Research On Fast Detection Of Seam-line For GF Image Via Image Segmentation

With the development of remote sensing in China,the application advantage of remote sensing data in regional macro research is more and more obvious,and the application scope is more and more extensive.Because of the improvement of spatial resolution of remote sensing data,more and more maps are needed in the same area,so mosaic becomes the most important task in remote sensing data processing.In order to ensure the image quality of mosaic,seam-line is needed in the process of mosaic.But now,the auto-generated of seam-line has been the bottleneck of GF data automatic production and it's quality won't satisfy the demand of research at present.A fast and feasible method is proposed for existing problems in the auto-generation of seam-lines in this study.Firstly,a fast and simple method is proposed to generate the effective range of GF data based on the characteristics of the sensing satellite data.Secondly,the overlapping region of the adjacent images and the initial seam-line are generated based on the effective range.Then,the morphological boundary of the image is obtained by watershed image segmentation and a feasible and a feasible method is proposed to carry out morphological boundary point and line information.Finally,the optimal seam-line is obtained through the Dijkstra algorithm.In this paper,the algorithm of fast detection of seam-line for GF image Via image segmentation use c++ language for implementation and take GF-2 data as research object.The experimental results show that the proposed method can be more rapid for the generation of the seam-line,The seam-line searches the morphological boundary of the ground objects,rarely across the features,avoids the complex features such as construction and has a better effect.