Author Affiliations
1 School of Architecture and Urban Planning, Hunan University of Technology, Zhuzhou, Hunan 412000, China2 School of Geosciences and Information-Physics, Central South University, Changsha, Hunan 410083, Chinashow less
Fig. 1. Schematic diagram of pseudo-grid. (a) Three-dimensional display; (b) two-dimensional display
Fig. 2. Construction process of pseudo-grid
Fig. 3. Flow chart of improved filtering algorithm
Fig. 4. Filtering flow chart based on CUDA
Fig. 5. Filtering result of sample 11. (a) DSM of sample data; (b) result after filtering with slope method; (c) result after filtering with the improved method; (d) real DEM provide by ISPRS; (e) error distribution map
Fig. 6. Comparison of type II error with different algorithms
Fig. 7. Comparison of processing time with different algorithms
Fig. 8. Experimental data. (a) Original point cloud data; (b) DSM grey-scale map of after meshing
Fig. 9. Experiment results. (a) Filtering result of this method; (b) filtering result of progressive TIN filtering algorithm; (c) filtering result of slope filtering algorithm; (d) DEM after filtering
Data | Data 1 | Data 2 | Data 3 |
---|
Sensor | ALS50-Ⅱ | ALS50-Ⅱ | ALS60 | Time | 2012.4 | 2012.5 | 2012.4 | Altitude /m | 1250 | 1250 | 2000 | Number | 185673 | 267859 | 381623 | Mean point density /m-2 | 1.6 | 2.5 | 3.2 | Coverage /km2 | 0.12 | 0.17 | 0.19 |
|
Table 1. Attributes of filtering data
Number | Max grid scale /m | Slope threshold | St | Si | Sm |
---|
1 | 75 | 15 | 20 | 15 | 40 | 2 | 75 | 15 | 25 | 15 | 40 | 3 | 75 | 15 | 20 | 20 | 40 | 4 | 75 | 15 | 15 | 15 | 40 |
|
Table 2. Parameters of data filtering
Experiential data | Number | Type I error /% | Type II error /% | Gross error /% |
---|
Data 1 | 1 | 10.56 | 5.78 | 7.78 | | | | 2 | 11.23 | 5.89 | | 8.12 | 3 | 12.35 | 5.23 | | 10.11 | 4 | 14.68 | 7.43 | | 12.35 | Data 2 | 1 | 9.56 | 5.51 | 5.88 | | | | 2 | 10.45 | 6.34 | | 8.12 | 3 | 13.45 | 5.68 | | 9.34 | 4 | 15.67 | 8.67 | | 9.58 | Data 3 | 1 | 10.34 | 6.45 | 9.17 | | | | 2 | 11.57 | 7.78 | | 9.35 | 3 | 14.56 | 5.69 | | 9.09 | 4 | 15.67 | 4.93 | | 8.74 |
|
Table 3. Filtering error statistics of three groups data under different parameters
Data | Filtering method | Type I error /% | Type II error /% | Gross error /% | Efficiency /s |
---|
| Progressive TIN | 13.58 | 7.32 | 9.12 | 18.4 | Data 1 | Slope filtering | 14.78 | 8.56 | 9.56 | 11.5 | | This method | 10.56 | 5.78 | 7.78 | 2.3 | | Progressive TIN | 11.85 | 4.60 | 9.89 | 17.7 | Data 2 | Slope filtering | 13.43 | 6.78 | 10.67 | 10.4 | | This method | 9.56 | 5.51 | 5.88 | 1.3 | | Progressive TIN | 12.58 | 4.47 | 10.01 | 19.7 | Data 3 | Slope filtering | 13.45 | 7.89 | 12.56 | 11.5 | | This method | 10.34 | 6.45 | 9.17 | 0.9 |
|
Table 4. Comparison table of filtering error and efficiency of different algorithms