Multi-user Collaborative Scheduling Scheme Research In Mass Satellite Image Visualization

The problem of massive satellite image visualization problem is a focus of attention ofgeography information systems. With the development of computer hardware, the problemsof massive satellite images visualization encountering in the rendering frame rate, largedata scheduling, and external memory organization have been basically solved. In theaspect of mass satellite image visualization for multi-user, Google earth put forward amulti-user solution which is depended on the powerful cluster rendering technology andcollaborative network transmission technology. However, in the limited resources of thestand-alone system, this problem becomes a new challenge. Because the data schedulingalgorithm directly influences the final scene rendering quality and user roaming experience.This thesis reconstructs the current scheduling algorithms of satellite image data forsingle-user, and analysis them comparatively. A collaborative scheduling scheme of satelliteimage data for multi-user in the stand-alone mode is proposed, which can improve theutilization rate of system resource and balance the user’s experience while ensuring thestability of system throughput. Finally, in the prototype system, the validity of this methodis verified through analyzing its efficiency and interactivity from a user study.The main works of this dissertation are as follows:1) Firstly, a study of the satellite image data scheduling scheme for single-user iscarried out. On the basis of the improvement of the traditional single-userscheduling scheme, two data scheduling algorithm faced to different applicationenvironment are designed and analyzed comparatively, which provide a basic ideato the problem of the multi-user massive satellite image data scheduling.2) Secondly, we research multi-user satellite image collaborative data schedulingalgorithm. For the image data request characteristics under multi-user background,we reconstruct the structure of buffer pool as well as overall rendering frame. Onthis basis, we design a collaborative scheduling scheme which is based on weighteddynamic priority principle. This scheme not only can meet user’s data requirementwho have higher priority, but also can ensure other users’ roaming experience.Finally, we design matching experiments, and analyze the experimental results.3) Lastly, a satellite images visualization system which can support multi-userobserving satellite image at the same time is constructed on the stand-alone situation. We design a different application scenario which use multi-touch screenas our interactive equipment, it can support multi-users’ collaborative operating instand-alone situation. Our system can also guarantee a higher value of the averageuser satisfaction under the fixed system throughput.