提出了一种基于灰关联分析的多景遥感影像最佳镶嵌线的检测算法。该算法根据中心像元与周围像元之间的临近像元效应,在两景影像重叠区域选择中心像元邻域作为参考序列与比较序列,利用斜率关联度计算两者之间的灰关联度,据此找到一条影像色调和纹理差异较小的镶嵌线。以四组两景相邻影像自动寻找镶嵌线为例进行了实验研究,取得了良好效果,验证了该方法的可行性与有效性。
This paper puts forward a novel algorithm based on grey relational analysis for detecting an optimal seam line in remote sensing image mosaic.First,the algorithm regards pixel neighboring sequence in overlapping area as referential sequence and compared sequence because of the relation of center point and neighboring points,then computes grey relational degree of referential sequence and compared sequence using slope relational degree,finally finds an optimal seam line which is less differences in color and texture between the images.Four sets of experimental results show the seam line detected by slope relational degree is more superior to that detected by normal method and verify the feasibility and validity of the new method.