Dynamic Multi-scale Transmission Of Vector Data Over The Internet

With the development of spatial information services and the growing popularity of spatial data over the internet, the data transmission time which is usually too long due to the massive spatial data and complex network environment becomes a technical bottleneck and restricts the development of spatial information services. Multi-scale transmission technology of spatial data is one effective way to improve network transmission efficiency and reduce data waiting time of the client. At present, the research on Multi-scale transmission of spatial data based on GRID/TIN structure has received rich results, while the topic on Multi-scale transmission of spatial data based on vector structure which is first proposed by the Bertolotto has gained less focuses and the technique for transmitting vector data has yet to be improved.Unfortunately, main technical schemes for multi-scale transmitting vector map data still exist deficienties. Scheme which generates multi-scale vector data based on on line simplification has flexibility, but the time performance is not ideal. Scheme which generates multi-scale vector data by the pre stored a few of fixed scale data has a better time performance, but the data redundancy and fexibility is insufficient. For the deficiencies of current technology, a dynamic multi-scale transmission model of vector data is proposed and is verified its feasibility and superiority through experiments in this paper. Vector data is organized into a multi-scale map structure and a hierarchical Hilbert R Tree index is established on server side in advance. Upon receiving the data request from the client, the server can quickly generate the multi-scale map data by hierarchical Hilbert R Tree index and multi-scale map structure, and quickly maintain the topological consistency of vector map by using an optimized monotone chains intersection algorithm.The main contents of this paper include:(1) multi-scale organization method of vector data. Spatial data operation process includes data retrieval, extraction and simplification, the multi-scale organization method is one effective way to quickly generate the multi-scale vector map data on server side. To shorten the data extraction time on server side, we proposed a hierarchical Hilbert R Trees index structure. The structure stores features on different levels of R Trees according to their hierarchical levels and supports data query based on their location and scale information. By adjusting the records number dynamically, the data capacity of leaf node is close to the disk block size that can optimize the data I/O access time. The method of storing a number of fixed scales of data on server side can meet the need of quick data retrieval, but has limitations as inflexibility and data redundancy. A multi-scale organization structure of vector map for multi-scale transmission is presented in the paper. The structure which is used to organize vector map on server side can quickly simplify vector map data with the minimum data redundancy.(2) The fast algorithm for maintaining topological consistency of vector map. Topological inconsistency exists in the process of multi-scale map generalization. The existing methods for maintaining topological consistency have a high time complexity and not suit for maintaining topological cosistency of massive map data. A fast algorithm for maintaining topological consistency is proposed in this paper, the procedure of the algorithm is to (a)avoid the occurrence of the first type of topological exception by preserving constraint points; (b)detect and eliminate the other types of topological exception by using an optimized monotone chains intersection method. Compared with the current existing algorithms, the algorithm proposed in this paper can guarantee the planar and non-planar topological consistency and has a lower time complexity.(3)The similarity measurement model for multi-scale spatial data. Concerning the dissimilarity between the original spatial data and the progressively transmitted spatial data, this paper proposes a model for measuring similarity of multi-scale spatial data. The model takes into account the differences of the topological and geometric characteristics and includes some measuring factors such as weighted similarity, similarity standard deviation and similarity distribution curve. The model uses weighted similarity to assess the overall similarity degree of feature datasets and uses similarity standard deviation, similarity distribution curve to assess the similarity differences between features. The similarity measurement model is verfied its validity in the experiment.The experiment system is developed based on the dynamic multi-scale transmission model of vector data, and the experimental results show that:(1)the model can realize’the multi-scale transmission of vector data in the premise of meet ing the requirements of real time response, dynamic generation and topological consistency, ﹖he multi-scale transmission mode can effectively reduce the data waiting time of the client and improve user experience.③The similarity measurement model can assess the overall similarity degree between different feature datasets and the differences among individuals of same feature dataset.