面向地形特征的DEM与影像纹理差异分析

发布时间：2018-05-01 21:56

本文选题：地形特征 + 纹理分析　；参考：《地球信息科学学报》2016年03期

【摘要】：纹理分析方法在宏观地形特征分析方面具有较大的优势与潜力,但当前缺少对DEM与影像数据纹理特征差异的系统分析研究。本文采用灰度共生矩阵为纹理量化模型,选取了8个不同地貌单元的样本数据,对DEM和遥感影像2类数据的纹理进行了特征值对比分析、纹理特征稳定性分析、纹理特征组间差异性分析。实验结果表明,在所测试的二阶角矩、对比度、方差、熵4个纹理指标中,DEM和影像的对比度特征值间具有显著的相关性;通过不同地貌样区纹理特征值对比分析发现,DEM数据在地形起伏较大区域纹理特征更为明显,遥感影像数据则受地表覆盖物影响较大;从地形特征的稳定性角度分析,DEM数据在丘陵和山地分析有优势,影像数据则在平原和台地分析表现更好;从地形特征差异性角度分析,DEM数据要优于影像数据。进一步采用光照模拟和坡度数据以增加DEM纹理信息,研究结果表明,DEM派生的2类数据在地形量化差异性方面改进明显,并大大优于影像数据。
[Abstract]:Texture analysis method has great advantages and potential in macroscopic terrain feature analysis, but there is no systematic analysis of the difference between DEM and image data. In this paper, the gray-scale co-occurrence matrix is used as the texture quantization model, and the sample data of eight different geomorphological units are selected. The texture characteristics of DEM and remote sensing images are compared and analyzed, and the stability of texture features is analyzed. Difference analysis of texture feature groups. The experimental results show that there is a significant correlation between Dem and the contrast eigenvalues of the image in the four texture parameters: second order angular moment, contrast, variance and entropy. It is found that the texture characteristics of Dem data are more obvious in the areas with large terrain undulation, while the remote sensing image data are greatly affected by the ground cover through the comparative analysis of texture characteristic values of different geomorphologic sample areas. From the angle of stability of terrain features, Dem data has advantages in hilly and mountainous analysis, while image data is better in plain and platform analysis, and Dem data is better than image data in terms of terrain characteristics difference. Furthermore, the illumination simulation and slope data are used to increase the DEM texture information. The results show that the two types of data derived from Dem are much better than the image data in terms of the difference in topographic quantification.
【作者单位】：南京师范大学虚拟地理环境教育部重点实验室;江苏省地理信息资源开发与利用协同创新中心;
【基金】：江苏高校优势学科建设工程资助项目国家自然科学基金项目(41171320、41271438、41401440)
【分类号】：P208

【参考文献】