In: Wagner W., Székely, B. (eds.): ISPRS TC VII Symposium - 100 Years ISPRS, Vienna, Austria, July 5-7, 2010, IAPRS, Vol. XXXVIII, Part 7B 
Now, if AB value is 13, this means that the coherence was low 
in period 1 and high in period 2. If BC value is 3 at that same 
place, meaning that the coherence remained high from period 2 
to period 3, this can mean that there were some works in that 
place (for instance, human activity such as creating built-up 
areas) which started in the first period (AB=13) and have been 
finished in the following periods (BC=3, referring to a built-up 
area, i.e. hard surface). The pixels where AB=13 and BC=33 are 
shown in white in Fig. 6 and zoomed in Fig. 7. 
5. INCLUSION OF KNOWLEDGE SOURCES 
In (Milisavljevic, Closson and Bloch, 2010), we mention the 
possibility of including knowledge sources. We develop this 
idea here. Namely, once CCD results are obtained, various 
knowledge sources can be taken into account in order to 
improve the final interpretation. These knowledge sources can 
be related to the sensors, such as their operational principles, or 
to the situation at hand, referring to the context - terrain type, 
land-use, historical background etc. We illustrate here the 
inclusion of three knowledge sources: 
1. area of interest (airport), extracted manually from a very high 
resolution image and shown in Fig. 8 - it can be included in the 
obtained results in a straightforward way, by taking into account 
only pixels that are within the white region; 
2. pedologically extracted robustness (i.e., soil hardness), 
presented in Fig. 9; while dark grey pixels represent areas with 
low robustness, high (white) values correspond to hard surface; 
it is difficult to perturb such a soil and detect human activities, 
so coherence is high in such an area and the reliability to detect 
human activities is low; therefore, regions with the highest 
robustness are superimposed to the classification results as areas 
with surely high coherence (label 3); 
3. slope (extracted from the digital elevation model created from 
contours of topographical maps), presented in Fig. 10, where 
the lighter the pixel, the higher the slope, so the lower the 
possibility to have an airfield; again, regions with the highest 
slope can be superimposed to the classification results as being 
of low interest while looking for human activities in an airport 
(so, labelled as 3). 
Figure 9. Pedologically extracted robustness. Black - no data; 
dark grey - low robustness; light grey - medium robustness; 
white - highest robustness 
Figure 10. Slope (extracted from the digital elevation model) 
Figure 11. Result of the inclusion of three knowledge sources in 
results shown in Figure 5 
As an illustration, Figure 11 contains the result of the inclusion 
of the three knowledge sources mentioned above to the result 
shown in Figure 5. 
Based on the way the knowledge sources are included, it is 
certain that the method itself would not be affected if they were 
not present. In addition, as these sources, in this particular case, 
affect only areas that are outside of the region within the blue