A NOVEL MINEFIELD DETECTION APPROACH BASED ON MORPHOLOGICAL DIVERSITY
Yuming Wang, Qian Song, Tian Jin, Xiaotao Huang, Hanhua Zhang · Electromagnetic waves · 2013
Battlefield surveillance is a common application of synthetic aperture radar (SAR), in which minefield detection is a challenging task.In this paper, a novel minefield detection approach is proposed via the morphological diversities between targets and background.Firstly, SAR image speckle is suppressed effectively by total variation, and targets edges are preserved well.Secondly, a nonlinear transform is introduced to map the special distributed targets, e.g., landmines, into spot targets.Lastly, the modification of morphological component analysis is adopted to improve the signalto-clutter ratio and separate the spot targets from image.The performance of the proposed approach is validated by using the data acquired over an airship mounted SAR system.