摘要
利用南京市1988、1998和2006年Landsat TM遥感影像,采用小波分析技术,对1988—2006年南京市中心城区的景观结构特征进行了研究.结果表明:应用小波分析技术分析景观样带数据时,取样宽度过小易导致检测结果的不稳定,而样带过宽则易导致较小尺度特征的丢失,应采用合适的样带宽度;基于TM影像(30m空间分辨率)的南京市城区样带宽度以4个像元为最佳.1988、1998和2006年,南京市中心城区都存在4个相对稳定的特征尺度域,分别对应于功能小区、小型街区、中型街区和大型街区;研究期间,大型街区的特征尺度几乎没有变化,而其它3个尺度域的变化相对较大,但各尺度域的变化趋势一致;各尺度域内特征尺度之间的差异性均逐渐变小,表明期间中心城区的结构特征越来越稳定,不同部位的空间结构趋同,且各尺度域的变化有较强的同步性.
Based on 3 scenes of Landsat TM image taken in 1988, 1998 and 2006, the changes of the structural characteristics of Nanjing central urban area were studied with one-dimensional wavelet analysis. The results showed that it was necessary to select an optimal sampling width when using wavelet analysis to detect the charaeteristic scales of landscape transects. Too narrow transects could cause lack of stabilization, while too wide ones could cause loss of some minor characteristic scales. From the TM images （30 m resolution） of Nanjing urban area, the optimal sampling width was identified as four pixels. The results from one-dimensional wavelet transform indicated that the landscape of Nanjing central urban area had four relatively stable characteristic scale domains in 1988, 1998 and 2006, corresponding to functional clumps, small blocks, middle-sized blocks, and large blocks, respectively. From 1988 to 2006, the scales of large blocks were nearly unchanged, while the other three domains all changed significantly with a similar trend. Moreover, the varia- tions among the characteristic scales within each scale domain declined from 1988 to 2006, indicating that the landscape structure of Nanjing central urban area tended to be more stable and accordant. The scale domains showed synchronous changes as well.
出处
《应用生态学报》
CAS
CSCD
北大核心
2008年第3期606-612,共7页
Chinese Journal of Applied Ecology
基金
国家自然科学基金资助项目(39670364)
南京大学人才引进培养基金项目(22410120)
关键词
等级结构
小波分析
样带法
最适宽度
尺度域
南京
hierarchical structure
wavelet analysis
transect
optimum width
scale domain
Nanjing.
