Spar 2011 Aldo Facchin

LiDAR and Stereo Imagery

Aldo Facchin – Geosoft Srl (Italy)

www.geosoft.it

Geosoft Srl - Italy

Who we are ?

An italian company focused on software products for …

• Map production

• 3D photogrammetry

• LiDAR

• Mobile Mapping (from vehicle)

Founded in 1992

Main markets (last 3 years) :

• 70-85 % Map, Survey and Engineering industry

• 15-30 % Public Administration, University and Research

Geosoft Srl - Italy

Our offer…

A family of ArcGIS Desktop stereoscopic extensions for :

Aerial photogrammetry and lidar (with a patented method)

Terrestrial Laser Scanning

Mobile Laser Scanning

Aerial Stereo Imagery and Lidar ArcGDS Stereo + LiDAR extension

Aerial Stereo Imagery and Lidar

Photogrammetry and Lidar are better if combined

Photogrammetry (3D imagery) alone is

not enough…

Optimal photointerpretation (image data

is continuous)

Use of RGB, CIR images with 8+ bit per

channel

Already experienced operators, well

known procedures and system

New digital cameras have bad B/H ratio

so Z quality is poor (respect to the

LiDAR data at the same

flight height)

Data quality is not geometrically

constant (Z in particular) because

depends on the operator (stereoscopic

perception)

Image correlation often have problems

exactly where we need (for ex. roof

edges)



LiDAR alone is not enough…

Optimal precision (especially Z) respect

to the photogrammetry at the same flight

height

Possibility to classify the points cloud

Data quality is geometrically constant

(Z in particular) because does not

depend on the operator (snap)

Data have low costs and short

acquisition times

Interpretation is very poor because the

data is discrete

Point density impacts a lot on cost

The points don’t have directly the color

of the objects

XY precision is poorer than Z



The solution :



The solution :

The operator move the cursor

observing the imagery stereo scene

The software update the cursor

height according with the point cloud,

using different alghoritms

Stereo imagery are used for XY and

photointerpretation

LiDAR point clouds are used for updating collimation mark

height (Z)

The point cloud is better than using DSM or other

resampled products

The operator has only the “check” the collimation

mark height (Z), not directly and carefully position it !

Live demo ArcGDS Stereo + LiDAR extension



Characteristics of ArcGDS LiDAR method (patent pending)

It’s a tool for the production of high quality 3D geodatabases.

Photointerpretation is available (photos, 3D...not orthophoto)

LiDAR acquisition allow much better Z precisions than photo-

mapping or orthophoto digitalization

Mapping (for ex. streets or buildings) is coherent with LiDAR

Operator checks Z during mapping and eventually changes it

manually

Is for everybody because not require a high photogrammetric

experience

Snap uses cloud classification

Transversal profile view helps operator for a better interpretation

Operator must control contemporary 3 windows



Key concepts :

Stereoscopy is for avaiable for all the users at low cost, inside ArcMap,

for images and point clouds

The time of proprietary CAD platform for maps data capture is over !

Combine stereo images and LiDAR point clouds to obtain the best

accuracy and efficiency

It’s the ideal tool for QC/QA after automatic feature extraction

phase

Can be bundled with raw data (images & point cloud) by data

providers

With this (patented) method…

• medium format cameras can be used

• low density point clouds can be used

Terrestrial Laser Scanning ArcGDS for TLS


During 3D feature extraction from laser scanner point clouds…

Operators spend a lot of time on moving and rotating the

scene

It’s easy to do mistakes on snapping laser point of different

planes

Often the images are used only to color the points

Use 3D glasses to save time, use images and point clouds as they are

The proposed solution :

Try to use 3D glasses to watch your point cloud !

Save time because you see your scene in 3D… don’t need

to continue to rotate

Use images “as images”



The solution :


The left view can be observed in 3D, rotated, zoomed, etc.

In the right view an image can be loaded to show the current

cursor position

Riegl V-line scanner are directly supported

Leica Scanstation C-10 internal camera support is under

development

Any oriented image can

be loaded

Collecting data inside a

geodatabase is the best

benefit !

Multiuser, versioning, etc.

are avaiable also in TLS

Why use a CAD ?


Live demo ArcGDS for TLS

Mobile Laser Scanning and Imagery ArcGDS for MLS


Use video to quickly navigate and 3D stereoscopic vision accurate data

collection

Now…

The current trend on MLS is to do all the work

using laser scanner(s) point cloud

The camera(s) are only for give color to point

cloud

Do you think to handle a mobile laser scanning

road network project with a CAD ?


Use video to quickly navigate and 3D stereoscopic vision accurate data

collection

Our idea…

Try to use camera image sequences (video) to

quickly navigate…

Use the point cloud only for the 3D precise

measure task

Use the images to recognize the objects you

have to collect

Use 3D glasses to save time



The interface :

Live demo ArcGDS for MLS

Conclusion

Images don’t replace point clouds and vice-versa

Use the best of each data type

Combine them

GIS is necessary for a good data collection

• for multiuser handling

• for versioning

• for data consistency

• etc.

There are no more advantages on using CAD for

3D

Thank you !

Grazie !

[email protected]

For further information, please visit Esri booth at SPAR International 2011.

mailto:[email protected]

Documents

Spar 2011 Aldo Facchin