The Doppler spectrum of the electromagnetic (EM) scattering field from the two-dimensional dynamic sea surface is calculated based on the composite scattering model. The two-dimensional dynamic sea surfaces are generally simulated as a superposition of large-scale gravity waves and small-scale capillary ripples. On this basis, the Doppler spectrum of the EM scattering field from the two-dimensional dynamic sea surface can be calculated based on the composite scattering model, which takes both the quasi-specular scattering and Bragg scattering mechanism into account. However, due to the high resolution and real-time dynamic complexity of the dynamic sea surfaces, the calculation of the Doppler spectrum will be computationally expensive and very time-consuming. In this paper, a GPU-based algorithm of Doppler spectrum was proposed by utilizing the Tesla K80 GPUs with diverse CUDA optimization techniques. The GPU-based Doppler spectrum implementation includes five optimization strategies: first, the temporary arrays are utilized to reduce the repeat float-points operations in the loop; then the device memory was effectively exploited to reduce the data transfer time between the CPU and GPU; the fast math compiler option was also utilized to further improve the computation performance of the Doppler spectrum calculation; finally the data transfer time between the device and host memories can be effectively hide by using the asynchronous data transfer (ADT). Compared to the CPU serial program executed on Intel(R) Core(TM) i5-3450 CPU, the GPU-based Doppler spectrum implementation can achieve a significant speedup of1200× .
Study of characteristics of sea clutter is very important for signal processing of radar, detection of targets on sea surface and remote sensing. The sea state is complex at Low grazing angle (LGA), and it is difficult with its large irradiation area and a great deal simulation facets. A practical and efficient model to obtain radar clutter of dynamic sea in different sea condition is proposed, basing on the physical mechanism of interaction between electromagnetic wave and sea wave. The classical analysis method for sea clutter is basing on amplitude and spectrum distribution, taking the clutter as random processing model, which is equivocal in its physical mechanism. To achieve electromagnetic field from sea surface, a modified phase from facets is considered, and the backscattering coefficient is calculated by Wu’s improved two-scale model, which can solve the statistical sea backscattering problem less than 5 degree, considering the effects of the surface slopes joint probability density, the shadowing function, the skewness of sea waves and the curvature of the surface on the backscattering from the ocean surface. We make the assumption that the scattering contribution of each facet is independent, the total field is the superposition of each facet in the receiving direction. Such data characters are very suitable to compute on GPU threads. So we can make the best of GPU resource. We have achieved a speedup of 155-fold for S band and 162-fold for Ku/Χ band on the Tesla K80 GPU as compared with Intel® Core™ CPU. In this paper, we mainly study the high resolution data, and the time resolution is millisecond, so we may have 10,00 time points, and we analyze amplitude probability density distribution of radar clutter.
The scattering characteristic of complex target from terrestrial and celestial background radiation has been widely used in such engineering fields as remote sensing, feature extraction, tracking and recognition of target thus having been an attractive field for many scientists for decades. In our method, the model of target is constructed using 3DMAX and the surface is divided into triangle facets firstly. Bidirectional Reflectance Distribution Function (BRDF) is introduced and MODTRAN is applied to calculate background radiation for a given time at a given place. Finally the scattering of each facet is added up to get the scattering of the target. As the background radiance comes in all directions and in a wide spectrum and the complex target always consists of thousands of facets, in general it takes hours to complete the calculation. Consequently this limits its use in the real time applications. Recent years have seen the continual development of multi-core CPU. As a result parallel programming on multi-cores has been more and more popular. In this paper, the openMP, Intel CILK ++, Intel Threading Building Blocks (TBB) are used separately to leverage the processing power of multi-cores processors. Our experiments are conducted on a DELL desktop based on an Intel I7- 2600K CPU running at 3.40 GHz with 8 cores and 16.0 GB RAM. The Intel Composer 2013 is employed to build the program. Also in OpenMP implementation, gcc is used. The results demonstrate that highest speedups for three parallel models are 5.06X, 5.02X, 5.15X respectively.
The problem of electromagnetic waves scattering at rough boundaries is of practical interest and has been addressed many times in different papers. Theories for investigation of rough surface scattering primarily two kinds of methods: numerical method and approximate method. As the classic analytical methods cannot calculate the electromagnetic scattering characteristics at Low Grazing Angle (LGA) accurately, in this paper, a novel method is presented by utilizing the Radar Cross Section (RCS) of the low grazing two-dimensional sea surface based on the triangles-based Physical Optics (PO) method. Firstly, the Monte Carlo method is applied to simulate the two-dimension rough sea surface in different wind speeds based on the PM sea spectrum. Then, the sea surface is generally meshed by 1/8~1/10 length of the incident wave. Secondly, the complex permittivity of the sea surface is calculated by two-Debye method and compared with the experiment. The physical optical is used to calculate the backscattering coefficient of the random rough sea surface. Considering the problem of low grazing, it is especially sampled more densely between the scattering angles 70°~90°. Then the self-shadow and inter-shadow of the sea surface at low grazing angle is taken into account, the Z-BUFFER method is used to judgment of the shadow effect. The numerical result is compared with the FEKO and good agreement is obtained. As the frequency increasing, the sea surface will have more triangles to be calculated, it will take more time. Finally, we propose a novel graphic processing unit (GPU)-accelerated decoding system. The result is 68.96 faster than its CPU counterpart.