Classification of ALS Point Clouds Using End-to-End Deep Learning

Winiwarter, Lukas; Mandlburger, Gottfried; Schmohl, Stefan; Pfeifer, Norbert

doi:10.1007/s41064-019-00073-0

Cited by 49 publications

(50 citation statements)

References 38 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…It is worth noting that half of the surveyed region is characterized by alpine areas (higher than 600 m a.s.l. ), particularly challenging for the classification task, as observed in (Winiwarter et al, 2019). However, our method achieved a good classification also in mountainous environments.…”

Section: Introductionmentioning

confidence: 67%

Efficient Large-Scale Airborne Lidar Data Classification via Fully Convolutional Network

Maset

Padova

Fusiello

2020

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

View full text Add to dashboard Cite

Abstract. Nowadays, we are witnessing an increasing availability of large-scale airborne LiDAR (Light Detection and Ranging) data, that greatly improve our knowledge of urban areas and natural environment. In order to extract useful information from these massive point clouds, appropriate data processing is required, including point cloud classification. In this paper we present a deep learning method to efficiently perform the classification of large-scale LiDAR data, ensuring a good trade-off between speed and accuracy. The algorithm employs the projection of the point cloud into a two-dimensional image, where every pixel stores height, intensity, and echo information of the point falling in the pixel. The image is then segmented by a Fully Convolutional Network (FCN), assigning a label to each pixel and, consequently, to the corresponding point. In particular, the proposed approach is applied to process a dataset of 7700 km2 that covers the entire Friuli Venezia Giulia region (Italy), allowing to distinguish among five classes (ground, vegetation, roof, overground and power line), with an overall accuracy of 92.9%.

show abstract

Section: Introductionmentioning

confidence: 67%

Efficient Large-Scale Airborne Lidar Data Classification via Fully Convolutional Network

Maset

Padova

Fusiello

2020

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

View full text Add to dashboard Cite

show abstract

“…Popular techniques like convolutional neural networks (CNNs) outperform most traditional classification techniques, especially in the area of pattern recognition. In the context of classification of 3D ALS point clouds, grid and voxel-based approaches based on 2D-and 3D-CNNs (Zhao et al, 2018;Schmohl & Sörgel, 2019), as well as single point based networks (Winiwarter et al, 2019), are used. While all these approaches have already proven their competitiveness compared to standard classification techniques, they still need abundant training data.…”

Section: Discussionmentioning

confidence: 99%

Archaeological Ground Point Filtering of Airborne Laser Scan Derived Point-Clouds in a Difficult Mediterranean Environment

Doneus

Mandlburger

Doneus

2020

Journal of Computer Applications in Archaeology

Self Cite

View full text Add to dashboard Cite

Digital terrain models (DTM) based on airborne laser scanning (ALS) are an important source for identifying and monitoring archaeological sites and landscapes. However, a DTM is only one of many representations of a given surface. Its accuracy and quality must conform to its purpose and are a result of several considerations and decisions along the processing chain. One of the most important factors of ALS-based DTM generation is ground point filtering, i.e., the classification of the acquired point-cloud into terrain and off-terrain points. Filtering is not straightforward. The resulting DTM is usually a compromise that might show the surface below very dense vegetation while losing detail in other areas. In this paper, we show that in very complex situations (e.g., strongly varying vegetation cover), an optimal compromise is difficult to achieve, and more than one filter with different settings adapted to the varying degree of vegetation cover is necessary. For practical reasons, the results need to be combined into a single DTM. This is demonstrated using the case study of a Mediterranean landscape in Croatia, which consists of open areas (agricultural and grassland), olive plantations, as well as extremely dense and evergreen macchia vegetation. The results are the first step toward an adaptive ground point filtering strategy that might be useful far beyond the field of archaeology.

show abstract

“…Obviously, point cloud transformation requires more computational effort and causes information loss, which hinders accurate classification. Deep learning can also work directly on the raw point cloud or graphs [30,31]. Qi et al [32] propose PointNet for the classification and recognition of point clouds.…”

Section: Point Cloud Classificationmentioning

confidence: 99%

An Object-Based Bidirectional Method for Integrated Building Extraction and Change Detection between Multimodal Point Clouds

2020

View full text Add to dashboard Cite

Building extraction and change detection are two important tasks in the remote sensing domain. Change detection between airborne laser scanning data and photogrammetric data is vulnerable to dense matching errors, mis-alignment errors and data gaps. This paper proposes an unsupervised object-based method for integrated building extraction and change detection. Firstly, terrain, roofs and vegetation are extracted from the precise laser point cloud, based on “bottom-up” segmentation and clustering. Secondly, change detection is performed in an object-based bidirectional manner: Heightened buildings and demolished buildings are detected by taking the laser scanning data as reference, while newly-built buildings are detected by taking the dense matching data as reference. Experiments on two urban data sets demonstrate its effectiveness and robustness. The object-based change detection achieves a recall rate of 92.31% and a precision rate of 88.89% for the Rotterdam dataset; it achieves a recall rate of 85.71% and a precision rate of 100% for the Enschede dataset. It can not only extract unchanged building footprints, but also assign heightened or demolished labels to the changed buildings.

show abstract

Classification of ALS Point Clouds Using End-to-End Deep Learning

Cited by 49 publications

References 38 publications

Efficient Large-Scale Airborne Lidar Data Classification via Fully Convolutional Network

Efficient Large-Scale Airborne Lidar Data Classification via Fully Convolutional Network

Archaeological Ground Point Filtering of Airborne Laser Scan Derived Point-Clouds in a Difficult Mediterranean Environment

An Object-Based Bidirectional Method for Integrated Building Extraction and Change Detection between Multimodal Point Clouds

Contact Info

Product

Resources

About