Image Parallel Computing Theory And Technology Research Based On C++AMP

With the rapid development of computer technology, digital image processing technology has been widely used in the field of meteorology, biology, national defense science, technology, engineering, manufacturing, simulation and prediction, games and other area. Due to the large volume of data and real-time requirements of the continuous upgrading, so how to fast and effective image processing has become a hot topic int the field of parallel computing.In the performance of different parallel computing platform, Microsoft's C++AMP accelerated massively parallel computing(C++ accelerated massive Paralleism. C++AMP) fully using the GPU's powerful computing ability and realize the universal parallel computing has gradually become researchers develop parallel computing software preferred.This thesis focuses on the C++AMP technology in the application of parallel computing in image. Firstly, the calculation theory and parallel image processing technology has carried on the summary to the existing parallel to the image histogram calculation as an example, introduces the concept and programming of C++AMP core model, and analyzed the experimental results.Secondly, research on edge detection, image processing, color serialization and other typical simplified algorithm, using C++AMP to achieve the image of heterogeneous parallel computing platform. Finally, through the optimization calculation, the optimization of memory access model, performance analysis and others strategy, make full use of the calculation ability of accelerator, improves the efficiency of parallel image processing algorithms. On this basis, this thesis carried out a detailed experiment, choose different calculation models under the condition of the accelerator, the experimental results show that the C++AMP computing performance to get a huge boost in heterogeneous.