A Segmentation Algorithm of Image Semantic Sequence Data Based on Graph Convolution Network

Image semantic data have multilevel feature information. In the actual segmentation, the existing segmentation algorithms have some limitations, resulting in the fact that the final segmentation accuracy is too small. To solve this problem, a segmentation algorithm of image semantic sequence data based on graph convolution network is constructed. The graph convolution network is used to construct the image search process. The semantic sequence data are extracted. After the qualified data points are accumulated, the gradient amplitude forms complete rotation field and no scatter field in the diffusion process, which enhances the application scope of the algorithm, controls the accuracy of the segmentation algorithm, and completes the construction of the data segmentation algorithm. After the experimental dataset is prepared and the semantic segmentation direction is defined, we compare our method with four methods. The results show that the segmentation algorithm designed in this paper has the highest accuracy.

Research on Calculation Method for Acoustic Scattering of Helicopter Noise

A new calculation method of helicopter rotor/fuselage acoustic scattering is developed. Firstly, a CFD analysis model is developed to simulate flow field of the rotor, which is based on the motional embedded grid system and RANS equations, and provides aerodynamic data for rotor noise calculation. Then, FW-H equations are employed to calculate the aeroacoustic characteristics of isolated rotor, and G 1A formulas are applied to calculate the rotor acoustic gradient to provide boundary condition for acoustic scattering. Based on these, the time-domain equivalent source method is applied to calculate acoustic scatter field, and the total acoustic field that considered the fuselage scatter is superposed by isolated rotor acoustics and the scatter one. Finally, the numerical simulations of helicopter main-rotor/fuselage and tail-rotor/fuselage scatter effect are conducted by using the developed models. The results indicate that the helicopter fuselage has important scatter effect on the high frequency acoustics of main rotor and tail rotor, and the acoustic scatter effect become more obvious with the smaller space between the main rotor (tail rotor) and fuselage.

Multiscattering inversion for low-model wavenumbers

A successful full-waveform inversion implementation updates the low-wavenumber model components first for a proper description of the wavefield propagation and slowly adds the high wavenumber potentially scattering parts of the model. The low-wavenumber components can be extracted from the transmission parts of the recorded wavefield emanating directly from the source or the transmission parts from the single- or double-scattered wavefield computed from a predicted scatter field acting as secondary sources. We use a combined inversion of data modeled from the source and those corresponding to single and double scattering to update the velocity model and the component of the velocity (perturbation) responsible for the single and double scattering. The combined inversion helps us access most of the potential model wavenumber information that may be embedded in the data. A scattering-angle filter is used to divide the gradient of the combined inversion, so initially the high-wavenumber (low-scattering-angle) components of the gradient are directed to the perturbation model and the low-wavenumber (high-scattering-angle) components are directed to the velocity model. As our background velocity matures, the scattering-angle divide is slowly lowered to allow for more of the higher wavenumbers to contribute the velocity model. Synthetic examples including the Marmousi model are used to demonstrate the additional illumination and improved velocity inversion obtained when including multiscattered energy.

Ultrasonic Imaging of a Crack Using Modified Time Reversal

A majority of the research in Structural Health Monitoring focuses on detection of damage. This paper presents a method of imaging crack damage in an isotropic material using the Time Reversal imaging algorithm. Inputs for the algorithm are obtained via computational simulation of the propagation field of a crack in a medium under tone-burst excitation. The approach is similar to existing techniques such as Diffraction Tomography which makes use of the multi-static data matrix constructed using scatter field measurements from the computational simulation. Results indicate excellent reconstruction quality and accurate estimation of damage size.

Efficient Plane Wave Excitation for 3-D Laguerre-Based FDTD Method

In the three-dimensional (3-D) Laguerre-based finite-difference time-domain method, each electric field variable has the relationship with the adjacent twelve electric fields. This results in the tedious modification of field components adjacent to the total-field/scatter-field boundary in analyzing scattering problems. In addition, the plane wave excitation requires much time in evaluating the expansion coefficient of incident field which involves integral of the weighted Laguerre polynomials with respect to time. In this letter, the plane wave is introduced by defining a set of equivalent currents on a closed Huygen's surface and a computationally efficient one-dimensional auxiliary propagator is presented to speed up the plane wave excitation. Numerical results indicated that the proposed method is valid.

A Hybrid Technique Based on the Combination of Multilevel Fast Multipole Algorithm and the Geometrical Theory of Diffraction

This paper proposes a hybrid technique for treating electromagnetic problems of scattering and radiation in which the source structure is described as an array of antennas. This strategy is based on the combination of the rigorous method multilevel fast multipole algorithm (MLFMA) and the high frequency technique geometrical theory of diffraction (GTD). Thanks to the use of MLFMA, the source can be discretized into several cubic regions considering each of them as a source point in order to reduce the number of times required to compute the ray tracing when GTD is applied to obtain the scatter field. In this analysis, objects with complex shapes are described by using nonuniform rational B-splines (NURBS) which is a very common way to model geometrical bodies. Numerical results that demonstrate the accuracy and efficiency in terms of CPU time are shown.

Velocity Imprecision in Finite-Beamwidth Shipboard Doppler Sonar: A First-Generation Correction Algorithm

The finite angular width of Doppler sonar beams introduces errors into the measurement of ocean velocity from moving ships. These errors are exacerbated as ship speed increases and as the acoustic scatter field becomes more inhomogeneous in space. Zooplanktonic scatters often reside in distinct quasi-isopycnal scattering layers. When measured from a moving ship, such layers appear to have a distinct velocity signature, even if the ocean is quiescent. Here, this effect is explored in a simple analytic model and a least squares algorithm is presented for remediating its signature. The algorithm corrects the measured velocity at any given depth, with the correction being a weighted sum of the depth gradient of log acoustic intensity at surrounding depths. The correction weights are a property of the specific sonar beam pattern. Once determined, they can be considered constant. The algorithm, when applied to a 50-kHz sonar on the Research Vessel (R/V) Roger Revelle, is found to remove roughly 90% of the error in vertical wavenumber spectra of shear. More sophisticated algorithms can be developed as experience with the present approach is gained.