When you specify a window, import captures only the window's
contents; use the `-frame' option to include the window manager
frame in the image.
-
To capture a particular window, including its window manager frame, and
write it to a PNG-format file, first type:
| | $ import -frame session-1.png RET
|
-
Then, left-click on the window you want to capture.
In this example, the capture is saved to a file called
`session-1.png'.
NOTE: The system bell rings once when the screen capture
starts, and twice when the captures finishes.
20.1.2 Taking a Screen Shot in a Console
To take screen shots in a virtual console, use cat to save the
contents of the device file corresponding to that virtual console; these
files are in the `/dev' directory, and are in the format
`vcsnumber', where number is the number of the virtual
console.
For example, if the target console is the first virtual console (which
you would see by typing ALT-F1), the device to
cat is `/dev/vcs1'.
-
To take a screen shot of the fourth virtual console, and save it to a
file called `screenshot', type:
| | $ cat /dev/vcs4 > screenshot RET
|
NOTE: You must have superuser privileges to access these files
(see section Administrative Issues).
Take the screen shot from a virtual console different from the one you
want to take a shot of; if you try to take it from the same console you
want to capture, the command line you give will be included in the shot!
(Kind of like having your thumb in front of the lens while taking a
photograph.)
Screenshots taken of virtual consoles, as shown here, are saved as text
files; you can't take screen shots of virtual consoles when graphics are
displayed.
20.2 Scanning Images
@sf{Debian}: `sane'
@sf{WWW}: http://www.mostang.com/sane/
SANE, "Scanner Access Now Easy," is the de facto Linux scanner
interface; use it to scan an image with a scanner and save it to a file.
SANE works with a wide array of scanning hardware, but make sure the
scanning hardware you want to use is compatible by checking the
Hardware HOWTO and
SANE's list of
supported scanners.
Once you have SANE running, you can scan images with SANE-aware
applications like the GIMP (see section Editing Images with the GIMP).
The following recipes describe use of the command-line scanimage
tool, which comes with the SANE package.
NOTE: As the acronym implies, getting a scanner to work on a
Linux system hasn't always been smooth going. The SANE interface is
completely open, and its developers are making sure that it is
generalized enough to be implementable on any hardware or operating
system.
20.2.1 Listing Available Scanner Devices
Before you can use a scanner device, you need to know its device
name. To get this name, use scanimage with the
`--list-devices' option.
-
To list available scanner devices, type:
| | $ scanimage --list-devices RET
device `umax:/dev/sgb' is a UMAX Astra 1220S flatbed scanner
$
|
In this example, there's one scanning device on this system, a UMAX
brand scanner that can be specified to scanimage by giving its
device name, `umax:/dev/sgb', as an argument to the `-d'
option.
To list the available resolutions and options supported by a particular
device, use the `--help' option along with the `-d' option
followed by its device name.
-
To list available options supported by the device listed in the previous
example, type:
| | $ scanimage --help -d 'umax:/dev/sgb' RET
|
NOTE: For all scanimage commands, specify the scanner
device you want to use by including the `-d' option with the device
name.
20.2.2 Testing a Scanner
To run diagnostic tests on a scanner to make sure that it can be
properly read from, use scanimage with the `--test' option.
-
To test the UMAX scanner listed previously, type:
| | $ scanimage --test -d 'umax:/dev/sgb' RET
|
20.2.3 Scanning an Image
@sf{Debian}: `netpbm'
@sf{WWW}: http://www.debian.org/Packages/stable/graphics/netpbm.html
Use
scanimage to scan an image. Most scanners let you specify the
x and
y values, in pixels, for the image size to scan,
starting from the top-left corner of the scanner bed. Give these
coordinates as arguments to the
`-x' and
`-y' options. Also,
give an argument to the
`--resolution' option to specify the scan
resolution, given in dpi ("dots per inch"). Common resolution values
include 72, 120, 300, and 600 dpi; 72 dpi is the most popular resolution
for use on the Web or for viewing on screen, and 204 dpi is often used
for images that you want to send on a fax machine.
