Abstract
This section will explain how to install a scanner through ScannerDrake, and how to use it with SANE and XSane (scanner interface software). We will also present a list of other scanner interface software you could use with GNU/Linux.
![]() | Tip |
---|---|
Please note that while support for scanners is very good, not all scanners are supported under GNU/Linux: before buying new hardware, it is always a good idea to visit Mandrakesoft's Hardware Database and the SANE home page to check for compatibility issues. |
ScannerDrake is the Mandrakelinux scanner
detection and configuration tool. This wizard will help you
install your scanner. Make
sure your scanner is powered on and launch ScannerDrake by
clicking on the Scanners entry of the
Mandrakelinux Control Center's Hardware section. You will first
be prompted to install some necessary software, simply click on
Yes.
The program will try to detect the manufacturer and model of your scanner. If it finds one which is ready to use it will show some information about it in the upper part of the wizard's main window. If the detection failed and no scanners were identified, you will be shown this window:
You can try to perform a new search by clicking on , which is useful if you have just plugged a scanner in. If the automatic detection fails, click on and look for the specific model you own by first expanding the manufacturer's section, then browsing through the list of available models.
After choosing the
appropriate model, you can leave the default Auto-detect
available ports option and click on
OK. If you have a parallel port scanner,
selecting /dev/parport0
in the pull-down list
should be the right choice.
Your scanner should now be installed and you should be ready to use the programs which come with SANE, XSane or other acquisition software.
![]() | Warning |
---|---|
Note that HP multi-function devices, such as the OfficeJet and PSC printers, must be configured through PrinterDrake. Please refer to Section 6, “Configuring Printers with PrinterDrake”. The scanning part of non-HP multi-function devices can be set up with ScannerDrake as a stand-alone scanner. |
To test that everything works correctly, launch xscanimage[3] from a terminal and try to acquire a picture from your scanner. You may first acquire a preview of the scanned image by clicking on the Preview window button, as shown in Figure 3.27, “Acquiring Images with xscanimage”.
Note that xscanimage may also be invoked directly from GIMP by choosing + + -> from the program menu.
ScannerDrake allows for scanner sharing between users connected via a LAN. Installation is very easy: just click on and either select or depending on what you want to do. With these buttons you can define which machines are allowed to use your scanners and which remote scanners you want to use.
While xscanimage is more than enough for your basic scanning needs, more experienced and/or graphic-oriented users will want to use a more sophisticated program. XSane[4] offers more options and a more informative display during the image acquisition process.
You can launch XSane through the + -> menu item. You'll see several windows pop up on the screen.
If the package
xsane-gimp
is installed, you will have a
GIMP plugin at your disposal. It will allow you to import
your images directly into GIMP for image retouching
tasks. To do this, simply choose
+ -> to launch XSane. You can
now scan your image and have it sent directly to
GIMP.
Most modern scanners boast high resolutions, such as 1200, 1600 or 2400 DPI (Dots Per Inch). But it would be a mistake to perform all of your scanning at the maximum available resolution. You will notice very little — if any — quality difference between a 300 and a 600 DPI image scan, but the file size will grow exponentially, up to many MBs of disk space for a single image file.
The resolution value should be chosen according to the device on which the image will be reproduced. For images to be viewed on computer monitors, e.g. for web sites, the resolution should be close to typical monitor resolution values, between 70 and 100 DPI. Higher values will result not only in bigger images, but the dimensions will also increase, so that an image scanned at 160 DPI instead of 80 will be about twice as large[5].
If you intend to print your images, a resolution of 300 DPI should be enough for most home printers. Increase this value if you have a very high quality printer.
Higher values should be chosen only for specific uses, such as enlarged images on very high quality printers, or quality scans of black and white originals. You will have to experiment a little until you're satisfied with the results.
By installing the
kdegraphics-kooka
and
ocrad
packages, you'll be able to use
Kooka, a simple graphical front-end to
SANE which is also able to perform
OCR tasks. To launch it, choose -> from the main
menu.
First you will be asked to choose the scanning device and then Kooka's main window will appear.
Figure 3.30, “Kooka's Main Window” shows a screen shot of a scanned image. To do so place the image or book you want to scan onto your scanner and click on at the bottom left of Kooka's window. For better results, you should choose the Grayscale Scan mode. Then adjust the resolution (200 DPI should be enough), select the part of the image to be scanned and click on .
To benefit from Kooka's OCR features, you must configure it. Access the -> menu, click on OCR, and select the OCR engine you want to use.
Once this is done you can click on
this icon ( -> ) and click on
. The resulting text will
appear along with a spell correction window.
Here is a list of other scanner interface software which is known to work under GNU/Linux.
Unfortunately, OCR (Optical Character Recognition) programs are not as common or sophisticated as their image acquisition counterparts. However there are at least two programs which are mature enough: Clara OCR and GOCR. You'll find all the necessary packages on the Mandrakelinux CDs.
Clara's graphical interface is very simple and doesn't require a specific desktop environment, but it will be necessary to train the program on the scan of a sample page. The training process can be quite tedious, and not very intuitive for the new user[6], so be sure to read the tutorial file. More experienced users should also refer to its Advanced User's Manual.
GOCR is another project
which has reached a usable status, but bear in mind that it is
still beta software. GOCR can read
images in many formats, and will produce a simple text file as its
output. It's a command-line tool, so if you want to use a
graphical front-end you will need to install the
gocr-gtk
package, although some image capture
applications such as XSane offer limited
OCR capabilities through it. You can launch it
through the + -> sub-menu.
Users of the FLTK (“Fast Light Tool Kit”) graphic user interface could try FlScan, a FLTK front-end for SANE.
For EPSON scanners, you could download Image Scan! for Linux, a scanner utility provided free of charge to GNU/Linux users by EPSON KOWA Corporation.
While multi-functional HP devices are configured using PrinterDrake, owners of these devices should have a look at the HP Linux Inkjet Project. The developers involved in the project aim at providing GNU/Linux support for most Hewlett-Packard OfficeJet, PSC, LaserJet, and PhotoSmart printer multi-function peripherals (MFPs).
[3] The sane-frontends
package must be installed.
[4] The xsane
package
must be installed.
[5] However, scanning at a higher resolution, and then reducing the resulting image size using graphic manipulation software such as GIMP, is one method often used to obtain better results than directly scanning at the desired final resolution.
[6] To quote the authors: “Clara OCR is not simple to use. A basic knowledge about how it works is required for using it”.