Fig. 1. Spatial distribution of noise. (a) Point clouds data nearby TLS; (b) vertical profile along one direction

Fig. 2. Intensity map of point cloud. (a) All point clouds; (b) point clouds after points with intensity less than 2σ removed

Fig. 3. Flow chart of point cloud denoising algorithm

Fig. 4. Denoising results for all point clouds. (a) Data 1, original data; (b)data 1, Cloud compare; (c) data 1, statistical filter & radius filter; (d) data 1, proposed method; (e) data 2, original data; (f) data 2, Cloud compare; (g) data 2, statistical filter & radius filter; (f) data 2, proposed method

Fig. 5. Denoising results for edges of partial buildings and partial electric wires from data 1. (a) Original data; (b) Cloud compare; (c) statistical filter & radius filter; (d) proposed method

Fig. 6. Denoising results for edges of partial buildings and partial electric wires from data 2. (a) Original data; (b) Cloud compare; (c) statistical filter & radius filter; (d) proposed method

Table 1. Parameters for proposed method and Cloud compare software

Table 2. Denoising accuracy of proposed method, statistical filter & radius filter and Cloud compare

Table 3. Operation efficiency and ratio of number of denoised points to original points of proposed method, statistical filter & radius filter and Cloud compare