朱庆, 于杰, 杜志强, 张叶廷. 面向对象的真正射影像纠正方法[J]. 武汉大学学报 ( 信息科学版), 2013, 38(7): 757-760.
引用本文: 朱庆, 于杰, 杜志强, 张叶廷. 面向对象的真正射影像纠正方法[J]. 武汉大学学报 ( 信息科学版), 2013, 38(7): 757-760.
ZHU Qing, YU Jie, DU Zhiqiang, ZHANG Yeting. An Object-Oriented True Ortho Image Rectification Method[J]. Geomatics and Information Science of Wuhan University, 2013, 38(7): 757-760.
Citation: ZHU Qing, YU Jie, DU Zhiqiang, ZHANG Yeting. An Object-Oriented True Ortho Image Rectification Method[J]. Geomatics and Information Science of Wuhan University, 2013, 38(7): 757-760.

面向对象的真正射影像纠正方法

An Object-Oriented True Ortho Image Rectification Method

  • 摘要: 现有"像素级"的真正射影像纠正方法由于没有充分考虑地物特征和影像像素间的关联关系,导致对DSM分辨率十分敏感,难以保持地物轮廓边缘特征的准确性和纹理结构的完整性,遮挡恢复和阴影补偿难以自动化处理,需要大量人工干预等问题。提出了一种面向对象的真正射影像纠正方法,主要内容包括:①物方和像方对象的定义及语义描述;②像方对象的全局可见性索引;③面向对象的真正射纠正和纹理优化采样。选择广东阳江地区多角度航摄影像进行了实验,结果表明,面向对象的真正射纠正方法既能保持准确的几何特征和完整的纹理结构,还能自适应地处理遮挡和阴影,为多角度高分辨率影像数据的自动化、智能化真正射纠正提供了一种有效的新途径。

     

    Abstract: After reviewing the existing true ortho image rectification methods,we analyze image rectification and texture resampling processes that use a pixel-oriented strategy.These strategies generate inaccurate geometric features,imperfect texture structures of man-made objects.Thus,occlusion recovery and shadow compensation are difficult to automatically process,requiring much manual intervention.Aiming to solve these problems,this paper proposes a novel object-oriented true-ortho rectification method,implementing a high-level object-oriented strategy instead of a low-level pixel-oriented strategy.There are three main steps:① definition of physical objects and image object containing geometric and semantic information.In 3-D physical space,all the objects in the horizontal plane of projection form a seamless surface comprised of a series of contiguous triangular facets,in this paper this surface is represented by Semantics constrained Triangulated Irregular Network(STIN).The physical objects are characterized by different aspects: property,outline,geometry and topological relation,extracted from 3-D city models and point clouds.The image objects are made up of several pixels with semantic descriptions for spectral information,2-D geometry and topological relations.They can be extracted by image segmentation,edge detection,and texture clustering.② establishment of a global hierarchical spatial index of image objects.After deriving visual correspondence relationship between 3-D physical objects and 2-D image objects,we use the index to organize all the information from physical and image objects efficiently and provide a foundation for high performance true-ortho rectification and optimized image sampling.③ object-oriented true-ortho rectification and optimal image resampling.True-ortho rectification is carried out based on the physical objects as STIN.And based on the semantic links in the global index,the image objects with semantic information on visibility,integrality,and radiometric features can be chosen for optimized sampling to adaptively handle occlusion and shadows.Experimental results revealed that the proposed method has the advantages of high accuracy,high efficiency and full automation.

     

/

返回文章
返回