sht 
ange 
ng it 
ed in 
1 and 
four 
jicted 
inally 
r any 
mally 
been 
    
  
  
  
© False Positives (FP) 
  
  
Figure 10. Change map after refinement by morphologic 
opening. The 16 areas in red constitute building change. 
4.3 Quantitative Evaluation of Proposed Approach 
In order to evaluate the performance of our approach, we 
captured all building changes between the two epochs 
manually. We found 12 newly constructed but no demolished 
building. 
The results of a comparison of our results against the manual 
acquisition are contained in table 1. TP (True Positive) refers to 
the number of changed building (constructed or demolished) 
that are extracted correctly, FP (False Positive) indicates 
building changes incorrectly extracted by the approach; 
whereas FN (False Negative) is the number of missed changes. 
  
TP FP FN 
  
11 5 1 
  
  
  
  
  
Table 1. Quantitative evaluation of proposed method 
As can be seen in table 1 we were able to automatically extract 
11 of the 12 new constructions (a TP was declared if 75% of the 
area of the new building was covered by pixels indicating 
change). While we still need to reduce the number of FP, we 
note that there is only one FN in the results. As we had pointed 
out, our system is supposed to act as an alarm system, thus FP 
can be quickly deleted by a human operator. 
5. CONCLUSIONS AND FUTURE WORK 
In this paper we have described an approach for automatically 
detecting building change from stereoscopic high resolution 
satellite images of two epochs and have shown preliminary 
results. Due to the ground resolution of only Im some problems 
occur in properly detecting changes and delineation of change 
footprints is not possible. Besides refining the current 
workflow, e.g. by making use of the better spatial resolution of 
the GeoEye DSM and by improving the DEM extraction, in 
future work we plan to reduce the number of false alarms by 
introducing more shape information for the buildings and 
building changes such as requiring them to be rectangular. In 
addition we want to be able to handle more complex building 
ground plans and changes of building parts, e. g. new annexes 
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIX-B7, 2012 
XXII ISPRS Congress, 25 August — 01 September 2012, Melbourne, Australia 
of an existing building. Another area of future work will 
comprise a more sophisticated consideration of the existing GIS 
data. Accurate georeferencing of the GIS data to the DSMs is 
one topic, a more detailed comparison between extracted blobs 
and the building ground plan is another one. The tasks will 
probably require a finer image resolution, as was also found by 
Rottensteiner (2008) and Champion (2009). Further, we also 
consider the possibility to use radiometric image information. 
Once the false alarm rate is further reduced, we aim to use the 
approach in an alarm system for building updates. 
ACKNOWLEDGEMENTS 
The authors thank the German Academic Exchange Service 
(DAAD) for financial support of this research. We are also 
grateful to Dr. Alobeid for valuable discussions and for 
providing software and technical support. 
6. REFERENCES 
Alobeid A., Jacobsen K., Heipke C., 2010: Comparison of 
Matching Algorithms for DSM Generation in Urban Areas from 
IKONOS Imagery, PE&RS 76 (9), 1041-1050. 
Alobeid A., Jacobsen K., Heipke C., Al Rajhi M., 2011: 
Building Monitoring with Differential DSMs. In: IntArchPhRS 
38 (4/W19). 
Bouziani, M., Goita, K., and He, D., 2010: Automatic change 
detection of buildings in urban environment from very high 
spatial resolution images using existing geodatabase and prior 
knowledge: ISPRS Journal of Ph. and RS 65, 143-153. 
Chaabouni-Chouayakh, H., and Reinartz, P., 2011: Towards 
Automatic 3D Change Detection inside Urban Areas by 
Combining Height and Shape Information: PFG 4, 205-217. 
Champion, N., 2007: 2D Building Change Detection from High 
Resolution Aerial Images and Correlation Digital Surface 
Models. In: IntArchPhRS 36 (3/W494A). 
Champion N., 2009: Detection of unregistered buildings for 
updating 2D databases, EuroSDR Off. Publ. No. 56, 7-54. 
Heipke, C., Koch, A., and Lohmann, P., 2002: Analysis of 
SRTM DTM - Methodology and Practical Results: Journal of 
the Swedish Soc. for Ph. and RS, 69-80. 
Im, J., and Jensen, J.R., 2005: A change detection model based 
on neighborhood correlation image analysis and decision tree 
classification: Rem. Sens. of Environment 99, 326-340. 
Le Bris, A., and Chehata, N., 2011: Change detection in a 
topographic building database using submetric satellite images. 
In: IntArchPhRS 38 (3/W22). 
Malpica, J.A., and Alonso, M.C., 2010: Urban changes with 
satellite imagery and LiDAR data, In: IntArchPhRS 38 (8). 
Niemeyer, J.; Rottensteiner, F.; Kühn, F.; Soergel, U., 2010: 
Extraktion geologisch relevanter Strukturen auf Rügen in 
Laserscanner-Daten. In: DGPF Jahrestagung, 298-307. 
Rottensteiner F., 2008: Automated updating of building 
databases from digital surface models and multi-spectral 
images. In: IntArchPhRS 37 (B3A) 265-270. 
Tian, J., Chaabouni-Chouayakh, H., Reinartz, P., Krauss, T., 
and d’Angelo, P., 2010: Automatic 3D Change Detection Based 
On Optical Satellite Stereo Imagery. In: IntArchPhRS 38 (7B).