Scanned output is sent to standard output, so to scan an image to a
file, redirect the standard output.
scanimage outputs images in the PNM ("portable anymap")
formats, so make sure that you have the `netpbm' package (installed
on most Linux systems by default); it's a useful collection of tools for
converting and manipulating these formats. The formats output by
scanimage are as follows:
| FORMAT |
DESCRIPTION
|
PPM |
Color images.
|
PBM |
Black and white images.
|
PGM |
Grayscale images.
|
Use the `--mode' option to specify the format of the output,
followed by one of the following arguments: `color' for color PPM,
`gray' for PGM grayscale, or `lineart' for black and white
PBM. Each scanner has a default mode; for most color scanners, the
default mode will be `color'.
-
To make a 72 dpi scan of a color image 200 pixels wide and 100 pixels
tall, using the UNIX scanner from previous examples, and writing to a
file called `scan.ppm', type:
| | $ scanimage -d umax:/dev/sgb --resolution 72 -x 200 -y 100 >
scan.ppm RET
|
-
To make a 300 dpi scan of a black and white image 180 pixels wide and
225 pixels tall, using the UMAX scanner from previous examples,
and writing to a file called `scan.pbm', type:
| | $ scanimage -d umax:/dev/sgb --resolution 300 --mode lineart
-x 180 -y 225 > scan.pbm RET
|
NOTE: The command lines in this recipe are split across two
lines because they're too long to fit on one, but type these commands on
one long line.
Once the image has been scanned and written to a file, you can edit it
just as you would any image.
20.3 Extracting PhotoCD Images
@sf{Debian}: `xpcd'
@sf{WWW}: http://user.cs.tu-berlin.de/~kraxel/linux/xpcd/
There are two methods to extract an image from Kodak
PhotoCD
(30)
("PCD"). If you are browsing the disc with the
xpcd tool, then
choose an image, extract a copy at the desired resolution, and save it
to a file, as described in
Browsing PhotoCD Archives.
You can also use pcdtoppm on a PCD file directly to extract an
image at a given resolution and save it to a file in PPM format. Use the
`-r' option to specify the resolution to extract, given as a
numeric argument from 1 (lowest resolution) to 5 (highest); if this
option is omitted, a value of 3 is assumed. Also give as arguments the
name of the PCD file to read from and the name of the PPM file to write
to.
-
To extract the highest resolution from the file `slack.pcd'
and save it to a PPM file named `slack.ppm', type:
| | $ pcdtoppm -r5 slack.pcd slack.ppm RET
|
20.3.1 Converting a PhotoCD Image
Once you extract a PhotoCD image and write it to a PPM format file, use
convert to convert it to another format and adjust or improve the
image (see section Converting Images between Formats).
To improve the image while you convert it to JPEG
format, specify no interlacing with the `-interlace' option, 50
percent image sharpening with the `-sharpen' option, and add an
optional border and annotation to the image with the `-border' and
`-comment' options.
-
To convert the file `slack.ppm' to non-interlaced JPEG, sharpen the
image, add a two-pixel by two-pixel border, and annotate the image,
type (all on one line):
| | $ convert -interlace NONE -sharpen 50 -border 2x2 -comment
'Bob was here' slack.pnm slack.jpeg RET
|
20.3.2 Removing PhotoCD Haze
@sf{Debian}: `gimp'
@sf{WWW}: http://www.gimp.org/
Extracted PhotoCD images are known to sometimes have a kind of "green
haze" over them; to remove it, open the image in the GIMP and adjust
the color levels with the
Auto Levels function. This technique,
adapted from a tip for using PhotoCD by
Philip Greenspun, works well for
improving any scanned or imported image.
-
To remove the "green haze" from a PhotoCD image, do the following:
-
First, open the extracted image in the GIMP (see section Editing Images with the GIMP).
-
Then, click through the
Image menu to the Colors submenu
and then to the Levels submenu, and choose Auto Levels.
-
Click
OK in the Levels window to accept the changes.
This document was generated
using texi2html
