119
DIGITAL AUTOMATIC ORTHOPHOTO PRODUCTION
WITH LASER LOCATOR AND AERIAL PHOTOGRAPHY DATA
Evgueny Medvedev
Opten Limited, Moscow, Russian Federation
Commission V, Working Group 4
KEY WORD: Laser scanning, digital aerial camera, orthophoto rectification, DTM, IMU
ABSTRACT
Application of traditional aerial survey technologies for topographic purposes has a number of principal problems.
The most serious problem is definitely impossibility of automatic use of classical stereophotogrammetry methods in
cases of complex terrain, urban areas, the areas with no visual texture and so on.
It is well recognized by now that laser locator methods are very efficient for such kinds of work. The proposed
approach is based on the following main principles:
1) Original 3D nature of laser locator data in combination with an absolute geodetic accuracy of 10-15 cm,
unachievable to traditional methods of aerial topography survey, allows to suggest a number of effective algorithms of
fully automatic recognition, delineation and positioning for the objects such as Digital Terrain Model (DTM), water surface
contours, forest arrays, single trees, buildings.
2) Laser locator survey can be successfully combined with digital aerial photography during both on-board data
acquisition and on-ground processing. Indeed, the application of Global positioning system (GPS) navigation complex +
inertial measurement unit (IMU) on board required for correct laser locator data interpretation makes it possible to record
3D coordinates of principal points and angles of optical axis orientation for each frame, which in turn provides the
99999999999999
Presented approach ensures automatic topographical mapping in scale 1:2000 - 1:5000 with total productivity of
200-300 km per day. The special software for automatic mapping and some examples of realization are described.
1. CLASSIC STEREOTOPOGRAPHIC METHOD:
ADVANTAGES AND PROBLEMS
Despite the permanent development of digital
photogrammetry methods the complete automation of
relief part of map production and orthorectification of
aerial photos can hardly be achieved within the classic
approach. Leaving alone the outdated and extremely
unproductive purely ground based methods of relief
depiction, a currently dominating based on aerial survey
data so called stereotopographic method of production
and upgrading topographic maps and plans, is
reviewed, that method allows both relief reproduction
and orthorectification in cameral conditions. An attempt
to automate the method faces the complications
conditionally divided in two categories:
1) Disadvantages of method. This category is made up
of a group of major photogrammetric problems, which
being put together mostly limit the method application in
topography in general. By recent times such were
rather considered as natural limits of application but not
problems. Reference to them as the problems is now
due to recent alternative solutions appearance.
Stereotopography method assumes the imperative on
ground geodetic plan/elevation support of aerial survey
results. Within such, a limited in volume geodetic survey
of the area relief is performed, that is conditioned by
high cost and low productivity of on-ground geodetic
works. The results on this stage, being spatial
coordinates of ground control points, are only used for
photo triangulation extension, that results in supporting
geodetic net spreading on all aerial photos of a project
thus providing each with 5-6 secondary control points.
This in turn allows absolute spatial orientation of
photos, and their mutual orientation in stereopairs, upon
which one may immediately begin retrieval of relief in
absolute coordinates and orthorectification. Such is, in
brief, a content of the classic stereotopography method
that is obviously well known to everyone aware of
photogrammetry in general.
The major conclusion that can be done of the above
brief description is - so produced relief model is only
with small degree (particularly in a volume of on-ground
geodetic support) provided with independent control
measurements. The rest of necessary information
appears within the stereotopography method itself, and
therefore by definition contains no mechanism of
correctness and accuracy control.
The above described technique may be, indeed, divided
into four general processes:
