A SAR Target Image Simulation Method With DNN Embedded to Calculate Electromagnetic Reflection

Electromagnetic (EM) scattering calculation is a very important part of most synthetic aperture radar (SAR) target image simulation methods. It affects the intensity echo signal to great extent, thus affecting quality final image. EM reflection models are usually approximate formulas derived under certain assumptions. The errors between these and actual situation can cause significant differences simulated images real images. To solve this problem, we propose novel modified SAR framework, in which deep neural network (DNN) embedded calculate reflection, so that DNN directly learn fit from First, single separated cumulative multiple signals each pixel. Thus, be replaced with models. Next, model trained backpropagation algorithm Finally, fitted an post-processing applied simulate different imaging angles. In functions limited calculating coefficient adding sidelobe speckle noise. still original method based on ray tracing, ensures correctness generalization method. Experiments show proposed significantly improve When normalized [0, 1], minimum mean square error Sandia laboratory implementation cylinders reach 0.003. visualization results curve convolution kernel used for consistent laws theoretical model. addition, when complex targets, similarity also improved.