面向地形匹配的多波束测深数据抽稀方法研究

New thinning method of multi-beam bathymetric data based on terrain matching

基于多波束测深的地形定位是水下潜器导航技术研究和发展的重点,多波束测深数据的高精度快速重采样是水下地形匹配定位的前提。传统的实时抽稀方法因对多波束测深数据模型的过分简化而效果欠佳。参考Douglas-Peucker算法和点云数据抽稀方法,采用角度-弦高联合准则对多波束每ping数据进行抽稀处理,参考导航地形图对抽稀后的多ping数据基于点云离散度进行二次抽稀处理,从而实现多波束测深数据的高精度快速抽稀处理。典型的数学仿真地形和实测多波束条带数据实验表明:文中提出的抽稀方法数据抽稀率仿真地形在85%以上,实测地形在90%以上,数据抽稀前后点云构成的曲面DEM误差在3%以内,并且算法实时性较好。

Terrain positioning based on multi-beam bathymetric data is the emphatic direction of research and development of underwa- ter vehicle navigation technology. The high accuracy and fast data thinning method is the presupposition and basis for the seabed terrain matching position technology. Since the traditional real-time data thinning methods oversimplify the multi-beam data model, they have poor performance. In this paper, a new multi-beam data thinning method has been provided, which is based on Douglas-Peucker algo- rithm and the traditional point cloud data thinning method. Each ping data of the multi-beam data is firstly thinned based on its angle and chord height, the density of the multi-ping thinned data is lessened further based on the point cloud dispersion with reference to the navigation topographic map. In the end, based on the mathematical simulation of typical terrain and the surveyed multi-beam swath da- ta, the experiments have been made to evaluate the efficiency and performance of the new thinning method. The experimental results show that the data thinning rate is more than 85% for the simulated terrain and 90% for the surveyed terrain data; the point cloud sur- face DEM error is less than 3% ; and the real-time performance is better.