摘要
本文面向灾害应急的无人机遥感影像快速拼接融合算法的研究,满足灾害应急对灾情影像数据时效性的需求。为研究算法的融合改进对无人机近红外遥感影像精度的提升问题,利用SURF计算效率高和MSERS能够提取完全仿射不变区域特征的特点,提出了无人机遥感影像特征提取与匹配的方法,该方法是基于SURF融合特征,在影像和尺度空间均具有稳定性且具有多尺度仿射不变的可靠性。并对MS-MSERS、MSERS+SIFT、MSERS+SURF三种算法进行配准效果验证。结果表明:MSERS+SURF的匹配正确率优于其他两种算法。该算法实现了MSERS的多尺度检测及拼接效率成果较好的平衡。
This article discusses the application of fast fusion algorithm to UAV remote sensing images; thus, fulfilling the needs of disaster emergency response. In order to enhance UAV remote sensing images and improve accuracy of near infrared imaging by fusion algorithm-making use of the characteristics of SURF （high calculation efficiency） and MSERS （ability to fully extract the affine invariant region features）-a method for extraction of remote sensing human images and matching, which is reliable and efficient in image and scale space, is put forward. And the effect of registration and verification of MS-MSERS, MSERS＋SIFT, MSERS＋SURF-three kinds of algorithms is ascertained. The results show that the matching accuracy of MSERS＋SURF is better than the other two algorithms. The algorithm achieves multi-scale detection of MSERS and a good balance between stitching efficiency and stitching results.
出处
《红外技术》
CSCD
北大核心
2018年第2期146-150,共5页
Infrared Technology
基金
国家自然科学基金。遥感云图-电磁波-热红外对地震等重大地质灾害的预测机理(41371492)及基于光谱参量耦合的光谱定标机理及其精度评价方法研究(20130001110046)
关键词
无人机
近红外光谱
配准算法
MSERS
SURF
unmanned aerial vehicle（UAV）,infrared spectrum,registration algorithm,MSERS,SURF
