论文部分内容阅读
在SAR/INS组合导航系统中,所获取的SAR图像可能存在严重的斑点噪声和几何变形,对此,提出一种基于CenSurE特征的SAR/INS组合导航用景象匹配算法.该算法针对惯性组合导航的工作特点进行设计.首先提取CenSurE特征和垂直的SURF描述符,利用夹角余弦相似度量方法进行特征匹配;然后,采用分组一致采样算法和最小二乘精确匹配算法获取高精度的航向和位置偏差信息.景象匹配性能评价实验表明,在匹配适应性、匹配速度、精度和鲁棒性等方面,CenSurE特征都很优越,可以满足SAR/INS景象匹配导航系统匹配修正的高性能要求.
In the SAR / INS integrated navigation system, there may be severe speckle noise and geometric deformation in the acquired SAR image, and a CenSurE-based scene matching algorithm for SAR / INS integrated navigation is proposed.This algorithm aims at inertial integrated navigation Firstly, the CenSurE feature and the vertical SURF descriptor are extracted, and the features of the cosine similarity measure method are used for feature matching. Then, the grouping consistent sampling algorithm and the least-squares accurate matching algorithm are used to obtain the high-precision heading and position deviations Information.Simulation experiments show that the features of CenSurE are superior in matching adaptability, matching speed, accuracy and robustness, which can meet the high performance requirements of match correction of SAR / INS scene matching navigation system.