| 20 |
|
and columns are changing fastest in the input, which can have any number |
| 21 |
|
of components per matrix element. |
| 22 |
|
.PP |
| 23 |
+ |
The input must have a valid header, including dimensions and format, |
| 24 |
+ |
and a header will always be produced. |
| 25 |
+ |
This is similar behavior to |
| 26 |
+ |
.I rmtxop(1). |
| 27 |
+ |
If the input header is missing, one can usually be added with an |
| 28 |
+ |
appropriate call to |
| 29 |
+ |
.I rcollate(1). |
| 30 |
+ |
If an output header is not desired, it can be removed using |
| 31 |
+ |
.I getinfo(1). |
| 32 |
+ |
.PP |
| 33 |
|
One cannot specify crop boundaries outside the matrix dimension limits. |
| 34 |
|
If a 0 value is specified for the number of rows to copy, then all rows |
| 35 |
|
will be copied, starting from row0. |
| 36 |
|
Similarly, a 0 value for the number of columns to copy implies all |
| 37 |
|
columns >= col0 will be copied. |
| 38 |
+ |
A negative count for the number of rows or columns cuts the |
| 39 |
+ |
corresponding amount off the end(s). |
| 40 |
|
.PP |
| 41 |
|
Using |
| 42 |
|
.I rcrop |
| 48 |
|
However, one must remember that rows correspond to Y |
| 49 |
|
coordinates, and the orientation is reversed from a standard |
| 50 |
|
picture, where Y=0 is the bottom scanline. |
| 51 |
+ |
Spectral pictures, as might be produced by |
| 52 |
+ |
.I rtrace(1) |
| 53 |
+ |
or |
| 54 |
+ |
.I rfluxmtx(1), |
| 55 |
+ |
are also supported. |
| 56 |
|
.PP |
| 57 |
+ |
All matrix formats are supported, including text (ASCII), binary float |
| 58 |
+ |
and double. |
| 59 |
|
The |
| 60 |
|
.I rcrop |
| 61 |
|
tool also works on encoded normal vector files produced by |
| 62 |
|
.I rcode_norm(1) |
| 63 |
|
and encoded depth files produced by |
| 64 |
|
.I rcode_depth(1). |
| 65 |
< |
However, it is not able to crop identifier files produced by |
| 66 |
< |
.I rcode_ident(1). |
| 65 |
> |
Data is copied verbatim, i.e., there is no internal conversion to |
| 66 |
> |
another representation, although white space may change in text files. |
| 67 |
|
.PP |
| 68 |
|
If no input file is given, data will be taken from the standard input. |
| 69 |
|
If no output file is specified, data will be copied to the standard output. |
| 51 |
– |
All matrix formats are supported, including text (ASCII), binary float |
| 52 |
– |
and double. |
| 53 |
– |
Data is copied verbatim, i.e., there is no internal conversion to |
| 54 |
– |
another representation, although white space may change in text files. |
| 70 |
|
.SH EXAMPLES |
| 71 |
|
To crop a matrix down to 15 rows by 19 columns, |
| 72 |
|
starting at row 10 and column 12: |
| 81 |
|
starting from upper-left corner: |
| 82 |
|
.IP "" .2i |
| 83 |
|
rcrop 0 0 1000 1500 orig.hdr cropped.hdr |
| 84 |
+ |
.PP |
| 85 |
+ |
Two ways of computing the inner product of row 15 in matrix 1 by |
| 86 |
+ |
column 27 in matrix 2 and extracting the result without a header, |
| 87 |
+ |
the second command being more efficient: |
| 88 |
+ |
.IP "" .2i |
| 89 |
+ |
rmtxop -fa mat1.mtx mat2.mtx | rcrop 15 27 1 1 | getinfo - |
| 90 |
+ |
.IP "" .2i |
| 91 |
+ |
rmtxop -fa '!rcrop 15 0 1 0 mat1.mtx' '!rcrop 0 27 0 1 mat2.mtx' | getinfo - |
| 92 |
|
.SH AUTHOR |
| 93 |
|
Greg Ward |
| 94 |
|
.SH "SEE ALSO" |
| 95 |
< |
evalglare(1), pcompos(1), rcalc(1), rcode_depth(1), |
| 96 |
< |
rcode_ident(1), rcode_norm(1), rcollate(1), rfluxmtx(1), rmtxop(1) |
| 95 |
> |
evalglare(1), pcompos(1), rcalc(1), rcode_depth(1), rcode_norm(1), |
| 96 |
> |
rcollate(1), rcomb(1), rfluxmtx(1), rmtxop(1), rsplit(1), rtrace(1) |
