TerraMatch Airborne Laser Data

Calibration workflow

TerraMatch can be used to solve the mirror scale parameter and the misalignment between the laser scanner and the inertial measurement unit (IMU). The misalignment is expressed as heading, roll and pitch angular correction values which need to be known for every laser scanner system.

Different laser scanners may require additional calibration parameters as well. Those must be solved using more manual methods with TerraScan or with system specific software.

Flight pattern

The optimal site for a calibration flight contains both flat and sloped surfaces which do not have disturbing surface objects such as low vegetation. The most commonly used target area for calibration is an airport as it is easily accessible and most often contains suitable surfaces, such as the run ways and sloped building roofs.

The minimum flight pattern for calibration is four flight passes over the same area in a cross like pattern where the slopes surfaces are located at the center of the cross, as shown in the figure below.

Recommended Airborne Calibration Pattern

Recommended Airborne Calibration Pattern

You may consider some additional flight passes which may improve the quality of the calibration:

  • An additional flight pass for which the sloped surfaces are located at the right of left edge of the corridor covered. This helps to differentiate pitch and heading from each other.
  • Additional flight passes at a higher or a lower altitude.

Processing steps

The processing of a calibration flight can be outlined with the following steps:

  1. Solve GPS trajectories.
  2. Compute xyz laser points with system specific software using the last known calibration values.
  3. Import trajectories into TerraScan and transform them to any coordinate system.
  4. Split any trajectories which overlap themselves.
  5. Import time-stamped laser points into TerraScan and transform them to the same coordinate system.
  6. Make sure that the flightline numbering of the laser points matches trajectory numbers (in TerraScan, use Deduce using time command from Flightline pulldown menu).

Continue according to the matching method.

  • Surface-to-surface matching:
    1. Classify low points.
    2. Classify ground separately for each flightline.
    3. (Optional) Classify some buildings separately for each flightline.
    4. Smoothen ground surface if most of it is asphalt or some other hard surface.
    5. Run Find Match and solve for heading, roll, pitch and mirror scale corrections for the whole data set.
    6. Add the result values to the correction values used when computing the xyz points in step 2.
  • Tie line-based matching:
    1. Search for tie lines of type Surface lines. This does not require any classification of the laser points.
    2. Run Find Tie Line Match and solve for heading, roll, pitch and mirror scale corrections for the whole data set.
    3. Add the result values to the correction values used when computing the xyz points in step 2. ” The laser scanner may have operation modes which make calibration easier such as profile mode. If such a mode is available, it should be used to solve the pitch correction value first. Then TerraMatch should be used to solve for heading and roll correction only.

Project workflow

Even though the system has been calibrated, you may still find systematic errors in project data. TerraMatch can be used with actual project data to solve mismatches between laser data from different flightlines or between laser data and known points.

All parallel flightlines covering a project area should have a crossing flightline at both ends. For large project areas it is recommended to fly at least two or more crossing flightlines over sloped open terrain in order to provide good data for the matching task. Known points (ground control points) should be distributed close to the corners or edges of the project area.

As the project data volume can be huge, it is desirable to minimize the number of steps in the processing workflow. You probably want to run TerraMatch only if you notice that there are significant mismatches in the laser data.

Another difficulty with project data is that you do not know the nature of the errors beforehand. Mismatches may be a result of mistakes made during the setup of GPS reference stations, during computing trajectories or during operation of the airborne system. At some point within the correction workflow you must establish what parameters need to be corrected.

Processing steps

The general project workflow can be outlined with the following steps:

  1. Solve GPS trajectories.
  2. Compute xyz laser points with system specific software using the last known calibration values.
  3. Import trajectories into TerraScan.
  4. Import time stamped laser points into TerraScan.
  5. Compare flightlines visually in cross sections. Try to locate sloped surfaces both along flight direction and perpendicular to flight direction.
  6. If no significant mismatches are visible, you may skip the consecutive steps and continue with the normal processing workflow.
  7. Classify low points.
  8. Classify ground for each fligthline. If necessary, exclude water areas from the ground class.
  9. (Optional) Locate areas which are best suited for matching (crossing flightlines, visible clean sloped surfaces). Create a TerraScan project with blocks only in those areas.
  10. (Optional) Classify buildings for each flightline at the locations which are best suited for matching.

Continue according to the matching method.

  • Surface-to-surface matching:
    1. Run Find Match and solve for heading, roll, pitch and mirror scale corrections for the whole data set.

Apply the correction if it is significant.

  1. Run Find Match and solve for elevation or roll + elevation correction for individual flightlines.

Apply the correction if it is significant.

  1. Run Find Fluctuations and solve for fluctuating elevation correction.

Apply the correction if it is significant.

  • Tie line-based matching:
    1. Search for tie lines of type Surface lines based on ground and (optional) building class.
    2. Run Find Tie Line Match and solve for heading, roll, pitch and mirror scale corrections for the whole data set.

Apply the correction to the tie lines and the laser data if it is significant.

  1. Check the tie lines for worse observations.
  2. Run Find Tie Line Match and solve for heading, roll, pitch and elevation corrections for flightline groups (if there are any). Depending on the system, solve for mirror scale corrections per flightline groups as well.

Apply the correction to the tie lines and the laser data if it is significant.

  1. Check the tie lines for worse observations.
  2. Run Find Tie Line Match and solve for heading, roll, pitch and elevation corrections for individual flightlines.

Apply the correction to the tie lines and the laser data if it is significant.

  1. Check the tie lines for worse observations.
  2. Run Find Tie Line Fluctuations and solve for fluctuating elevation corrections.

Apply the correction to the laser data if it is significant.

” After each Apply corrections step you should check the flightlines visually in cross sections or using distance coloring to determine if the correction step improved the data. You have to decide if the correction was good or if you need to go back one step and try solving parameters with different settings.

Multi-day projects

Use GPS standard time to avoid conflicts between flight sessions with identical GPS seconds-ofweek time stamps. There are tools in TerraScan for converting between GPS seconds-of-week and GPS standard time and vice versa. The tools are applicable to trajectories and to laser data.

Apply a group number to trajectories for each flight session (day 1 = trajectory group 1, day 2 = trajectory group 2, etc.). Optionally apply also a quality tag to trajectories of a flight session, e.g. if flight conditions are worse on one of the days.

Reduce the amount of data for processing as much as possible. This includes thinning of trajectories during the import, but also reducing the number of blocks for matching in a TerraScan project. Use only blocks that are suitable for the matching task and only classes that are necessary (e.g. ground, building classes).

For adding a new data set to an already matched data set, apply quality “bad” to the trajectories of the new data set and quality “normal” or “good” to the trajectories of the old data set. Use Find Match or Find Tie Line Match and solve for corrections for the “bad” flightlines only.

Share

GeoCue Group Support

GeoCue Group Support has written 798 articles

2 thoughts on “TerraMatch Airborne Laser Data

  1. Avatar Venkat says:

    Hello,
    I have Las file. There is flight line difference between two strip s of 0.26m. I don’t have trajectories with me. I need to adjust it, Any one have idea?

    Cheers,
    Venkat

Leave a Reply