Method for perfectly splicing pillow-shaped hyperbolic edge images

The invention discloses a perfect splicing method for a pillow-shaped hyperbolic edge image, relates to the field of aerospace optical remote sensing image processing, and solves the problems of low splicing precision and local data redundancy of aerospace optical image splicing in a new stage. The method comprises the following steps: constructing a splicing model of a plurality of pillow-developing images; carrying out feature matching and information alignment between influences; and local redundant information is re-integrated and eliminated. According to the method, the utilization efficiency of large-breadth remote sensing image data is effectively improved, and a new thought is provided for a method for splicing multiple continuous large-breadth remote sensing images shot by a spacecraft. 一种枕形双曲边缘影像的完善拼接方法涉及航天光学遥感影像处理领域，解决了新阶段下航天光学影像拼接面临的拼接精度不高且局部数据冗余的问题。该方法包括：多幅枕兴影像拼接模型构建；影响间特征匹配及信息对齐；局部冗余信息再整合与剔除。有效提高了大幅宽遥感影像数据的利用效率，为航天器拍摄的连续大幅宽遥感影像多幅影像拼接方法提供了新的思路。