Distributed Network Image Processing System and Transmission Control Algorithm

With the increasing use of Internet technologies, image data is spreading more and on Internet. Whether it a social network or search engine, large amount generated. By studying distributed processing system transmission control algorithm, this paper proposes accurate gradient calculation method based SIFT algorithm. It concluded that performance proposed algorithm slightly better than original so implemented. On basis reducing dimension features effectively reduced. comparing influence retrieval in single-machine environment feature extraction rate, proved uses five nodes to construct achieves best results terms machine cost performance. The random Gaussian orthogonal matrix analyzed with good stability OMP has convergence reconstruction MH-BCS-SPL works best, PSNR decreases very smoothly process packet loss rate from 0.1 0.6. experimental show different bases behave differently under images. Overall, BCS-SPL series greatly improved effect compared traditional