readme

PDF417编码的算法的源程序

 The pdf417_encode program will take an input text file and convert
it to postscript output in pdf417 barcode format.  Type pdf417_enc -h
will give you a help list.  If you need the pdf417_enc in the old
3.9 format, you can make it by using Makefile.old.

 Usage: pdf417_enc [-t type] [-r nrows] [-c ncols] [-e ec_level] 
                          [-x X] [-y Y] [-q QZ] infile outfile

  [-t type] - You can now specify the type of output that is generated.
             Valid types are pbm, raw, ps, psbits,gif and eps.
             The default type is eps.
             The -t is optional.  The raw type will
             cause no output, but will put the barcode in an internal
             array of bits.  The option -t psbits will give only the
             bitmap portion of the postcript output.

 [-r nrows]  Number of rows in the barcode.
             This is an optional field.  If specified, it should < 90.
             However a special value of rows = 999 will cause the 
             barcode to grow to the size needed to fit the data.
             If no rows are specified, the default = 999 rows.

 [-e ec_level]  Specifies the error correction level to be used.
              Must be from 1 to 8.  The default if not specified is 5.
              This option can only be used if rows and columns are
              also specified.
        
              The ec_level will cause 2 * ( 2 ** ec_level) codewords
              to be used for error correction.  An ec_level of 5 will
              use 2 * ( 2 ** 5) or 64 codewords to be used for error
              correction.

 
 [-c ncols]  This is the number of columns in the barcode. Note that if
              rows is supplied, then columns must also be supplied.
              Columns must be less than 30.
              The default if no columns are given is 8.  

 [-x X]       Specifies the width in mils of the 1 unit bar.

 [-y Y]       Specified the height in multiples of the X value, of the
              barcode row.

 [-q QZ]      Specifies the width of the quiet zone thickness as a multiple
              of the X value.
              This applies to the ps and eps output.

  infile - The input file to the program.  The format for this file
           is given below.  If the infile = _ then stdin is used.

  outfile - The output file.  Note that for the postscript output this 
            is not a printable postscript file and lacks the headers necessary
            to print.  The ex.ps file is included to show how this
            postscript can be included in a printable postscript file.
            If the outfile = _ then stdout is used.

   To compile:   Modify the Makefile to reflect your system
                    ie UNAME = { Linux, FreeBSD, Other }
                 Also, you can modify it to select either a
                 libpdf417enc.a or libpdf417enc.so

                 make

                    will make the binary executable pdf417_enc

                 make lib

                    will make a callable libary libpdf417enc.a 
                    and a test program called lib_test

                 make clean
                     
                    remove all .o files and results of previous compile

                 make lib_test

                    creates a small program that tests the libpdf417enc.a

                 make libpdf417enc.a or make libpdf417enc.so

                    create the callable library for pdf417 routines
                    
 
    
   If the rows and columns are not specified the number of rows defaults
to 24 and the number of columns defaults to 8.


 The pdf program takes as input a file that consists of two columns.
The first column has in capital letters BC, TC, NC and MC.  BC indicates that
the ascii text will be byte compressed.  TC indicates that the text
will be text compressed and NC indicates that the text will be
numerically compressed.  MC is only used for specifying the File ID
and segment index for a Macro PDF ( multiple symbol) barcode.  
 You can embed special characters in the
Text compress mode (TC) by using \CR for carriage return, \LF for 
new line, \DQ for double quote and \NL for  carriage return followed by
line feed.   Also, use \BS for backslash itself.

The second column will have a text string beginning with and ending
with quotes.  For BC the text must consist of the numbers 0-9 and letters
A-F ( Hex 0-9, A-F).  Also, in BC if the length of the string is odd,
a zero will be prepended to the string to make it and even number.
For TC the text can consist of any of the legal
text compress charactors in the pdf417 standard.  For NC, the 
letters 0-9 are allowed in the text.

An example of a valid input file are given below.

BC "000197380300"
TC "062S_"
BC "940400F0BA54"
BC "0A00260200B5"
BC "0A3101DB6401"
BC "0064FB20BE0C"
BC "6159ADD80700"
BC "D08905003131"
BC "303081FCA4C1"
BC "004A9ABFC647"
BC "D057E68B8415"
BC "8464AF4255A5"
BC "09A2F5C14CC8"
BC "DD74DC72D3DD"
BC "DD402C90698B"
TC ""
TC ""


   I have included in this release some Linux Intel binaries for doing
 decoding of the pdf417_barcodes.

 The first program is scan_image.   
 This can read in the output of pdf417_encode and put out the 
 original text used to create the input.  Right now it only works for
 the output file I create with my pdf417_encode program.  Look at the
 example run.scan in this directory.

 The second program is called scan_data.  This performs the decode 
 function on data that is supplied as output from a  frontend program
 that reads  in a graphics ( pnm ) image file of a barcode.  This
 barcode graphics file is what might be read in using a flatbed scanner
 to scan in the barcode.  The format of the data input to scan_data
 consists of the row and column, followed by the code value ( ranging
 from 0 to 2787 ( 929 * 3). The code value = 9999 indicates that the
 module at that and row and column could not be decoded ( erasure ).


  The data is the output of the program called frontend.  Frontend can
 read in a graphics file ( .pnm ).  The .pnm file is obtained by scanning
 in a barcode using a flatbed scanner.  The frontend program will locate
 the barcode in the image ( must be surrounded by white space ) and 
 will figure out the slant of the barcode in the image.   The slant will
 be corrected and the barcode is then analyzed.  

Usage:
   frontend in.pnm [1-7]

     in.pnm is the input pnm file.
     Optionally followed by a number from 1-7

      If number is odd,  = write all the mods in the files outmodulex_y.pgm
                             for all x,y = row and column numbers
      If number has 2 bit set = write the final unslanted barcode in outf.pgm
      If number has 4 bit set = write final barcode columns in outcolx.pgm
                                write the final barcode in out.pgm

   so  for example

     frontend in.pnm 7
    
   will write out all the module and column .pgm files and the outf.pgm file.
   plus produce and out.data file for use with scan_text.

   See the examples in the directory  frontend.
 

   Finally, I have a pdf417_decode program that works simularly to the
  frontend, but combines the functionality of the scan_data program as well.


    Usage: pdf417_dec in.pnm [0-7]

   This will readin a scanned .pnm file with a barcode on it.  It will
 locate the barcode on the page,  unslant it and attempt to decode it
 ( with proper error correction if needed ).  The output text will be
 in the file out.txt.   Note that the .pnm used as input must be a 
 greyscale image ( ie a .pgm file).  The gimp can create a .pnm output
 file that is color also, so you MUST set gimp to output greyscale if
 you process your image thru the gimp.   The input file must have some
 amount of whitespace around the border to work also.

  I have provided a couple example in the directory pdf417_dec of the
 usage of this program.



  PDF417 Library

    You can now make a shared calleable library of the pdf417_enc.
  To generate the library type 

             make lib

   This will create libpdf417enc.a. This shared library can be copied
 to /usr/local/lib.

   An example of the use of library is given in the c code lib_test.c.
 The include file for the calleable routines is in pdf417.h
 and a script for compiling the lib_test example is in compile.lib_test.


  Mark Brand has contributed a pre-processor for pdf417_enc that implements
the appendix P of pdf417 spec.  It determines the optimum encoding of
a symbol for a given input text.   You can find the code for this
int the pdf417_prep subdirectory.  It is also incorperated into the 
pdf417_lib with the call to pdf417_prep_to_raw( const char* buffer, len);
This will take a raw buffer of text and use annex P to create the
the raw_input array used by pdf417_en and pdf417_en_new to create a symbol.