摘要
多源遥感数据融合是遥感技术向纵深发展的必然趋势。本文对多源遥感数据融合算法的应用特点,从基于像元的融合、特征的融合以及决策级融合3个层次上进行了详细的分析,并以河北丰宁县为例,说明遥感数据融合方法在遥感信息提取中的具体应用:对所用数据进行预处理,然后对1999年Landsat TM数据进行主成分变换,前3 个主成分占总信息量的97.8%,主成分逆变换后的结果影像更清晰,层次更丰富。信息提取时选择TM全色和主成分变换后的多光谱数据融合后的影像,波段4、3、2和波段5、4、3的彩色合成方案,并对植被指数和穗帽变换后的绿度指数进行了分析,遥感影像与DEM以及与GIS空间数据的信息融合也可以提高遥感信息提取的精度。最后分析了多源遥感数据融合尚待解决的问题及努力方向。
Multi-source remote sensing data fusion is the development trend of remote sensing technology in depth. This paper analyzes in detail algorithmic application characteristics of multi-source remote sensing data from three levels of pixel-based, feature-based and decision-based fusion processings. Take Fengning County for example, specific applications of remote sensing data fusion methods in information extraction are illuminated. The data used in this study is firstly pre-processed, then the principal components of Landsat TM data in 1999 are analyzed, the first three principal components account for 97.8% of the total information, the resulted image of inversed principal components transformation is clearer and has more abundant levels. To extract information from remote sensing image, we select the fusion image from Landsat TM pan and multi-spectral bands after principal components transformation, color composition scheme of bands 4, 3, 2 and bands 5, 4, 3, and vegetation index and greenness index after tasseled cap transformation are analyzed, the remote sensing image information fusion with DEM and spatial data of GIS database can also improve the accuracy of remote sensing information extraction. Problems to be resolved and future direction of multi-source remote sensing data fusion are put forward.
出处
《地球信息科学》
CSCD
2005年第3期97-103,共7页
Geo-information Science
基金
国家自然科学基金项目(40171001)河北师范大学科学研究基金项目(L2003Q13)
关键词
多源数据
遥感
信息融合
multi-source data
remote sensing
information fusion
