径向基核函数在DEM内插中的适用性分析

发布时间：2018-03-19 20:59

本文选题：径向基函数　切入点：核函数　出处：《测绘科学》2017年02期 　论文类型：期刊论文

【摘要】：针对空间插值中不同核函数在不同地形条件下的适用性问题,该文选取了平原、丘陵和山地3种地貌类型区,采用5种常用的径向基核函数分别生成若干分辨率的DEM,通过高程统计参数、空间自相关特性和等高线套合分析DEM的精度差异。综合而言,平原地区中多重对数函数、多重二次曲面函数、反多重二次曲面函数和自然三次样条曲面函数插值效果没有显著差别,薄板样条曲面函数相对略差;丘陵地区选择多重二次曲面函数和薄板样条曲面函数插值效果最好,多重对数函数和自然三次样条曲面函数次之,反多重二次曲面函数最差;山地地区采用多重二次曲面函数插值能取得较为满意的效果,其他4种核函数逊之。
[Abstract]:Aiming at the applicability of different kernel functions in spatial interpolation under different terrain conditions, three types of geomorphologic regions, plain, hilly and mountainous, are selected in this paper. Five kinds of commonly used radial basis function are used to generate several resolution dems respectively. The accuracy difference of DEM is analyzed by height statistic parameters, spatial autocorrelation characteristics and contour alignment. In general, multiple logarithmic functions in plain area are obtained. There is no significant difference in the interpolation effect between the functions of multiple Quadric surfaces, inverse Quadric surfaces and Natural cubic Spline surfaces, and the function of thin plate spline surfaces is relatively poor. In the hilly area, the interpolation effect of multi-quadratic surface function and thin plate spline surface function is the best, the multi-fold logarithmic function and the natural cubic spline surface function are the second, and the inverse multi-quadratic surface function is the worst. In the mountainous area, the interpolation of multi-quadratic surface function can obtain satisfactory results, but the other four kernel functions are inferior to each other.
【作者单位】：南京工业大学测绘科学与技术学院;
【基金】：国家自然科学基金项目(41101360)
【分类号】：P208

【相似文献】