摘要
针对飞机形状特点,提出一种基于面向对象分析形状特征的高分辨率遥感影像飞机提取方法。首先通过区域增长的影像分割方法获得备选对象,从对象的质心和主轴方向出发建立象限划分,根据各象限像素分布直方图计算飞机形状特征因子。利用飞机形状特征因子,结合光谱灰度值,采用简单的阈值提取即可有效地提取飞机。实验证明该方法具有较好的鲁棒性和适应性。
Aircraft is a special man-made object with distinct shape.In this paper,an aircraft extraction method is presented based on shape feature by object-oriented analysis from high resolution remote sensing image.Firstly,the candidate objects are acquired by region-growing segmentation of the image.Centroid and primal axis are calculated for all objects.Sectors partition is implemented by the centroid and orientation of primal axis.A novel shape feature index is calculated by pixels distribution histogram in different sectors.Aircraft can be then effectively extracted by setting simple thresholds for the shape feature index and spectral value.Experiment shows the method is of good robustness and adaptation.
出处
《遥感信息》
CSCD
2013年第6期3-6,共4页
Remote Sensing Information
基金
国家自然科学基金(41171326)
关键词
影像分割
形状特征
飞机提取
image segmentation
shape feature
aircraft extraction
