关键词: 矢量地图数据, WebGIS, 离散余弦变换, 平均误差, 空间数据压缩

Abstract:

With the constant enhancement of data acquisition capability, an increasing amount of spatial data is available to WebGIS. However, the efficiency of spatial data transmission cannot meet the current need for fast access to spatial data. In such a context, the compression of spatial data is important for reducing the space of data storage and improving the efficiency of data transmission and processing for WebGIS. Traditional methods of vector data compression are usually based on spatial relations, such as the distances or angles between different vertices, to determine which vertices are redundant. Whereas in the field of image compression, the method based on the techniques of transform coding (in frequency domain) is more frequently-used. This paper proposes a novel method of spatial vector data compression in frequency domain. This method is composed of several steps. To be specific, we transform the coordinate values of x and y to frequency domain coefficients respectively using Discrete Cosine Transform (DCT), and quantize these coefficients according to the relevant specific thresholds, which could restrict the average distortions (root-mean-square error, RMSE) of the reconstructed vector data. This quantizing method does not need quantization tables and is adaptive to large thresholds that often cause problems such as compressing a polygon or line feature into a point or a polygon feature into a line. In the final stage of the proposed vector data compression method, the Huffman coding is used in a similar way to the corresponding part of JPEG standard. The proposed compression method was implemented by C#.NET and was applied to compress vector data to test its performance from the aspects of compression ratio and geometric shape distortion. The results of the tests show that the proposed method is a feasible solution for vector data compression. It is flexible in terms of restricting distortion, and it can achieve large compression ratios as well as retain the main geographic shape´s characteristics which are derived from the original vector data, within the compressed vector data.

Key words: vector data map, WebGIS, DCT, average error, spatial data compression