捷联图像导引系统传感器数字复合式稳像方法  

Strapdown Image Homing System Sensor and Digital Compound Image Stabilization

作　　者：方丹[1] 郭晓冉 崔少辉[1] 

机构地区：[1]军械工程学院导弹工程系,河北石家庄050003 [2]军械技术研究所,河北石家庄050000

出　　处：《微电子学与计算机》2016年第3期50-55,共6页Microelectronics & Computer

基　　金：武器装备军内科研项目(装司字[2012]665)

摘　　要：针对传感器稳像与数字稳像各具优势而又存在不足的问题,提出了传感器稳像与数字稳像相结合的复合式稳像方法,实现两种方法之间的优势互补.首先,利用SIFT算法检测图像中的特征点,并在特征点匹配部分将陀螺数据解算出的像面位移量所指向的匹配点位置作为特征点匹配时的初始搜索位置,指示当前帧在参考帧中以该匹配点位置为中心的一定大小矩形区域内搜索匹配特征点;然后,利用欧式距离最短的几对匹配点在仿射变换模型下计算出帧间变换矩阵;最后,利用基于视觉暂留机制的参考帧更新方法,对每一幅当前帧图像进行重建,完成视频序列的复合式稳像.实验结果表明,该方法在保证较高稳像精度的同时,降低了数字稳像方法的计算量,提升了运算速度,可满足捷联图像导引系统对稳像精度及实时性的要求.For the sensor image stabilization and digital image stabilization each has its advantages and insufficient, a compound image stabilization algorithm is proposed. First of all, the feature points are extracted using SIFT algorithm, and the indicated position by gyro data is made as the center point of initial sourcing position in a certain size rectangular area to search for matching feature points. Then, the interframe transformation matrix under affine transformation model is calculated using the shortest euclidean distance of matching point pairs. Finally, using the reference frame update method based on the mechanism of persistence of vision, each current frame image is reconstruction, and the compound stabilization algorithm is completed. Experimental results show that the compound image stabilization method is reducing the computational method of digital image stabilization and improving the speed of operation, it can meet the requirement for image stabilization precision and real-time performance of the strapdown image guidance system.

关 键 词：图像稳定 传感器稳像 数字稳像 复合式 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]