The Offshore Munitions Problem
Safe access to large areas of the sea floor is a prerequisite for offshore economic development. However, this ability is compromised by the presence of submerged ammunition around the globe. Over the past 140 years, ammunition entered the sea in a variety of entry modes such as naval battles, aerial bombing and mine laying. Additionally, large quantities were dumped after the Second World War. What appears unreasonable today, was considered a safe and permanent means of disposing of weapons for many decades. Even today, in times of peace, naval target practice adds to the amount of munitions in the sea.
To this day, more than 1,6 million tons of both chemical and conventional (i.e. explosive) munitions remain in German territorial waters alone. The OSPAR Quality Status Report listed 148 dumpsites in the North-East Atlantic Ocean and North Sea. The reported number of encounters of munitions in these areas amounts to over 900 in some years. As such, these dumped munitions constitute a serious safety risk for users of our coasts and seas. The legacy is not only an obstacle to the offshore construction industry, it also affects sectors like coastal tourism and fishing. On top of this, it will impede deep sea mining endeavours in the future. Finally, it may pose a threat to marine life due to gradual leakage of carcinogenic TNT and its metabolites into the sea water.
The Challenge of Munitions Detection
Conventional approaches for offshore munitions detection are time consuming and therefore costly. There is no stringent industry wide standard for data acquisition and data handling, and this has resulted in high diversity and heterogeneity in process chains.
The trend of an ever faster growing amount of data during munitions detection will continue and is already part of the general big data challenge. Single data sets alone are already 'big' and analysing several different layers is very time consuming (magnetic, multibeam, multibeam backscatter, sub-bottom profiler, side-scan sonar, and various properties of these datasets). Because of this, data are often analysed separately and the final 'classification' of an object often still depends on subjective (i.e. human) interpretation which highly affects repeatability.
It is the aim of BASTA to increase both accuracy and cost-efficiency of munitions detection, thereby tackling the processes of data acquisition, data processing and data interpretation alike.