顾及小波变换的土地利用变化与经济因子的多尺度相关性分析  

Multi-scale Correlation Analysis on Land-Use Change and Economic Factors in a Typical Region of Wavelet Transforming

作　　者：陈江平[1] 韩青[1] 胡晶[1] 张鹏林[1] 

机构地区：[1]武汉大学遥感信息工程学院,武汉市珞喻路129号430079

出　　处：《武汉大学学报（信息科学版）》2013年第9期1118-1121,1130,共5页Geomatics and Information Science of Wuhan University

基　　金：国家科技支撑计划资助项目(2012BAJ15B04);国家自然科学基金资助项目(61172175);国家自然科学青年科研基金资助项目(40801152)

摘　　要：以武汉市为研究区域,借助小波分析工具,揭示了土地利用变化的特征尺度。通过小波方差-尺度图可以辨别出,样区中64m尺度域上的小波方差为极值点,能较好地反映土地利用变化的空间格局。在特征尺度分析的基础上进行土地利用变化与经济影响因子多尺度相关性分析,结果表明,各经济影响因子的对土地利用变化的制约性随尺度的增大不断增强,小尺度上经济因子与土地利用变化的相关性较小;在特征尺度上,一般经济因子的相关系数较大,原因在于特征尺度上土地利用的空间结构信息丰富,能较好地反映经济因子与土地利用变化的相关性;大尺度上随着信息的不断合并,多数经济因子相关系数增长缓慢并逐渐趋于稳定;多尺度相关性分析反映了经济影响因子作为宏观制约因素影响样区土地利用的变化,也验证了特征尺度分析的有效性。The analysis of multi-scale land-use change and the driving force factors behind it has become an important direction for research In this study, a wavelet analysis tool was ap- plied to analyze the multi-scale correlation between land-use change and economic factors based on characteristic scale analysis. The results showed that, in the study area, the scale of 64m was regarded as the characteristic scale and optimal to identify land-use heterogenei- ty. Wavelet variance revealed local information but failed to describe the general spatial pat- tern of land-use at a finer scale, and shapely raised for combination of information along with the upscaling. The results indicated that correlation between land-use change and economic factors was scale-dependent： the correlation coefficient values were smaller at a finer scale and reached the extreme at the characteristic scale. However, at a coarser scale, the correla- tion coefficient values of most economic factors became flat . This analysis suggests that eco- nomic factors effecting land-use change are macro constraints, and at the same time, shows the effectiveness of characteristic scale analysis. The coefficients among different factors are also different, under high-frequency wavelet coefficients there was strong positive correlation between population and land-use change at each scale, but a weak negative correlation in Per Capita income of rural households, suggesting that population was the most influential factor in land-use change. This study shows that wavelet analysis is a powerful tool for multi-scale correla- tion analysis, and can effectively reveal the multi-scale spatial patterns in land-use change.

关 键 词：土地利用变化 小波变换 经济因子 相关性分析 

分 类 号：P208[天文地球—地图制图学与地理信息工程]