摘要
通过实验,系统分析了已有的滤波、边缘检测、亚像素边缘定位和中心定位算法,从精度和时间效率两方面对比了各种算法的性能,并提出了矩算子亚像素边缘定位的改进措施。在此基础上提出了一种新的应用于InSAR基线测量的目标定位算法,新算法通过"目标匹配、分割→噪声滤除→边缘粗定位→亚像素边缘定位→目标中心定位"的设计思想,达到了InSAR系统规定的指标:0.05 pixel的定位精度和50 ms的运行周期。实验表明该算法具有良好的抗干扰性能,较适用于复杂的太空环境。
This paper deals with locating the synthetic targets precisely and dynamically. Based on experiments,authors analyze common methods such as filtering, edge detection, sub-pixel edge localization and central localization and compare them according to precision and time efficiency. Furthermore, the performance of moment operators is improved. More importantly,the paper presents a scheme applied to the baseline determination of dual- antenna InSAR. The new scheme consists of target matching, segmentation, noisy filtering,edge coarse localization,sub-pixel localization and central localization. The aim of 0.05 pixel accuracy and 50ms' period is achieved. The experiments indicate that the algorithm proposed is anti-noise outstandingly and more applicable to the complicated circumstance of the outer space.
出处
《遥感技术与应用》
CSCD
2008年第5期537-544,共8页
Remote Sensing Technology and Application
基金
国防科工委民用航天背景型号项目(20078194-02)