遥感数据的高斯金字塔尺度上推方法研究
发布时间：2018-12-13 05:28
【摘要】：尺度转换是遥感信息科学领域的研究热点,其传统研究方法大多局限于统计模型,对数据的空间结构信息考虑较少,很难满足遥感数据的多尺度表达要求。基于此,针对遥感数据的尺度不一致问题,本文提出了一种利用高斯金字塔的图像模糊特性进行遥感数据尺度上推的方法,在对金字塔每一层的数据高斯模糊的基础上,通过多次连续的降采样,得到一系列不同尺度的数据,从而满足实际应用的空间分辨率要求。为了验证本文所提方法的有效性,本文选择Landsat7 ETM影像和ASTER GDEM为研究数据进行尺度上推,并与传统的最邻近、双线性以及立方卷积等方法进行了实验对比,采用均值、方差、均方根误差、平均绝对误差等评价指标,以及相同分辨率的ASTER GDEM和SRTM DEM的等高线套合结果来衡量高斯金字塔方法的性能。实验结果表明,本文使用的高斯金字塔尺度上推方法能够有效地实现连续遥感数据的尺度转换,在保持遥感数据局部细节特征的基础上,较好地保持了原始遥感数据的信息量以及空间结构特征。
[Abstract]:Scale conversion is a hot topic in the field of remote sensing information science. Its traditional research methods are mostly confined to statistical models, and the spatial structure information of data is less considered, so it is difficult to meet the requirements of multi-scale representation of remote sensing data. Based on this, aiming at the problem of scale inconsistency of remote sensing data, this paper proposes a method to push up the scale of remote sensing data by using the fuzzy characteristics of Gao Si's image. A series of data of different scales are obtained by several successive demultiplexing, which meet the spatial resolution requirements of practical applications. In order to verify the validity of the proposed method, Landsat7 ETM image and ASTER GDEM are selected as the data to be used for scaling up, and the experimental results are compared with the traditional methods such as nearest neighbor, bilinear and cubic convolution. The mean value, variance and variance are used. The performance of Gao Si pyramid method is evaluated by the RMS error, the mean absolute error and the contour fitting results of ASTER GDEM and SRTM DEM with the same resolution. The experimental results show that Gao Si's pyramid up-scaling method can effectively realize the scale conversion of the continuous remote sensing data, and on the basis of preserving the local detail features of the remote sensing data, The information content and spatial structure of the original remote sensing data are well maintained.
【作者单位】： 中国科学院遥感与数字地球研究所;中国科学院大学;中国地质大学(武汉)计算机学院;中国科学院地理科学与资源研究所;
【基金】：中国科学院遥感与数字地球研究所所长基金项目“基于动态追踪树的区域计算型GIS空间分析并行化研究(Y6XS6300CX)”
【分类号】：P237
本文编号：2375968
[Abstract]:Scale conversion is a hot topic in the field of remote sensing information science. Its traditional research methods are mostly confined to statistical models, and the spatial structure information of data is less considered, so it is difficult to meet the requirements of multi-scale representation of remote sensing data. Based on this, aiming at the problem of scale inconsistency of remote sensing data, this paper proposes a method to push up the scale of remote sensing data by using the fuzzy characteristics of Gao Si's image. A series of data of different scales are obtained by several successive demultiplexing, which meet the spatial resolution requirements of practical applications. In order to verify the validity of the proposed method, Landsat7 ETM image and ASTER GDEM are selected as the data to be used for scaling up, and the experimental results are compared with the traditional methods such as nearest neighbor, bilinear and cubic convolution. The mean value, variance and variance are used. The performance of Gao Si pyramid method is evaluated by the RMS error, the mean absolute error and the contour fitting results of ASTER GDEM and SRTM DEM with the same resolution. The experimental results show that Gao Si's pyramid up-scaling method can effectively realize the scale conversion of the continuous remote sensing data, and on the basis of preserving the local detail features of the remote sensing data, The information content and spatial structure of the original remote sensing data are well maintained.
【作者单位】： 中国科学院遥感与数字地球研究所;中国科学院大学;中国地质大学(武汉)计算机学院;中国科学院地理科学与资源研究所;
【基金】：中国科学院遥感与数字地球研究所所长基金项目“基于动态追踪树的区域计算型GIS空间分析并行化研究(Y6XS6300CX)”
【分类号】：P237
【相似文献】
相关期刊论文 前1条
1 韩鹏;龚健雅;李志林;程亮;;遥感影像空间尺度上推方法的评价[J];遥感学报;2008年06期
，本文编号：2375968
本文链接：https://www.wllwen.com/kejilunwen/dizhicehuilunwen/2375968.html
