Ultrafast Object Detection on High Resolution SAR Images

Synthetic Aperture Radar (SAR) images provide a good means of analyzing large areas despite fog, clouds, etc. High resolution SAR images enable the detection of a wide variety of relevant targets to detect potential threats and assess the situation. However, the vast amount of information and the peculiarities of SAR imagery require a huge amount of trained experts and/or time to comb through the data. In [1], we demonstrated an AI based approach to detect and classify objects in SAR images with a deep neural network and modern training techniques. In this publication, we investigate, how much the neural network can be reduced without sacrificing the classification performance. An ablation study is shown demonstrating the relevant parts of the neural network.