结合多波束点云强度和高程信息滤波算法研究

Study on Filtering Algorithm Combining Multibeam Point Cloud Intensity and Elevation Information

针对含有鱼群或水草等生物水域精确地形获取困难,存在大量贴近地表的噪声点且难以滤除等问题,提出一种结合多波束点云数据强度和高程信息的滤波算法。首先根据多波束原始点云数据的强度信息对数据进行初步分类,然后利用高程四分法选取地形种子点,最后根据种子点与周围点云数据的强度和高程信息,实现数据的准确分类,达到滤波效果。试验结果与形态学滤波算法、自适应坡度滤波算法对比分析,结果表明所提算法在含有植被和鱼群等生物水域滤波效果明显,Ⅰ类、Ⅱ类误差降低了6%和10%左右,总误差降低了7%左右,验证了所提算法的有效性和可行性。

Aiming at the difficulty of obtaining accurate terrain of biological waters including fish schools or water plants,there are a lot of noise points close to the surface and it is difficult to filter out,a filtering algorithm combining multibeam point cloud data intensity and elevation information is proposed.Firstly,the data is initially classified according to the intensity information of the original multibeam point cloud data,then the terrain seed points are selected using the elevation quartile method,and finally,the accurate classification of the data is achieved according to the strength and elevation information of the seed points and the surrounding point cloud data.Filtering effect.The test results are compared with the morphological filtering algorithm and the adaptive gradient filtering algorithm.The results show that the proposed algorithm has obvious filtering effect in biological waters containing vegetation and fish,and the errors of Types Ⅰand Ⅱ are reduced by about 6%and 10%.The error is reduced by about 7%,which verifies the effectiveness and feasibility of the proposed algorithm.