大量空间数据可视化的算法

在大量应用中，空间数据的采集和使用都是通过其空间位置进行的。在大小有限的屏幕上显示大量的空间数据常会因为邻结点的高度重叠而导致较差的视觉效果。该文提出一种可视化大量空间数据的新方法。其主要思想就是用未被使用的像素点巧妙地代替重叠的数据点。介绍了两种解决方法：(I)将重叠的数据点放置在最近的未被占用的像素点上；(2)将数据点沿着屏幕填充曲线移动。接着，发展了一个更为成熟的方法叫栅格划分算法。最后，提出了一种有待进一步研究的新方法;基于最近邻接方法的聚类方法，并指出了今后研究的方向。

Algorithms to Visualizing Large Spatial Data Set

When a large amount of spatial data on a limited screen are shown, there is often a poor visualization because of the high degree of overlapping of data points. This paper introduces an algorithm to visualizing large spatial data sets. Its main idea is to use unoccupied pixels instead of overlapping data points. It presents two ways: (1) put the overlapping data points on thee unoccupied pixels which is nearest to the original data points; (2) move the overlapping data points along some screen curves. Then a more sophisticated algorithm which is called Grid is presented. At last, a new visual clustering method for spatial data and the aspects which are under developing are presented .