1Pnmhisteq User Manual(0) Pnmhisteq User Manual(0)
2
3
4
6 pnmhisteq - histogram equalize a PNM image
7
8
10 pnmhisteq
11
12 [-gray]
13
14 [-noblack] [-nowhite]
15
16 [-rmap pgmfile]
17
18 [-wmap pgmfile]
19
20 [-verbose]
21
22 [pnmfile]
23
24
25
27 This program is part of Netpbm(1).
28
29 pnmhisteq increases the contrast of a PGM or PPM image through the
30 technique of "histogram equalization."[1]
31
32 pnmhisteq computes a histogram of the luminosity of the pixels in the
33 image. It then calculates a mapping between each luminosity and a new
34 luminosity such that it spreads out intensity levels around histogram
35 peaks and compresses them at troughs. I.e. it moves pixels around in
36 the histogram so as to make it flat. It applies that mapping to the
37 input image to produce the output image. The effect of this is that
38 the image has equal numbers of pixels at each possible intensity level,
39 which means it uses the available levels of intensity more efficiently
40 and thereby has more visible detail.
41
42 Mathematically, the luminosity mapping is this: Assume the pixels are
43 sorted by luminosity into B buckets numbered from 0 (lowest luminosity)
44 to B-1. N[i] is the number of pixels in bucket i. T is the total num‐
45 ber of pixels (sum of N[i] over all i). W is the luminosity of white.
46
47 pnmhisteq replaces an input pixel whose luminosity falls into bucket j
48 with one whose luminosity is:
49
50
51 j
52 ---
53 \
54 > (N[i] / T) * W
55 /
56 ---
57 i=0
58
59 Considering a grayscale image for simplicity, this means that pixels in
60 the most luminous bucket become white. Pixels in the 10th per centile
61 of luminosity become 10% of white.
62
63 pnmhisteq maps a single luminosity in the input to a single luminosity
64 in the output. That means if pixels A and B both have luminosity .2 in
65 the input, and pixel A has luminosity .4 in the output, pixel B also
66 has luminosity .4 in the output. And since the luminosities in the
67 input are not continuous, the luminosities in the output aren't either
68 and pnmhisteq doesn't meet the ideal of having exactly the same number
69 of pixels of each luminosity in the output.
70
71 If you're processing a related set of images, for example frames of an
72 animation, it's generally best to apply the same luminosity mapping to
73 every frame, since otherwise you'll get distracting frame-to-frame
74 changes in the brightness of objects. pnmhisteq's -wmap option allows
75 you to save, as a PGM image, the luminosity map it computes from an
76 image. The -rmap option causes pnmisteq to use such an image as its
77 luminosity map.
78
79 So you can run pnmhisteq with -wmap on a composite you created with
80 pnmcat of the images you intend to process. Then, you can run pnmisteq
81 with -rmap on each of the individual images, using the luminosity map
82 you generated from the composite.
83
84 Use pnmhistmap to see the result. Run a color image through ppmtopgm
85 first so that you see a histogram of the luminosity instead of his‐
86 tograms of the three color components. It should generally show a flat
87 histogram. But because of the quantization effects described above,
88 you might see high bars interleaved with low bars, with the local aver‐
89 age being flat. To see local averages, use the -width option of
90 pnmhistmap.
91
92
93
95 You can abbreviate any option to its shortest unique prefix.
96
97
98
99 -gray When processing a color image, only gray pixels (those with
100 identical red, green, and blue values) are included in the his‐
101 togram and modified in the output image. This is a special pur‐
102 pose option intended for images where the actual data are gray
103 scale, with color annotations you don't want modified. Weather
104 satellite images that show continent outlines in color are best
105 processed using this option. The option has no effect when the
106 input is a graymap.
107
108
109 -noblack
110 Do not include black pixels in the equalization. The black pix‐
111 els in the output are exactly the black pixels in the input and
112 the number of black pixels does not affect the color of any
113 other pixels.
114
115 Sometimes, black isn't as much a color as a background or anno‐
116 tation for the real colors, so you want to treat it specially
117 this way. Think of a picture of stars, which is nearly all
118 black, but with lots of stars of different brightness. You want
119 to change the brightnesses of the stars to maximize contrast
120 between them, but if you considered the blackness to be signifi‐
121 cant, all the stars would end up close to full white.
122
123 This option was new in Netpbm 10.70 (March 2015).
124
125
126 -nowhite
127
128 Same as -noblack, but for the white pixels.
129
130 This option was new in Netpbm 10.70 (March 2015).
131
132
133 -rmap mapfile
134 Process the image using the luminosity map specified by the PGM
135 file mapfile.
136
137 The PGM image, usually created by an earlier run of pnmhisteq
138 with the -wmap option, contains a single row with number of col‐
139 umns equal to the maxval (greatest intensity value) of the image
140 plus one. Each pixel in the image is transformed by looking up
141 its luminosity in the corresponding column in the map file (col‐
142 umn number = luminosity) and changing it to the value given by
143 that column.
144
145
146 -wmap mapfile
147 Creates a PGM file mapfile, containing the luminosity map com‐
148 puted from the histogram of the input image. This map file can
149 be read on subsequent runs of pnmhisteq with the -rmap option,
150 allowing a group of images to be processed with an identical
151 map.
152
153
154 -verbose
155 Prints the histogram and luminosity map on Standard Error.
156
157
158
159
161 Histogram equalization is effective for increasing the visible detail
162 in scientific imagery and in some continuous-tone pictures. It is
163 often too drastic, however, for scanned halftone images, where it does
164 an excellent job of making halftone artifacts apparent. You might want
165 to experiment with pnmnorm and pnmgamma for more subtle contrast
166 enhancement.
167
168 The luminosity map file supplied by the -rmap option must have the same
169 maxval as the input image. This is always the case when the map file
170 was created by the -wmap option of pnmhisteq. If this restriction
171 causes a problem, simply adjust the maxval of the map with pamdepth to
172 agree with the input image.
173
174 If the input is a PBM file (on which histogram equalization is an iden‐
175 tity operation), the only effect of passing the file through pnmhisteq
176 will be the passage of time.
177
178
180 pnmnorm(1), pnmcat(1), pamdepth(1), pnmgamma(1), pnm(1),
181
182
183
184 [1] Russ, John C. The Image Processing Handbook. Boca Raton: CRC
185 Press, 1992. Pages 105-110.
186
188 This manual page was generated by the Netpbm tool 'makeman' from HTML
189 source. The master documentation is at
190
191 http://netpbm.sourceforge.net/doc/pnmhisteq.html
192
193netpbm documentation 22 March 2015 Pnmhisteq User Manual(0)