1RECOVERJPEG(1) RECOVERJPEG(1)
2
6 recoverjpeg - recover jpeg pictures from a filesystem image
7
9 recoverjpeg [options] device
10
12 Recoverjpeg tries to identify jpeg pictures from a filesystem image.
13 To achieve this goal, it scans the filesystem image and looks for a
14 jpeg structure at blocks starting at 512 bytes boundaries.
15 Salvaged jpeg pictures are stored by default under the name im‐
17 ageXXXXX.jpg where XXXXX is a five digit number starting at zero. If
18 there are more than 100,000 recovered pictures, recoverjpeg will start
19 using six figures numbers and more as soon as needed, but the 100,000
20 first ones will use a five figures number. Options -f and -i can over‐
21 ride this behaviour.
22 recoverjpeg stores the recovered pictures into the current directory.
24 If you want it to store them elsewhere, just go to the directory you
25 want recoverjpeg to save the images into (using the cd command at the
26 shell prompt) and start recoverjpeg from there, or use the -o option.
27 Note that device is not necessarily a physical device. It may also be
29 a file containing a copy of the faulty device in order to reduce the
30 actual processing time and the stress imposed to an already defective
31 hardware. dd(1) or ddrescue(1) may be used to create such a working
32 copy.
33
35 -h Display an help message.
36 -b blocksize
38 Set the size of blocks in bytes. On most file systems, setting
39 it to 512 (the default) will work fine as any large file will be
40 stored on 512 bytes boundaries. Setting it to 1 maximize the
41 chances of finding very small files if the filesystems aggre‐
42 gates them (UFS for example) at the expense of a much longer
43 running time.
44 -d formatstring
46 Set the directory format string (printf-style, default: use the
47 current directory). When used, 0 will be used for the 100 first
48 images, 1 for the 100 next images, and so on. The goal of this
49 option is to circumvent the directory size limit imposed by some
50 file systems.
51 -f formatstring
53 Set the file name format string (printf-style, default: "im‐
54 age%05d.jpg"). It is used with the image index as an integer
55 argument.
56 -i integerindex
58 Set the initial index value for image numbering (default: 0).
59 -m maxsize
61 Maximum size of extract jpeg files. If a file would be larger
62 than that, it is discarded. The default is 6 MiB.
63 -o directory
65 Change the working directory before restoring files. Use this
66 option to restore files into a directory with enough space in‐
67 stead of the current directory. This option can be repeated.
68 -q Be quiet and do not display anything.
70 -r readsize
72 Set the readsize in bytes. By default, this is 128 MiB. Using
73 a large readsize reduces the number of system calls but consumes
74 more memory. The readsize will automatically be adjusted to be
75 a multiple of the system page size. It must be greater than the
76 maxsize parameter.
77 -s cutoffsize
79 Set the cutoff size in bytes. Files smaller than that will be
80 ignored.
81 -S skipsize
83 Set the number of bytes to skip at the beginning of the filesys‐
84 tem image. This can be used to resume an interrupted session,
85 in conjunction with -i. The number of bytes may be rounded down
86 to be a multiple of a memory page size in order to improve per‐
87 formances.
88 -v Be verbose and describes the process of jpeg identification. By
90 default, if this flag is not used, recoverjpeg will print a
91 progress bar showing how much it has analyzed already and how
92 many jpeg pictures have been recovered.
93 -V Display program version and exit.
95 All the sizes may be suffixed by a k, m, g, or t letter to indicate
97 KiB, MiB, GiB, or TiB. For example, 6m correspond to 6 MiB (6291456
98 bytes).
99
101 Recover as many pictures as possible from the memory card located in
102 /dev/sdc:
103 recoverjpeg /dev/sdc
105 Do the same thing but ignore files smaller than one megabyte:
107 recoverjpeg -s 1m /dev/sdc
109 Recover as many pictures as possible from a crashed ReiserFS file sys‐
111 tem (which does not necessarily store pictures at block boundaries) in
112 /dev/sdb1:
113 recoverjpeg -b 1 /dev/sdb1
115 Do the same thing in a memory constrained environment where no more
117 than 16MB of RAM can be used for the operation:
118 recoverjpeg -b 1 -r 16m /dev/sdb1
120
122 Copyright (c) 2004-2016 Samuel Tardieu <sam@rfc1149.net>. This is free
123 software; see the source for copying conditions. There is NO warranty;
124 not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
125 If recoverjpeg saves your day and you liked it, you are welcome to send
127 me the best rescued ones by email (please send only 800x600 versions of
128 the pictures) and authorize me to put them online (indicate which con‐
129 tact information you want me to use for credits).
130
132 recovermov(1) sort-pictures(1) remove-duplicates(1)
133
135 Recoverjpeg does not include a complete jpeg parser. You may need to
136 use sort-pictures afterwards to identify bogus pictures. Some pictures
137 may be corrupted but have a correct structure; in this case, the image
138 may be garbled. There is no automated way to detect those pictures
139 with a 100% success rate.
140
142 Samuel Tardieu <sam@rfc1149.net>.
143Recoverjpeg User Manuals November 12, 2016 RECOVERJPEG(1)