鲁棒Hausdorff距离在SAR/惯性组合导航图像匹配中的应用研究

Image matching based on robust Hausdorff distance in SAR/INS inte grated navigation system

分析了适用于SAR景象匹配辅助导航系统的图像匹配算法的特点 ,提出了一种新型Hausdorff距离的鲁棒型图像匹配算法 .首先分析了传统的Hausdorff距离算法的不足 ,然后引入了一种能有效消除高斯噪声影响的新算法 .同时针对此算法图像匹配时全局搜索的不足 ,提出了扫描终止法技术 ,以有效解决图像匹配实时性的问题 .针对高斯噪声的大小对图像匹配的影响进行了仿真实验 .仿真结果表明 ,新图像匹配算法能大大提高合成孔径雷达图像的匹配精度和实时性 ,对组合导航系统的精度提高有重要的意义 ,是一种行之有效的方法 .

Based on the characters adapted to the SAR image matching in navigation system, a new robust Hausdorff distance image matching algorithm is proposed. The limita tions of conventional Hausdorff distance measures are analyzed and a new algorit hm which can eliminate the Gaussian noise in image is presented. Aiming at the glob al search shortage in image matching, a scan termination method is also proposed to solve the real-time problem. The effect of Gaussian noise on image matching was simulated and the results show that the proposed algorithm is more efficien t than the conventional ones, and can also improve the integrated navigation sys tem accuracy.