Marlborough - Blenheim LiDAR 1m DSM (2014)

Survey Specification:

-Scanner: Optech Sensor
-Flying Height: 900m above lowest ground
-Scan Angle: ±20 degrees
-Scan Frequency: 40Hz
-Pulse Rate: 70 or 75kHz
-Points Per M2: 1.4

Data Acquisition:

The LiDAR was acquired between 23 to 28 February 2014 and on 3 May 2014
using NZAM’s Optech LiDAR sensor systems.

The data in the tidal influence area was collected within 2 hours either side of low tide.

The data was collected flying 900 metres above lowest ground, and using a
scan angle of 20 degrees either side of nadir. The outgoing laser pulse rate
(PRF) was set at either 70kHz or 75kHz and mirror scan frequency 40 Hz. The
settings were selected to create a dataset with a minimum 1.4 point per square
meter pulse density in open ground.

New Zealand Aerial Mapping used a reference mark we established at
Woodbourne (NZAM0161) for the collection of GNSS base station data
during the aerial data acquisition.

Independent of the aerial survey work Sounds Surveying Ltd field surveyed a
series of topographic features in open ground, to be later used to verify the
accuracy of the processed ground dataset. This field survey work was
conducted between 10 and 13 March 2013. The field survey work is in terms
of LINZ geodetic marks AB83, AD84, BE0G, AD5R, A3TV, B3ED, A7GH,
BAK6, A867, ACYW, APW4 and BH3D.

Data processing:

The LiDAR sensor positioning and orientation (POS) was determined using
the acquired GPS/IMU datasets and Applanix POSPac software. This work
was all undertaken in NZGD2000 geodetic reference system using the data
collected at the LINZ geodetic reference marks for the differential processing.

The POS data was combined with the LiDAR range files and used to generate
LiDAR point clouds in NZTM map projection. This process was undertaken
using Optech LiDAR Mapping Suite software. As well as creating the point
clouds this software also checks the relative fit between points in the
individual swath overlap areas. All subsequent data processing steps were
undertaken using TerraSolid’s LiDAR processing software modules TerraScan
and TerraModeler.

The point cloud was checked for completeness of project coverage and pulse
density. Automated routines tailored to the projects land cover and terrain
were then setup and run on the data to create a ground classified point cloud.
This data was converted from ellipsoidal heights to Nelson 1955 vertical datum
using the LINZ NZGeiod09 grid and offset model.

Comprehensive manual editing of the LiDAR point cloud data was
undertaken to improve the quality of the automatically classified ground
point dataset. This editing involved visually checking over the data and
changing the classification of points into and out of the ground point dataset.
As part of this process LiDAR returns from water bodies were removed from
the ground point dataset and supplementary points were added to help support
hydro flow characteristics around and along water bodies.

The positional accuracy of the processed data has been checked by overlaying
the Sounds Surveying field surveyed features on the LiDAR dataset displayed
with its intensity values. The data was found to fit well in position. The
height accuracy of the classified ground points were checked by calculating
height difference statistics between a TIN of the LiDAR ground points and the
field surveyed points. The standard deviation statistic for the data is +/-0.06m.

The data from New Zealand Aerial Mapping was supplied in NZTM map projection
and the Nelson 1955 vertical datum, the point cloud datasets had been
tiled into NZTopo50 1:2000 tiles.

Classification of the point cloud followed the classification scheme below:

1 - Unassigned
2 - Ground
14 - Above_Ground

Data Reprocessing:

In 2017 the data was reprocessed by AAM New Zealand for LINZ relative
to the NZVD2016 vertical datum, NZGD2000 horizontal datum, NZTM map projection, with the NZGeoid2016. This data was supplied as 1:1000 nominal scale
(720m high x 480m wide tiles per full NZ Topo50 sheet).

Sounds Surveying Ltd undertook ground surveying to support the original work.
For this project, the Sounds Survey converted this dataset to NZVD2016. When doing so, they took accord of any local adjustments to geodetic marks that had been applied to the original field surveying coordinates.

Classification of the point cloud followed the classification scheme below:

2 - Ground
14 - Non Ground

The starting point for processing was the Classified point clouds produced for
Marlborough DC based on NZGeoid09. This data has been reverted to ellipsoidal
heights, then NZGeoid2016 has been applied to convert the data to orthometric
heights. The data was then compared to the ground surveying NZVD2016 test
points, and an appropriate final Z shift applied to each block.

Ground point data had been checked to check site data obtained by the field survey, this was assumed to be error free. The test points were distributed across the mapping area and located on clear open ground. Difference statistics yielded for Reference Project Site Awatere: RMS 0.035m and for Wairau: RMS 0.045m

All point attribute data (intensity, return number, GPS time) in the original data
points have been imported into the new las file dataset. Lakes and large rivers were
hydroflattened in the bare earth digital elevation model. The deliverables to LINZ were:

1m gridded bare earth digital elevation model (DEM)
1m gridded digital surface model (DSM)
Classified point cloud

Classified point cloud Data hosted by OpenTopography was re-classified: the Above_Ground (14) points were reclassified as Unassigned classification (1)