Automatic DTM extraction from airborne LiDAR based on expectation-maximization

Hui, Zhenyang; Li, Dajun; Jin, Shuanggen; Ziggah, Yao Yevenyo; Wang, Leyang; Hu, Yincui

doi:10.1016/j.optlastec.2018.10.051

Cited by 40 publications

(20 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…POT was further refined with 1 • increments from POT-2 • to POT+2 • . Figure 14 shows the accuracies of the proposed algorithm under CR = 0.5 m against other 31 algorithms [2, [24][25][26][27][31][32][33][34]37,38,[41][42][43][44]47,[54][55][56][57][58][59][60][61]. The average total error of the proposed algorithm under CR = 0.5 m provides competitive performance, below four filtering algorithms, but above 27 other algorithms.…”

Section: Testing With Isprs Datasetmentioning

confidence: 97%

Filtering Airborne LiDAR Data Through Complementary Cloth Simulation and Progressive TIN Densification Filters

et al. 2019

View full text Add to dashboard Cite

Separating point clouds into ground and non-ground points is a preliminary and essential step in various applications of airborne light detection and ranging (LiDAR) data, and many filtering algorithms have been proposed to automatically filter ground points. Among them, the progressive triangulated irregular network (TIN) densification filtering (PTDF) algorithm is widely employed due to its robustness and effectiveness. However, the performance of this algorithm usually depends on the detailed initial terrain and the cautious tuning of parameters to cope with various terrains. Consequently, many approaches have been proposed to provide as much detailed initial terrain as possible. However, most of them require many user-defined parameters. Moreover, these parameters are difficult to determine for users. Recently, the cloth simulation filtering (CSF) algorithm has gradually drawn attention because its parameters are few and easy-to-set. CSF can obtain a fine initial terrain, which simultaneously provides a good foundation for parameter threshold estimation of progressive TIN densification (PTD). However, it easily causes misclassification when further refining the initial terrain. To achieve the complementary advantages of CSF and PTDF, a novel filtering algorithm that combines cloth simulation (CS) and PTD is proposed in this study. In the proposed algorithm, a high-quality initial provisional digital terrain model (DTM) is obtained by CS, and the parameter thresholds of PTD are estimated from the initial provisional DTM based on statistical analysis theory. Finally, PTD with adaptive parameter thresholds is used to refine the initial provisional DTM. These contributions of the implementation details achieve accuracy enhancement and resilience to parameter tuning. The experimental results indicate that the proposed algorithm improves performance over their direct predecessors. Furthermore, compared with the publicized improved PTDF algorithms, our algorithm is not only superior in accuracy but also practicality. The fact that the proposed algorithm is of high accuracy and easy-to-use is desirable for users.

show abstract

Section: Testing With Isprs Datasetmentioning

confidence: 97%

Filtering Airborne LiDAR Data Through Complementary Cloth Simulation and Progressive TIN Densification Filters

et al. 2019

View full text Add to dashboard Cite

show abstract

“…Crosilla et al [29] employed the sequential skewness and kurtosis analysis of elevation and intensity point distribution values to filter and classify point clouds. Taking the filtering as a separation of mixed Gaussian models, Hui et al [30] proposed a threshold-free algorithm based on expectationmaximization. However, the statistical-based filters are prone to classify the non-ground points as the ground points when the number of the former are greater than that of the latter [28].…”

Section: Related Workmentioning

confidence: 99%

Multi-Level Interpolation-Based Filter for Airborne LiDAR Point Clouds in Forested Areas

et al. 2020

View full text Add to dashboard Cite

Over the past decades, plenty of filtering algorithms have been presented to distinguish ground and non-ground points from airborne LiDAR point clouds. However, with the existing methods, it is difficult to derive satisfactory filtering results on rugged terrains with dense vegetation due to the low-level penetration ability of laser pulses. Therefore, a multi-level interpolation-based filter is developed in this paper. The novelty of the algorithm lies within its usage of multi-scale morphological operations and robust z-score to correctly select ground seeds as more as possible, and a terrain-adaptive residual threshold to adapt to various terrain characteristics. Rural samples provided by International Society for Photogrammetry and Remote Sensing (ISPRS) were employed to assess the performance of the proposed method and its results were compared with 15 filtering algorithms developed in recent 5 years (2015-2019). Results show that the proposed method with the optimized parameters produces the best accuracy with the average total error and kappa coefficient of 1.89% and 87.88%, respectively. We further filtered high-density point clouds in six forested areas with different vegetation covers and terrain slopes. Results demonstrate that the proposed algorithm is more accurate than the well-known filtering methods including morphological-based filter, progressive TIN densification filter (PTD), improved PTD and cloth simulation filter, with the average total error decreased by 26.2%, 19.9%, 3.8% and 40.4%, respectively. Moreover, the DEMs of the proposed method have lower average root mean square errors than the four classical filters. Therefore, the proposed method can be considered as an effective ground filtering algorithm for airborne LiDAR point clouds in forested areas. INDEX TERMS Filtering, point cloud, interpolation, digital elevation model.

show abstract

“…However, details of the experiments and parameter settings are not always given and results for a same method on a same dataset can greatly vary. For example, for progressive triangulation on ISPRS dataset 1, (Bigdeli et al, 2018) found an error of 10.21% while (Hui et al, 2019) obtained an error of 28.21%. Indeed, results are sensitive to parameter settings and parameters are usually adjusted according to different variables such as the point density and the type of terrain.…”

Section: Geometrical Approachesmentioning

confidence: 99%

Ground Point Filtering From Airborne Lidar Point Clouds Using Deep Learning: A Preliminary Study

Janssens-Coron

Guilbert

2019

Int. Arch. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

<p><strong>Abstract.</strong> Airborne lidar data is commonly used to generate point clouds over large areas. These points can be classified into different categories such as ground, building, vegetation, etc. The first step for this is to separate ground points from non-ground points. Existing methods rely mainly on TIN densification but there performance varies with the type of terrain and relies on the user’s experience who adjusts parameters accordingly. An alternative may be on the use of a deep learning approach that would limit user’s intervention. Hence, in this paper, we assess a deep learning architecture, PointNet, that applies directly to point clouds. Our preliminary results show mitigating classification rates and further investigation is required to properly train the system and improve the robustness, showing issues with the choices we made in the preprocessing. Nonetheless, our analysis suggests that it is necessary to enrich the architecture of the network to integrate the notion of neighbourhood at different scales in order to increase the accuracy and the robustness of the treatment as well as its capacity to treat data from different geographical contexts.</p>

show abstract

Automatic DTM extraction from airborne LiDAR based on expectation-maximization

Cited by 40 publications

References 32 publications

Filtering Airborne LiDAR Data Through Complementary Cloth Simulation and Progressive TIN Densification Filters

Filtering Airborne LiDAR Data Through Complementary Cloth Simulation and Progressive TIN Densification Filters

Multi-Level Interpolation-Based Filter for Airborne LiDAR Point Clouds in Forested Areas

Ground Point Filtering From Airborne Lidar Point Clouds Using Deep Learning: A Preliminary Study

Contact Info

Product

Resources

About