Changes between Version 11 and Version 12 of Procedures/AlsprocProcessing


Ignore:
Timestamp:
Sep 2, 2014 10:28:02 AM (5 years ago)
Author:
stgo
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • Procedures/AlsprocProcessing

    v11 v12  
    7272Note that you can only run one of pitch/roll/heading as incrementing at present, this is to stop confusing files but could be added in if it is felt it is needed.
    7373
    74 [[Be certain about your -inc value as this can result in very large numbers of files being produced.]]
     74'''Be certain about your -inc value as this can result in very large numbers of files being produced.'''
    7575
    7676You will then see a section for each flightline. process_line should be set to true for each flightline unless there was some problem, in which case the maintainer for alsproc should be contacted.
     
    9696}}}
    9797
    98 Log files will be kept for each flightline in the directory specified by log_dir in the configuration file, for your use if alsproc should suffer from some processing error.
     98To specify which flightline:
     99
     100
     101{{{
     102alsproc -f 140306_130407 2013144.cfg
     103}}}
     104
     105or to run on-grid:
     106
     107{{{
     108alsproc_submit.py 2014238.cfg
     109}}}
     110
     111Log files will be kept for each flightline in the directory specified by log_dir in the configuration file, for your use if alsproc should suffer from some processing error. If you are running on the grid separate log files directly from the gridnodes will be outputted as well, these can be of use when checking pre-processing errors.
    99112
    100113== Pitch and Roll Errors ==
    101 First you will need to generate an initial dataset to base some observations off
     114Follow the instructions above using the roll/pitch/heading correct values as these will be from the boresight for the current time period. Once you have set off generation wait for the files to create. This can take a while; 20 minutes to 20 hours depending on the size of the dataset and load on the network, it will be noticeably faster on the grid.
    102115
    103 {{{
    104 airborne@pmpc1322:/users/rsg/arsf/arsf_data/2013/flight_data/uk/GB13_05-2013_144_Glenmore/processing/als50# generate_alsproc_config
    105 Gathering details about project...
    106 Using calibration file (/users/rsg/arsf/calibration/lidar/20130219/als50.cfg)...
    107 Collating details for .cfg file...
    108 Writing .cfg file
    109 Scanning flightline 130524_130407... ok
    110 Scanning flightline 130524_130807... ok
    111 Scanning flightline 130524_131239... ok
    112 Scanning flightline 130524_131711... ok
    113 Scanning flightline 130524_132140... ok
    114 Scanning flightline 130524_132606... ok
    115 Scanning flightline 130524_133035... ok
    116 Scanning flightline 130524_133540... ok
    117 Scanning flightline 130524_134001... ok
    118 Scanning flightline 130524_134447... ok
    119 Scanning flightline 130524_134927... ok
    120 Scanning flightline 130524_135443... ok
    121 Scanning flightline 130524_135900... ok
    122 Scanning flightline 130524_140340... ok
    123 Scanning flightline 130524_140918... ok
    124 Scanning flightline 130524_141340... ok
    125 Scanning flightline 130524_141645... ok
    126 Scanning flightline 130524_142158... ok
    127 Scanning flightline 130524_142738... ok
    128 all done
    129 }}}
     116After creation is complete inspect the initial files to check for pitch (along track shifts) and roll (across track shifts) errors, it is unlikely that the boresight parameters will get you a perfect solution first time. Once you have an idea of the error you are looking at start using start/end/inc values to correct.
    130117
    131 Then edit the resulting .cfg file and set the appropriate fields. Example:
     118On early iterations, you may identify just a few flightlines and focus your efforts on those, but you '''must''' check flightlines from across the entire dataset, as pitch and roll has been known to change over the course of a flight in the most unfortunate cases.
    132119
    133 {{{
    134 output-projection = +init=epsg:27700 +nadgrids=/users/rsg/arsf/usr/share/proj/OSTN02_NTv2.gsb +geoidgrids=/users/rsg/arsf/usr/share/proj/osgm02_wgs.gtx
    135 sol-file = %(project_dir)s/posatt/ipas_honeywell/proc/20130524_103914.sol
    136 sup-file = %(project_dir)s/posatt/ipas_honeywell/proc/20130524_103914.sup
    137 }}}
    138 
    139 output-projection is set according to global position. This is a UK flight so I have selected the British National Grid.
    140 
    141 
    142 sol-file and sup-file have been set according to what was created during the navigation processing.
    143 
    144 
    145 Then process the dataset. This can take a while; 20 minutes to 20 hours depending on the size of the dataset and load on the network. On early iterations, you may identify just a few flightlines and focus your efforts on those, but you '''must''' check flightlines from across the entire dataset, as pitch and roll has been known to change over the course of a flight in the most unfortunate cases.
    146 
    147 {{{
    148 alsproc 2013144.cfg
    149 }}}
    150 
    151 Then go into the dataset and look for rooftops or other features both along and across areas of overlap. It should look like this:
     120The best method is to search for rooftops or other recognisable features both along and across areas of overlap using lag. It should look like this:
    152121
    153122[[Image(pitch_error_249-2012.jpg, align=center)]]