Bomb funnel! ...or not?

The interpretation of explosive bomb craters is an essential part of aerial image analysis for the preliminary detection of explosive ordnance. How well these objects can be recognised depends primarily on the quality and scale of the available aerial imagery. If both factors are favourable, it is generally possible to identify explosive bomb craters on the basis of their characteristics (ejected material, hollow shape). This is particularly the case if the bombing took place shortly before the aerial photograph was taken and the funnels were not filled in again between the time they were dropped and the aerial photograph was taken. Older bomb craters that were partially or completely filled in, on the other hand, can often only be recognised as round, flat objects with blurred edges in wartime aerial photographs (see Fig. 1 for an example). Identification as bomb craters is often no longer possible beyond doubt, especially as objects shaped in this way can also be attributed to non-military circumstances.

In areas used for forestry, it is the hole cutting (circular clearing to rejuvenate the forest stand) that creates aerial structures that - especially if they are small and irregular - can easily be mistaken for "aged" bomb craters (see Fig. 3 for an example).

Other - albeit less frequently occurring - objects that have the potential to be confused with older/overprinted bomb craters are, for example, circular field irrigation systems or old barrows. In both cases, flat, circular ground structures can be seen in the aerial image, which can bear a clear resemblance to infilled bomb craters (see Fig. 4 for an example). 

Another important interpretation feature when analysing aerial images to detect bomb craters is the geometric arrangement of the individual objects. For example, certain distribution patterns can provide a strong indication that image content that has typical characteristics of bomb craters in terms of shape and size can be attributed to an origin other than bombing.

In order to avoid confusion with image features of civilian origin when detecting explosive bomb craters, it is necessary to obtain aerial images of several time periods in order to be able to carry out a multi-temporal comparison of the utilisations at the location in question. This is the only way to ensure the most reliable object identification possible; the risk of misinterpretation is minimised. A monotemporal aerial image evaluation with consideration of a single aerial image time slice considerably increases the probability of misinterpretations and is not to be considered technically meaningful. A prerequisite in this context is the availability of the corresponding image material in the relevant archives.