Brandt, Steffen
Result of MATCH_T analysis
—9—- Fláche 1 / 12
—4— Fläche 1/24
—$— Fläche 2 / 12
—£&— Fláche 2 / 24
—$— Fláche 3 / 12
—$é-—- Fláche 3 / 24
—0— Fláche 4 / 12
—Ék—- Fláche 4 / 24
—9$— Fláche 5 / 12
—£&— Fláche 5 / 24
Square mean fo differences [m]
eeeséspppce
Different settings in MATCH-T
Figure 3. Statistical Analysis of the Elevation Accuracy in Dependence
on the Scan Resolution (12 or 24 um) and Topography (areal — area 5)
32 Automatic DTM Generation using OrthoMax as an Example
For the second project discussed here, the program package OrthoMAX (VISION International Inc., Earth City, USA)
was used to automatically reconstruct a DTM from high-resolution scanned aerial images of slag heaps on the basis of
homologous points. For a detailed description of the program, the individual steps to be followed and the algorithms
implemented in the software, please see ERDAS.
The main activities of DSK focus on the monitoring of slag heaps which are being filled. This involves an annual
photogrammetric evaluation of the slag heap site and of the current body of the heap combined with the creation of a
digital terrain model, a volume calculation and the determination of the remaining capacity based on the planned heap.
For further investigations a digital terrain model was correlated under OrthoMAX and its accuracy was analysed using a
reference model from the analytical photogrammetry.
Particular attention for this contribution was paid to the accuracy of the breaklines. As a rule, the geometry of the body
of a slag heap is especially suited to the detailed study of this, due to the prominent berms. In this example the body of
the slag heap was divided into 4 assessment zones due to the different characteristics of the surfaces (for example, the
northern zone is being filled, whilst the southern zone is already completely recultivated). The overall result of all 4
zones together is shown in the table below.
Standard deviation OrthoMax
Massenpunkte
Umring
Bruchkante
Square mean of
differences [m]
cO > N OO A O
cm 10 cm 30 cm 49 cm
Search radius
Figure 4. Standard Deviation Reference Measurement / OrthoMax (Massenpunkte = masspoints, Umring = border
polygon, Bruchkante = breakline)
On the one hand, it is evident that in the clearly visible area (characterised by the “border polygon” measurements)
accuracies in the order of 20 to 30 cm can be achieved. On the other hand, it was the breaklines which again showed a
significantly lower level of accuracy.
216 International Archives of Photogrammetry and Remote Sensing. Vol. XXXIII, Part B7. Amsterdam 2000.