Semi-Global Filtering of Airborne LiDAR Data for Fast Extraction of Digital Terrain Models

Hu, Xiangyun; Ye, Lizhi; Pang, Shiyan; Shan, Jie

doi:10.3390/rs70810996

Cited by 30 publications

(32 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We compare the deep CNN model with the popular commercial software TerraSolid TerraScan, Mongus's parameter-free ground filtering algorithm in 2012 [1], SGF [10], Axelsson's algorithm, and Mongus's connected operators-based algorithm in 2014 [29] on the ISPRS benchmark dataset. TerraScan uses the TIN-based filtering method; this software produces a significantly low average total error when a set of tunable parameters of the data is processed using the algorithm.…”

Section: Results and Comparison With Other Filtering Algorithmsmentioning

confidence: 99%

See 1 more Smart Citation

Deep-Learning-Based Classification for DTM Extraction from ALS Point Cloud

2016

Self Cite

View full text Add to dashboard Cite

Airborne laser scanning (ALS) point cloud data are suitable for digital terrain model (DTM) extraction given its high accuracy in elevation. Existing filtering algorithms that eliminate non-ground points mostly depend on terrain feature assumptions or representations; these assumptions result in errors when the scene is complex. This paper proposes a new method for ground point extraction based on deep learning using deep convolutional neural networks (CNN). For every point with spatial context, the neighboring points within a window are extracted and transformed into an image. Then, the classification of a point can be treated as the classification of an image; the point-to-image transformation is carefully crafted by considering the height information in the neighborhood area. After being trained on approximately 17 million labeled ALS points, the deep CNN model can learn how a human operator recognizes a point as a ground point or not. The model performs better than typical existing algorithms in terms of error rate, indicating the significant potential of deep-learning-based methods in feature extraction from a point cloud.

show abstract

Section: Results and Comparison With Other Filtering Algorithmsmentioning

confidence: 99%

“…For instance, semi-global filtering (SGF) [10] employs a novel energy function balanced by adaptive ground saliency to adapt to steep slopes, discontinuous terrains, and complex objects. Then, the SGF uses semi-global optimization to determine labels by minimizing the energy.…”

Section: Introductionmentioning

confidence: 99%

Deep-Learning-Based Classification for DTM Extraction from ALS Point Cloud

2016

Self Cite

View full text Add to dashboard Cite

show abstract

“…One group is focused on developing advanced filtering and interpolation algorithms by fully exploiting local contextual information [9][10][11][12][13]. Specifically, an adaptive approach to employ different interpolation methods based on the complexity of local terrain was designed in Maguya et al (2013) [9].…”

Section: Introductionmentioning

confidence: 99%

“…Similarly, an adaptive threshold was employed in filtering non-ground points in Su et al (2015) [10]. A novel energy function balanced by adaptive ground saliency was used to adapt to steep slopes, discontinuous terrains, and complex objects in the filtering process to identify ground points in Hu et al (2015) [11]. A strategy based on segmentation using smoothness constraint was introduced by Zhang and Lin (2013) [12] to iteratively expand ground seed surfaces into surrounding smooth terrains as much as possible.…”

Section: Introductionmentioning

confidence: 99%

An Integrated Approach to Generating Accurate DTM from Airborne Full-Waveform LiDAR Data

Gumerov

Wang

et al. 2017

Remote Sensing

View full text Add to dashboard Cite

Abstract:In this study, full-waveform LiDAR data were exploited to detect weak returns backscattered by the bare terrain underneath vegetation canopies and thus improve the generation of a digital terrain model (DTM). Building on the methods of progressive generation of triangulation irregular network (TIN) model reported in the literature, we proposed an integrated approach where echo detection, terrain identification, and TIN generation were carried out iteratively. The proposed method was tested on a dataset collected by a Riegl LMS Q-560 scanner over a study area near Sault Ste. Marie, Ontario, Canada (46 • 33 56 N, 83 • 25 18 W). The results demonstrated that more terrain points under shrubs could be identified, and the generated DTMs exhibited more details in the terrain than those obtained using the progressive TIN method. In addition, 1275 points across this study area were surveyed on the ground and used to validate the proposed approach. The estimated elevations were shown to have a strong linear relationship with the measured ones, with R 2 values above 0.98, and the RMSEs (Root Mean Squared Errors) between them were less than 0.15 m even for areas with hilly terrains underneath vegetation canopies.

show abstract

“…Lastly, each segment is filtered with semiglobal algorithm in order to remove surface objects, which will produce the final DTM. The proposed method is similar to the semi-global filtering recently proposed by Hu et al (Hu et al, 2015). However, a locally calculated balance coefficient is added in the smooth term ( ) smooth E s of energy function proposed by Hu, which makes the energy function rely on local slope heavily, thus it's impropriate to use in DSM filtering.…”

Section: Introductionmentioning

confidence: 99%

Automatic Extraction of DTM From Low Resolution DSM by Twosteps Semi-Global Filtering

Zhang¹,

Zhang²,

Zhang³

et al. 2016

ISPRS Ann. Photogramm. Remote Sens. Spatial Inf. Sci.

View full text Add to dashboard Cite

ABSTRACT:Automatically extracting DTM from DSM or LiDAR data by distinguishing non-ground points from ground points is an important issue. Many algorithms for this issue are developed, however, most of them are targeted at processing dense LiDAR data, and lack the ability of getting DTM from low resolution DSM. This is caused by the decrease of distinction on elevation variation between steep terrains and surface objects. In this paper, a method called two-steps semi-global filtering (TSGF) is proposed to extract DTM from low resolution DSM. Firstly, the DSM slope map is calculated and smoothed by SGF (semi-global filtering), which is then binarized and used as the mask of flat terrains. Secondly, the DSM is segmented with the restriction of the flat terrains mask. Lastly, each segment is filtered with semi-global algorithm in order to remove non-ground points, which will produce the final DTM. The first SGF is based on global distribution characteristic of large slope, which distinguishes steep terrains and flat terrains. The second SGF is used to filter non-ground points on DSM within flat terrain segments. Therefore, by two steps SGF non-ground points are removed robustly, while shape of steep terrains is kept. Experiments on DSM generated by ZY3 imagery with resolution of 10-30m demonstrate the effectiveness of the proposed method.

show abstract

Semi-Global Filtering of Airborne LiDAR Data for Fast Extraction of Digital Terrain Models

Cited by 30 publications

References 34 publications

Deep-Learning-Based Classification for DTM Extraction from ALS Point Cloud

Deep-Learning-Based Classification for DTM Extraction from ALS Point Cloud

An Integrated Approach to Generating Accurate DTM from Airborne Full-Waveform LiDAR Data

Automatic Extraction of DTM From Low Resolution DSM by Twosteps Semi-Global Filtering

Contact Info

Product

Resources

About