Automatic selection of potential WWII bombed areas by using spatial data

Abstract

The reconstruction of military defense systems, (e.g. World War II defense lines) is generally based on military object identification and mapping. Since unexploded bombs can be still dangerous today, detecting bomb craters can be useful in creating hazard maps. The most significant problem is managing the large amount of relevant data. Therefore, there is a strong demand for automatically select the potential danger zones and also automate the entire processing workflow. Automatic methods have been developed to reduce the areas of interest. In addition, this method can be applied in other tasks as well, for example in settlement detection.

Automatic Identification of Bomb Craters and their Potential Location

“…The development had two stages; the previous process worked effectively for those areas where the ratio of the wooded areas was relatively low to total investigated areas. Testing the Emmerich am Rhein sample data (6% forest coverage) the original data was reduced to 30%, but when checking the Rheine dataset (20% forest coverage) with the same method, the remaining data was still 75% of the original data (Juhász and Neuberger 2018). The unified tile-size based calculation was the main disadvantage of the method, because a significant number of non-categorisable areas was left even after the morphological cleaning; thus the methodology was further improved.…”

“…It is able to select areas of interest from a large amount of integrated geodata and efficiently identify and count bomb craters with minimal userinteraction. Previous versions of the methods have been published (Neuberger et al 2017;Juhász and Neuberger 2018), and now the details of the further improvements are presented in this article. Finally, some examples are shown for the more general use of these methods.…”

Világháborús bombakráterek automatikus térképezése

“…Ha a részarány egy bizonyos határérték alatt van, akkor további, kisebb rácselemek következnek. Így egy adaptív eljárást kapunk, amely folyamatosan zárja ki a vizsgálatból a hasznos információt nem tartalmazó "üres" területegységeket, valamint a megtartott rácselemek esetében a minél nagyobb potenciális vizsgálati területarányra törekszik (Juhász -Neuberger 2018).…”