摘要
针对现有光学影像与合成孔径雷达(SAR)影像之间匹配存在精度差、效率低等问题,提出了一种基于面特征的影像自动匹配方法.用中值代替均值的方法改进了Lee滤波器,用于对SAR影像相干斑噪声进行抑制.通过Canny算子提取出边缘,并用形态学膨胀对不连续边缘进行连接,在此基础上用区域生长法提取光学和SAR影像中的封闭区域作为面特征.设计了一种综合考虑面积、周长、中心点位置等属性的代价函数对面特征进行交叉匹配,筛选出匹配上的大目标,并将其中心作为匹配控制点.通过Envisat ASAR影像与Landsat ETM+影像进行的匹配试验表明,该方法不需要人工干预,可以实现自动匹配,且匹配精度小于1个像素.
Aiming at resolving the problem of lower-precision and labor-consuming in matching for synthetic aperture radar （SAR） and optical remote sensing images, a matching method based on region features was proposed. Speckle noises in SAR images were suppressed by a modified Lee filter which takes the advantage of median filter into account. Morphological dilations were used to link image edges extracted by Canny operator. Image segmentation based on region growing was used to extract closed regions in SAR and optical image as invariable features. Precise feature matching was guided by a cost function, which considering region areas, perimeters, positions of center points, etc. After cross matching, the center points of the matched region were used as matching control points. The experimentation of registering Envisat ASAR and Landsat TM images shows the whole workflow could be accomplished automatically with a precision of lower than 1 pixel.
出处
《中国矿业大学学报》
EI
CAS
CSCD
北大核心
2007年第6期843-847,共5页
Journal of China University of Mining & Technology
基金
国家高技术研究发展计划(863)项目(2005AA131021
2003AA135118)
关键词
面特征
SAR影像
自动匹配
region feature
SAR image
automatic matching
