1mkbitmap(1) General Commands Manual mkbitmap(1)
2
6 mkbitmap - transform images into bitmaps with scaling and filtering
7
9 mkbitmap [options] [filename...]
10
12 mkbitmap reads an image, and applies one or more of the following oper‐
13 ations to it, in this order: inversion, highpass filtering, scaling,
14 and thresholding. Each operation can be individually controlled and
15 turned on or off.
16 The principal use of mkbitmap is to convert color or greyscale images
18 into a format suitable as input for other programs, particularly the
19 tracing program potrace(1). It is particularly useful for converting
20 scanned line art, such as cartoons, handwritten text, etc., to high-
21 resolution bilevel images.
22 Highpass filtering can be used to ensure that foreground features such
24 as lines and text are preserved, while at the same time compensating
25 for uneven background. Optional blurring can be applied to smooth out
26 the image and remove visual noise. Scaling is important because a
27 scanned greyscale image contains more visual detail than a bilevel
28 image at the same resolution. By scaling the image to a higher resolu‐
29 tion (using interpolation) before thresholding it, some of this detail
30 is preserved. Thresholding means converting a greyscale image to a
31 bilevel image using only black and white pixels. Pixels that are darker
32 than a certain threshold value are converted to black. Optional inver‐
33 sion is useful if the input image shows bright features on dark back‐
34 ground, such as a picture of chalk drawings on a blackboard.
35 Supported input formats are PNM (PBM, PGM, PPM) and BMP. The output
37 formats are PBM for bitmaps, and PGM for greymaps.
38
40 General options:
41 -h, --help print help message and exit.
42 -v, --version print version info and exit.
44 -l, --license print license info and exit.
46 Input/output options:
48 filename If filename arguments are given, then mkbitmap will by
49 default create one output file for each input filename
50 given. The name of the output file is obtained from the
51 input filename by changing its suffix to ".pbm" or
52 ".pgm". If the name of the input file and output file
53 would be identical, then an additional suffix "-out" is
54 appended to the output filename. If no filename argu‐
55 ments are given, then mkbitmap acts as a filter, reading
56 from standard input and writing to standard output. A
57 filename of "-" may be given to specify reading from
58 standard input; the output for this argument will then
59 be written to standard output. Each input file may con‐
60 tain one or more images.
61 -o filename, --output filename
63 write output to this file. All output is concatenated
64 and directed to the specified file. This overrides the
65 default behavior of creating one output file for each
66 input file. A filename of "-" may be given to specify
67 writing to standard output.
68 Image processing options:
70 -x, --nodefaults
71 Turn off default options. Normally, the following
72 options are preselected by default: -f 4 -s 2 -3 -t
73 0.45. The -x option disables these defaults; thus,
74 mkbitmap -x does nothing but copy a greyscale image from
75 the input to the output. Other processing options can
76 then be added one by one; e.g., mkbitmap -xf10 does only
77 highpass filtering, mkbitmap -xt0.5 does only threshold‐
78 ing, etc.
79 -i, --invert Invert the input image. If this option is chosen, it is
81 applied to the image before any other operation. It is
82 used to deal with white-on-black images, such as photo‐
83 graphs of chalk drawings on a blackboard. Note that the
84 behavior of this option is not in general the same as
85 inverting the output bitmap, unless the thresholding
86 value is also inverted.
87 -f n, --filter n
89 Apply a highpass filter to the image. This filter is
90 approximately Gaussian and non-directional. The effect
91 is to preserve small detail while compensating for back‐
92 ground gradients. The parameter n is a radius (in pix‐
93 els) which corresponds approximately to the size of
94 details which should be preserved. More precisely, the
95 filter is implemented by subtracting a blurred version
96 of the image from the original image. The parameter n is
97 equal to the standard deviation of the blur. The output
98 of the filtering step is a normalized image whose aver‐
99 age brightness is exactly 0.5. The default filter radius
100 is 4.
101 -n, --nofilter Turn off highpass filtering.
103 -b n, --blur n Blur the image. The effect is to smooth out fine details
105 and to reduce visual noise in the image. The parameter n
106 is the blurring radius, and should be chosen small (1 is
107 a good value to start with). This is implemented as an
108 approximately Gaussian, non-directional blur with stan‐
109 dard deviation proportional to n. Blurring is applied
110 after the highpass filter, but before scaling and
111 thresholding. If this option is not given, the default
112 is not to apply any blurring.
113 -s n, --scale n
115 Scale the image by an integer factor n>0. Scaling is
116 done after highpass filtering, but before the threshold‐
117 ing step. A scaling factor of 1 indicates that no scal‐
118 ing is to be done. Otherwise, interpolation is used to
119 fill in the in-between pixels. If the output of mkbitmap
120 is to be used as input to a tracing program such as
121 potrace, a scaling factor of 2 is recommended. This pre‐
122 served the right amount of detail for the tracing algo‐
123 rithm to work well. If a scaling factor of 1 is used,
124 too much detail is lost. If a scaling factor of 3 or
125 higher is used, the interpolation tends to "invent"
126 detail which was not present in the original image, thus
127 preventing potrace from doing a good job.
128 -1, --linear Use linear interpolation when scaling to a higher reso‐
130 lution. This is slightly faster, but less nice, than the
131 default cubic interpolation.
132 -3, --cubic Use cubic interpolation when scaling to a higher resolu‐
134 tion. This is the default. It is slower than linear
135 interpolation, but leads to better results.
136 -t n, --threshold n
138 Set the threshold grey value for bilevel conversion. The
139 parameter n is a brightness value between 0 for black
140 and 1 for white. Any pixels below this brightness will
141 be converted to black (thus, smaller values of n will
142 lead to whiter output).
143 -g, --grey Disable bilevel conversion. If this option is given,
145 processing stops after the scaling step and a greymap is
146 output.
147
149 The exit status is 0 on successful completion, 1 if the command line
150 was invalid, and 2 on any other error.
151
153 1.16
154
156 Peter Selinger <selinger at users.sourceforge.net>
157
159 mkbitmap is distributed as part of the potrace package, and the latest
160 version is available from http://potrace.sourceforge.net/. This site
161 also contains documentation and information on how to obtain support.
162
164 potrace(1)
165
167 Copyright (C) 2001-2019 Peter Selinger
168 This program is free software; you can redistribute it and/or modify it
170 under the terms of the GNU General Public License as published by the
171 Free Software Foundation; either version 2 of the License, or (at your
172 option) any later version.
173 This program is distributed in the hope that it will be useful, but
175 WITHOUT ANY WARRANTY; without even the implied warranty of MER‐
176 CHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU General
177 Public License for more details.
178 You should have received a copy of the GNU General Public License along
180 with this program; if not, write to the Free Software Foundation, Inc.,
181 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. See also
182 http://www.gnu.org/.
183Version 1.16 September 2019 mkbitmap(1)