坡度理论分布与分辨率的关系

发布时间：2018-03-30 18:41

本文选题：分辨率　切入点：平滑程度　出处：《西北大学》2017年硕士论文

【摘要】：坡度是很多地表过程的重要影响因子,也是水文学、土壤学、遥感图像处理等学科中重要的地表形态指标之一。以往对于坡度分布的研究主要侧重于坡度的统计分布与地形类型间的关系。然而,基于DEM数据的坡度理论模型研究中,DEM分辨率对坡度理论分布的影响至关重要。本文选取五个典型样区,试图探讨在不同地形类型区,坡度理论分布与分辨率的关系,并深入探讨分辨率的两个组分——地形平滑和采样间距对高程梯度(p、q)、高程梯度联合方差(σg)、坡度相关统计特征和坡度理论分布的影响。本文所得主要结论如下:(1)基于信息含量法和采样定理法对选取的五个典型样区DEM适宜分辨率进行了确定,结果为:拜泉样区适宜栅格尺寸为20m,延安、绥德和横断山样区适宜栅格尺寸为10m,遂宁样区适宜栅格尺寸为5m。(2)随着分辨率的降低,高程梯度p、q的图像表面逐渐变得平缓和光滑,图像表面信息含量逐渐减少;p、q值域和标准差、高程梯度联合方差以及坡度统计特征值(平均值和标准差)随分辨率的降低均逐渐减小。分辨率的两个组分(平滑程度和采样间距)对以上结果均有影响,但地形平滑作用都是主要影响因素,且在地形较为复杂的地区,地形平滑的作用更加明显。(3)坡度频率和累积频率与分辨率的关系相似:随分辨率的降低,频率和累积频率曲线逐渐向低坡度段移动,坡度值域逐渐减小,频率峰值区也逐渐变窄。根据直方图相似度指数(HS)的结果显示,五个样区的HS值均大于0.8,黄土丘陵区(延安和绥德)的HS最高,说明坡度理论分布模型在对坡度分布的模拟效果整体很好,黄土丘陵区效果更佳。相比之下,极陡山地模拟效果稍欠佳。
[Abstract]:Slope is an important factor in many surface processes, as well as hydrology, soil science, In the field of remote sensing image processing and other disciplines, one of the important indicators of surface morphology. The previous study of slope distribution mainly focused on the relationship between the statistical distribution of slope and the type of terrain. However, The influence of Dem resolution on the theoretical distribution of slope is very important in the study of slope theoretical model based on DEM data. In this paper, five typical samples are selected to study the relationship between the theoretical distribution of slope and the resolution in different terrain types. The effects of topographic smoothing and sampling spacing on the elevation gradient, the joint variance of elevation gradient (蟽 g), the statistical characteristics of slope correlation and the distribution of slope theory are discussed in depth. The main conclusions of this paper are as follows: 1. The information content method and sampling theorem method are used to determine the appropriate DEM resolution of the selected five typical samples. The results showed that the suitable grid size was 20 m in Baiquan sample area, 10 m in Yan'an, Suide and Hengduan sample areas, and 5 m 路m ~ 2 in Suining sample area.) with the decrease of resolution, the image surface of elevation gradient PQ gradually became smooth and smooth. The surface information content of the image gradually decreases the range and standard deviation. The joint variance of elevation gradient and the statistical eigenvalue (mean and standard deviation) of slope decrease with the decrease of resolution. The two components of the resolution (smoothness and sampling spacing) have influence on the above results. However, topographic smoothing is the main influencing factor, and the effect of topographic smoothing is more obvious in areas with complex terrain. The relationship between slope frequency and cumulative frequency and resolution is similar: with the decrease of resolution, The curve of frequency and cumulative frequency gradually moves to the low slope, the range of slope decreases gradually, and the peak region of frequency becomes narrower. According to the histogram similarity index (HS), the results show that, The HS values of the five sample areas are all greater than 0.8, and the HS of the loess hilly region (Yan'an and Suide) is the highest, which indicates that the slope theoretical distribution model is effective in simulating the slope distribution as a whole, and the effect in the loess hilly area is better than that in the loess hilly region. The simulation effect of very steep mountain is not good.
【学位授予单位】：西北大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：P208

【相似文献】