基于矢量数据的土地利用类型分维数计算程序设计及其应用

Programming on Calculating Fractal Dimension of Land Use Types for 2D Vector Data

当前用来计算分形维数的原始数据通常是栅格图像,但是现有的土地利用数据大多以矢量形式存放,若将其转为栅格形式后计算分维数,则将面临基于栅格图像所得到的分形维数的计算精度与像元尺寸以及图像自身大小有较大关系,并且在对栅格图像进行不规则形状提取时容易造成数据丢失等问题。鉴于此,本文利用Visual C#完成了基于Windows平台的矢量数据分形维数计算程序设计,并利用该程序对20世纪90年代三个时期的秦岭中段和鄂东南地区土地利用类型的分维数进行了求算。结果表明该程序是一种方便、实用而且精确的分维数计算方法。此外,从三期各类土地利用类型的稳定性指数来看,在秦岭中段地区表现为:居民点及工矿用地>未利用土地>水田>旱地>草地>林地>水域,而鄂东南地区则表现为:水田>未利用土地>居民点及工矿用地>旱地>林地>草地>水域,这说明在上述两个地区,林地、水域、草地的稳定性都相对较差,它们的潜在变化趋势也较大。

Nowadays the fractal dimension is often calculated on raster data, but most of existing land-use data is stored as vector data in fact. If these vector data are converted to images to calculate fractal dimension, perhaps some pixels with inaccurate grey values will be resulted from the ＂GRID＂ structure of raster data. The precision calculated on raster data is closely related to the size of pixel and grid image. In this paper, a computation program for the fractal dimension of 2D vector data based on Windows platform has been designed by using Visual C sharp. Now the program has been successfully applied to land-use data of the middle Qinling Mountains and the southeast of Hubei Province in China. The results show that the program is a convenient, reliable and precise method for the fractal dimension of 2D vector data. The land use and land cover types arranged in the descending order of their average stability indices are as follows. 1） The fractal dimensions of the middle Qinling Mountains are in the order of construction land 〉 unutilized land〉 paddy field 〉 dry upland 〉 grassland 〉 forestland 〉 water area in 1990, 1995 and 2000. 2） The fractal dimensions of the southeast of Hubei Province in China are in the order of paddy field 〉 unutilized land 〉 construction land 〉 dry upland 〉 forestland 〉 grassland 〉 water area in 1990, 1995 and 2000. In a word, forestland, water area and grassland have poorer stability and higher possibility of changes in the two study areas.