Search

Nowadays, benefiting from its strong capability of nondestructive detection, the ground-penetrating radar (GPR) has been applied to detect and reconstruct underground targets and has drawn lots of attention both in military and civilian fields. However, in the processing of GPR imaging, due to the dispersion errors caused by random distribution of various particles in soil, conventional imaging methods have disadvantages of low signal-noise ratio (SNR), low resolution, and unbalanced amplitude. In this article, to achieve high resolution and high veracity on underground targets’ 3-D reconstruction, we improved the conventional reverse time migration (RTM) algorithm in perspective of medium constitutive relationship. Besides, we extended the improved RTM method in 3-D environments and reconstructed the 3-D structure of several underground targets. To make the RTM algorithm suitable for stepped frequency continuous wave (SFCW) GPR system, we generated three excitation signal models and analyzed the effect of different excitation signals on imaging performance. Finally, through the quantitative analysis of simulation and on-vehicle experimental results, we found that the 3-D images generated by improved RTM method had higher resolution, smaller measurement error, and higher veracity than those of conventional RTM method. [ABSTRACT FROM AUTHOR]

Reconstructing a 3D shape from a single-view image using deep learning has become increasingly popular recently. Most existing methods only focus on reconstructing the 3D shape geometry based on image constraints. The lack of explicit modeling of structure relations among shape parts yields low-quality reconstruction results for structure-rich man-made shapes. In addition, conventional 2D-3D joint embedding architecture for image-based 3D shape reconstruction often omits the specific view information from the given image, which may lead to degraded geometry and structure reconstruction. We address these problems by introducing VGSNet, an encoder-decoder architecture for view-aware joint geometry and structure learning. The key idea is to jointly learn a multimodal feature representation of 2D image, 3D shape geometry and structure so that both geometry and structure details can be reconstructed from a single-view image. To this end, we explicitly represent 3D shape structures as part relations and employ image supervision to guide the geometry and structure reconstruction. Trained with pairs of view-aligned images and 3D shapes, the VGSNet implicitly encodes the view-aware shape information in the latent feature space. Qualitative and quantitative comparisons with the state-of-the-art baseline methods as well as ablation studies demonstrate the effectiveness of the VGSNet for structure-aware single-view 3D shape reconstruction. [ABSTRACT FROM AUTHOR]

A rectangular tape helix slow-wave structure with infinitesimal thickness and finite width in free space is investigated. With the expansion of surface currents in the helix and the applications of the modified Marcatili’s method, as well as average power flow matching method at the boundaries, the dispersion properties and the interaction impedance for transverse antisymmetric modes in a rectangular tape helix immersed in free space are obtained. It is shown that, compared with the results of the simplified sheath model by previous researchers, higher accuracy has been obtained between the calculation results of the present theory and the data obtained from HFSS, and the validity of the present theory is further demonstrated by comparison with experiments. The improved characteristic equations hold scientific and practical significance in the design and performance evaluation of such plane slow-wave structure in the application of compact traveling-wave tubes. The distribution characteristics on the cross section of the longitudinal electric field fundamental component are also discussed based on this theory. [ABSTRACT FROM AUTHOR]