| 基于中值滤波加权修正的多波束声呐测深数据趋势面滤波方法 |
| |
| 引用本文： | 张兴伟,潘国富,张济博.基于中值滤波加权修正的多波束声呐测深数据趋势面滤波方法[J].海洋科学,2018,42(7):32-39. |
| |
| 作者姓名： | 张兴伟 潘国富 张济博 |
| |
| 作者单位： | 国家海洋局第二海洋研究所,国家海洋局第二海洋研究所,国家海洋局第二海洋研究所 |
| |
| 基金项目： | 上海市科学技术委员会科研计划项目(14DZl204600); 国家海岛保护专项资金 |
| |
| 摘 要： | 针对传统趋势面滤波方法中多项式拟合曲面系数向量的求取和作为阈值的均方根误差的求取都受到异常数据的影响,使该方法在异常测深数据较多的情况下滤波效果不佳的问题,提出了一种中值滤波加权修正的改进方法。在构造趋势面之前,对水深数据进行加权修正,以前后两次修正后数据的拟合优度的变化量作为是否进行下一步水深修正的依据,利用最终修正后的水深数据求取多项式拟合曲面系数向量和均方根误差,大幅降低了异常数据的影响,具有很强的抗差性。经仿真模拟数据和多波束实测数据滤波试验,该方法在异常数据较多的情况下依然良好,能够保持良好的滤波效果,明显优于传统趋势面滤波;同时,该方法能够保持较高的运算效率,适用于海量多波束测深数据的自动滤波。
|
| 关 键 词： | 多波束测深 趋势面滤波 抗差性 中值滤波 拟合优度 |
| 收稿时间： | 2018/3/9 0:00:00 |
| 修稿时间： | 2018/4/19 0:00:00 |
Trend surface filtering method for multibeam sonar echosounding data based on median filter weighed correction |
| |
| ZHANG Xing-wei,PAN Guo-fu and ZHANG Ji-bo.Trend surface filtering method for multibeam sonar echosounding data based on median filter weighed correction[J].Marine Sciences,2018,42(7):32-39. |
| |
| Authors: | ZHANG Xing-wei PAN Guo-fu and ZHANG Ji-bo |
| |
| Abstract: | Aiming at the problem of poor performance of traditional trend surface filtering when multibeam echosounding datasets contain too many outliers because the coefficient vector of the polynomial fitting surface and the root mean square error which is used as threshold are both affected by outliers, an improved method of median filter weighed correction is proposed. Before constructing the trend surface, water depth data is weighted corrected, the variation of the goodness of fit of the data before and after the depth correction is used as the basis for whether the water depth correction is to be carried out, the final corrected water depth data is used to calculate the coefficient vector of polynomial fitting surface and root mean square error. Greatly reducing the impact of outliers, this method has good robustness to outliers. The simulation results and real multibeam sounding data test results show that the proposed method has good filtering effect in the case of a lot of outliers, and obviously better than the traditional trend surface filtering; at the same time, this method can maintain high computational efficiency and is suitable for automatic filtering of massive multibeam echosounding data. |
| |
| Keywords: | multibeam echosounding trend surface filtering robustness median filtering goodness of fit |
| 本文献已被 CNKI 等数据库收录！ |
| 点击此处可从《海洋科学》浏览原始摘要信息 |
| 点击此处可从《海洋科学》下载免费的PDF全文 |
|
