Yuya Arima

In this research, we conduct scattering analysis for a simplified manmade objects model and investigate the polarimetric scattering characteristics, as a fundamental study for accurately detecting oriented manmade area in polarimetric SAR (PolSAR) image. The Finite-Difference Time-Domain (FDTD) method is utilized here. The model used in the analysis is made of buildings parts (four dielectric rectangular parallelepipeds) and a ground part (a circular plate), to include multiple scattering contribution between buildings in this analysis. As the ground part, three different models showing three different surface conditions are considered: 1) dielectric ground model with flat surface, 2) conductor ground model with flat surface, and 3) dielectric ground model with rough surface. Characteristics of polarimetric correlation coefficient in HV and LR bases for the models are in detail investigated for squint angle variation.

The polarimetric observation data of Pi-SAR2, an airborne polarimetric synthetic aperture radar, are very valuable because of its high resolution, but it cannot be used effectively because the data are not well calibrated with respect to elevation. Therefore, we have calibrated the data according to the observation conditions. The Pi-SAR2 observation data are very large due to its high resolution and require sufficient computational resources to be calibrated. We utilized the AI Bridging Cloud Infrastructure (ABCI) to calculate them. This paper reports on the calibration and scattering power decomposition of the Pi-SAR2 observation data using the ABCI.

Pi-SAR2 is an airborne polarimetric synthetic aperture radar operated by the National Institute of Information and Communications Technology. The polarimetric observation data of Pi-SAR2 are very valuable because of its high resolution, but it cannot be used effectively because the data are not well calibrated with respect to elevation. Therefore, we have calibrated the data according to the observation conditions. The Pi-SAR2 observation data are very large due to its high resolution and require sufficient computational resources to be calibrated. We utilized the AI Bridging Cloud Infrastructure (ABCI), constructed and operated by the National Institute of Advanced Industrial Science and Technology, to calculate them. This paper reports on the calibration, scattering power decomposition, and orthorectification of the Pi-SAR2 observation data using the ABCI.

An active millimeter-wave imaging system using complex-valued self-organizing map (CSOM) has potential applications in moving-target imaging by acquiring and classifying complex textures. However, interference in the coherent observation causes harmful effects such as amplitude fluctuation and phase distortion, resulting in the deterioration of the visualization quality achieved by the CSOM. In this paper, we propose the introduction of feature refinement using a complex-valued auto-encoder into the feature extraction process of our imaging system to suppress these effects. We show that the complex-valued auto-encoder is extremely useful for extracting features appropriately even under such adverse influences and improves the clustering performance. We also experimentally investigate the influence of the number of hidden-layer neurons on the auto-encoder performance to discuss the robustness of our system.

Millimeter wave exhibits relatively straight propagation as well as high penetration into dielectric materials, such as plastics, cloth, and paper. In security imaging, we use these features to discover weapons concealed under clothes. Self-organizing map (SOM), a type of neural networks, can map high-dimensional data on any dimension with unsupervised learning. It is utilized for clustering and visualization of high-dimensional data. Previously, we proposed a millimeter-wave imaging system for moving targets consisting of a 1-D array antenna, a parallel front end, and a complex-valued SOM to deal with complex texture. Experiments demonstrated its high performance in the visualization. In this paper, we investigate the dependence of the visualization performance on its configuration parameters as well as processing parameters. We reveal the effect of the modulation-frequency number and the window size. We also discuss the effective depth range for visualization and a tradeoff relationship between the measurement time and the visualization quality.