| In recent years, with the rapid development of the large-scale scientific computing and graphic images,the large amount of data has far exceeded the capabilities of researcher to process data in many areas. In the field of volume rendering, the large-scale growth of volume data, eagerly awaited the processing power of computing devices can be further improved. Of course, the upgrading of the hardware device itself is the most effective way, but this can not be achieved without restriction in most conditions. The development of the distributed technology has played a good role in promoting volume rendering research. To take full advantage of the variety of computing resources distributed in the environments of web, it is not only possible, but is effective for us to introduce web services technology into the area of volume rendering and research the distributed algorithm of volume rendering.On the one hand, with the development of distributed computing, web services, in particular, as a new distributed technologies, provide a new way of thinking for the research for the volume rendering. As a new distributed computing technology, web Services propose service-oriented distributed computing model. They implement the integration and the interaction among the heterogeneous applications.The system developed with web services are also cross-platform, language independent, loosely coupled. Therefore, it is becoming possible for us to research the new volume rendering algorithms based on web services. In this way, we can use the distributed computing technology to accelerate the process of rendering.On the other hand, the process of volume rendering has distinct stages. Including re-sampling and image synthesis. Obviously, web services are very suitable for each stage. These services can be run on different machines in the environment of the distribution e, sharing resources in the network. Therefor,it is is not only possible, but is effective for us to introduce web services technology into the area of volume rendering and research the distributed algorithm of volume rendering.In view of this, the main contents of this thesis is as follows:(1) we analyse and study the principle of the ray casting algorithm, including the re-sampling and image synthesis two important stages.On this basis, we propose a modified ray casting algorithm-distributed parallel ray casting algorithm (Distributed Parallel Ray Casting Algorithm, DPRC).(2) we proof that it is feasible to intrduce Parallel technology into the Ray Casting Algorithm.And give the differences between Ray Casting Algorithm and DPRC algorithm. Then, we describe the principles of DPRC algorithm implementation, including the sub-division of tasks, dynamic task allocation strategy, task assignment at first time and final synthesis of the image, and we also clearly show the overall process of the DPRC algorithm. Finally, the DFRC algorithm is implemented, including the implemention of parallel strategy in DPRC algorithm and the implemention of the web services in DPRC algorithm.(3) In order to verify the DPRC algorithm,we construct a Web services-based volume rendering system (WSVRS). As a distributed computing system, WSVRS provide a basic development platform for the querying and importing of services in volume rendering. WSVRS use of open source distributed problem solving environment-Triana as the underlying support environment. Triana provides effective management of Web services and a strong workflow engine, more importantly, Triana workflow support for parallel execution,which provides a strong support for the runing of DPRC algorithm.(4) We verify the effectiveness of the DPRC algorithm by examples in the Campus Network. The effectivenesstest is tested mainly from two aspects:distributed execution and parallel execution. Quantitative analysis is done from the speedupand and efficiency of the parallel algorithm. Experiment shows that the DPRC algorithm accelerate the effect of volume rendering, achieving significant results. |
PDF Full Text Request