Studies On Comprehensive Visibility Analysis Model And Parallel Algorithm Based On DEM

Visibility analysis, an indispensible tool of spatial analysis in GIS applications, plays an important role in geoscience analysis. However, existing visibility analysis based on GIS method pays more attention on single-point, static points and straight line rather than multi-points, moving points and line of complex geometrical morphology. Therefore, it is necessary to establish a comprehensive visibility analysis model which considering the synthetic action of all the key elements of visibility issues. In addition, it would play a significant role in the theoretical innovation as well as improving spatial analysis abilities of GIS methods. Moreover, the existing GIS softwares based on serial computation are confronted with technical bottleneck in visibility analysis tools, especially when it deals with the huge volume of spatial data. The key technology of parallel computing of digital terrain analysis focuses on how to parallelize visibility analysis algorithms in distributed parallel computing environment. Based on the integration of key elements of visibility analysis, this paper conclude a theoretical and methodological system of comprehensive visibility analysis, construct a comprehensive visibility analysis model and adopt the parallel visibility algorithms to solve the high complexity visibility problems.The mainly contents and research achievements of this paper are as follows:(1) In terms of algorithm principles, algorithm procedures and extended theorical models of visibility analysis, this paper constructs a comprehensive visibility analysis model (CVIM for short) focusing on the specific application and basic theory. This model contains conceptual model and mathematical model with three core elements, i.e. analysis object, sight attributes and constrains. On the basis of aforementioned processes, the visible issues could be expressed in detail, and the visibility problems could be dissected and solved systematically.(2) Mount Lushan in Jiangxi Province is selected as a case study area for exploring the applicability of CVAM. Several pratical applications, i.e. multi-points inter-visibility analysis, fixed view-angle analysis, visibility analysis of moving points and curve sight-line, have been designed by using the proposed model. And then, landscape planning issues have been discussed, including landscape evaluation, optimum sites identifying and ropeway planning.(3) A granularity model for the parallel computing has been proposed in this paper, including several sub-models, i.e. data granularity model, task granularity model, performance granularity model and fault-tolerance granularity model. The granularity model may provide theoretical basis for the specific algorithm parallelization. Due to data-intensive is the main characteristic of parallel visibility analysis, the data granularity model would be treated as the key point of this paper. The experimental results show that the proposed method appears a good feasibility and stability performance, and the dynamic scheduling strategy achieves a preferable efficiency result, which would be expected to ease the low-efficiency problems of current distributed parallel computing.(4) This paper has designed parallel algorithms of visibility analysis for high complexity of analysis algorithm and big volume of DEM data, and investigaed versatile DEM management and scheduling mechanism for various parallel visibility algorithms. By using the data-parallel strategy, this paper proposed a versatile spatial data management and dynamic scheduling algorithm. Then the data management, buffer size and computing data granularity have been illustrated in detail. Finally, the parallel algorithms have been discussed, and the performance and extendibility of the algorithm has been proved.On the basis of the above research, this paper explores a new research method of extensional visibility analysis issues based on CVAM, which includes basis theoretical concept, analysis steps and model designing. CVAM may represent the visible issues synthetically as well as provide a comprehensive cognize for spatial tracing of complex trajectories. Simultaneously, a granularity model focusing on the computing mechanism has been proposed for the parallel computing. The results show that CVAM can effectively describe and analyze the characteristics of terrain visibility issues at multi levels. The parallel algorithm based on granularity model can achieve high efficiency and low communication, which preferably remit the bottleneck of communication and efficiency under distributed parallel computing environment.