Research Of Parallel Image Fusion Processing Technology Based On Multi-core Processors

The subject of this thesis is to combined image fusion technology with parallel computing, which are high-performance image fusion processing technology. In the subject, sensors from the same or different for the same target shooting multiple images are blended into one to meet the specific needs of the image ,which can effectively combine the advantages of the different images to enhance the image information analysis and extraction capabilities.Thus more accurate, more comprehensive, more applications value target image would be got. The using of image fusion technology can effectively improve the image resolution, increase the amount of information and reduce the image noise. Meanwhile image fusion has greatly value in many areas, but the amount of information and data is growing even bigger which is needed to be processed by image fusion. In this context, the combination of image fusion and parallel processing technology is becoming extremely urgent.In this paper, cluster parallel processing technology is drawn upon the in the multi-core processors, a parallel image fusion platform is built under a single virtual node. Because of image fusion with the compute-intensive, require high precision characteristics,it also require real-time or quasi real-time image fusion in practical applications. At present a general study of parallel image fusion technology is mainly based on cluster computing systems because of the cluster system needs to work simultaneously on multiple computers, so it will be limited in a number of practical applications. The emergence of multi-core processors make parallel computing can be implemented on stand-alone, thus the application of image fusion technology becomes much broader.In this paper, the parallel image fusion system developed is divided into image pre-processing, image fusion, fusion results showed that three parts. Development platform is a quad-core processors machine, respectively; the platform will be regarded as four 4-processor nodes, each node in accordance with a MPI communication standard for data communication.VC + + 6.0 is used as a parallel development environment and Mallat algorithm is used in the paper, in the end of the thesis, simulation experiments for the image fusion system is processed. It is introduced by examples in parallel processing technology to achieve image fusion process. a more satisfactory speedup and efficiency is obtained through the experimental data. Thus it is improved the integration of multi-core processors, parallel image processing technology which under the validity and usefulness.