Bilinear pooling with poisoning detection module for automatic side scan sonar data analysis

Side-scan sonar (SSS) images are difficult for automatic analysis due to the acoustic measurement parameters as well number of different objects that can be distant. In addition, there is a risk seabed application may attacked. For this purpose, we propose solution based on convolutional neural networks with bilinear pooling in order achieve higher values classification accuracy. Bilinear merge data from two and return results. The first network’s branch receives original image second one after applying superpixel method. This approach allows focus types features. introduced mechanism poisoning detection analyze results network. evaluation process, used real SSS obtained between water channels Szczecin city north-western Poland. importance scientific research indicates accuracy safety measurements performed.