强光束干扰下低轨高清遥感图像无缝拼接方法改进研究

发布时间：2019-04-23 18:27

【摘要】：由于受到强光束的干扰,低轨高清遥感图像拼接处的拼接边界容易出现突变。当前图像拼接方法因不能避免外界因素的干扰,无法实现图像的无缝拼接。为此,提出一种新的强光束干扰下低轨高清遥感图像无缝拼接方法。通过Retinex理论对强光束干扰下低轨高清遥感图像进行预处理,利用原始低轨高清遥感图像的照射光得到物体的反射性质,以获取物体自身的面貌。介绍了尺度不变特征变换(SIFT)算法的基本思想,通过SIFT特征匹配法对遥感图像进行配准处理;针对SIFT法效率低的弊端,利用区域分块法对其进行改进。通过幂比值拉伸方法对拼接缝进行消除。实验结果表明,采用所提方法得到的拼接图像拼接处无重叠,较为平滑,在保证拼接图像信息量、清晰度和整体质量的同时,效率较高。
[Abstract]:Because of the interference of strong beam, the splicing boundary of low-orbit high-definition remote sensing image is easy to change. Current image mosaic methods can not achieve seamless image mosaic because they can not avoid the interference of external factors. In this paper, a new high-definition remote sensing image stitching method with strong beam interference is proposed for low-orbit high-definition remote sensing images. The Retinex theory is used to pre-process the low-orbit high-definition remote sensing image with strong beam interference. The reflection property of the object is obtained by using the illuminating light of the original low-orbit high-definition remote sensing image, so as to obtain the appearance of the object itself. The basic idea of scale invariant feature transform (SIFT) algorithm is introduced in this paper, and the remote sensing image registration is processed by SIFT feature matching method, aiming at the disadvantage of low efficiency of SIFT method, the region block method is used to improve it. The splicing seam is eliminated by power ratio tensile method. The experimental results show that the splicing image obtained by the proposed method has no overlap and is relatively smooth. It can ensure the amount of information, clarity and overall quality of the splicing image, and the efficiency is high at the same time.
【作者单位】：商丘工学院信息与电子工程学院;
【基金】：河南省教育厅科学技术研究重点项目(12B520047)资助
【分类号】：TP751

【相似文献】