Abstract:Filtering is one of the core post-processing steps for airborne LiDAR point cloud. In recent years, the morphology-based filtering algorithms have proven to be a powerful and efficient tool for filtering airborne LiDAR point cloud. However, most traditional morphology-based algorithms have difficulties in preserving abrupt terrain features, especially when using larger filtering windows. In order to suppress the omission error caused by protruding terrain features, this paper proposes an improved morphological algorithm based on multi-level kriging interpolation. This algorithm is essentially a combination of progressive morphological filtering algorithm and multi-level interpolation filtering algorithm. The morphological opening operation is performed with filtering window gradually downsizing, while kriging interpolation is conducted at different levels according to the different filtering windows. This process is iterative in a top to down fashion until the filtering window is no longer greater than the preset minimum filtering window. Fifteen samples provided by the ISPRS commission were chosen to test the performance of the proposed algorithm. Experimental results show that the proposed method can achieve promising results not only in flat urban areas but also in rural areas. Comparing with other eight classical filtering methods, the proposed method obtained the lowest omission error, and preserved protruding terrain features better.

No abstract

No abstract

The standard forward transformation equation plays a major role in coordinate transformation between global and local datums. Thus, it is a prerequisite step in the forward conversion of geodetic coordinates into cartesian coordinates in coordinate transformation from global to local datum and vice versa. Numerous studies have been carried out on converting cartesian coordinates to geodetic coordinates (reverse procedure) through the application of iterative, approximate, closed form, vector-based and computational intelligence algorithms. However, based on literature covered pertaining to this study, it was realized that the existing researches do not fully address the issue of applying and testing alternative techniques in the case of the forward conversion. Hence, the purpose of this present study was to explore the coordinate conversion performance of two different artificial neural network approaches (backpropagation artificial neural network (BPANN) and radial basis function neural network (RBFNN)) and multiple linear regression (MLR). The statistical findings revealed that the BPANN, RBFNN and MLR offered satisfactory prediction of cartesian coordinates. However, the RBFNN compared to BPANN and MLR showed better stability and more accurate prediction results. Furthermore, in terms of maximum three-dimensional position error, the RBFNN attained 0.004 m while 0.011 and 0.627 m were achieved, respectively, by MLR and BPANN. By virtue of the success achieved in this study, the main conclusion drawn here is that RBFNN provides a promising alter-B Hu Youjian 123 Math Geosci native in the forward conversion of geodetic coordinates into cartesian coordinates. Therefore, the capability of artificial neural network as a powerful tool for solving majority of function approximation problems in geodesy has been demonstrated.