Classification of Airborne Laser Scanning Data Using Geometric  Multi-Scale Features and Different Neighbourhood Types

Blomley, Rosmarie; Jutzi, Boris; Weinmann, Martin

doi:10.5194/isprs-annals-iii-3-169-2016

Cited by 34 publications

(19 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Previous studies used 3D labelled LiDAR data benchmarks (Blomley et al, 2016;Vosselman et al, 2017) or 2D labelled LiDAR data benchmarks that were extended to 3D labelled points (Niemeyer et al, 2014), which were not available for the tested study area. Also, the 3D reference points could be provided by the supplier or manually labeled (Xu et al, 2014); however this is time consuming and might introduce errors from the human operator or analyst.…”

Section: Evaluation Of Classification Resultsmentioning

confidence: 99%

“…In 3D point classification of the airborne LiDAR data, multi-class labeling has become an essential topic for 3D city modeling, change detection, map updating, disaster evaluation and emergency purposes. However, most recent studies focus on the geometric characteristics described by the LiDAR data (Mallet et al, 2011;Xu et al, 2014;Blomley et al, 2016;Vosselman et al, 2017). Studies rarely reported using the intensity LiDAR data along with the geometric features extracted from the LiDAR data for the purpose of urban areas classification.…”

Section: Acknowledgmentsmentioning

confidence: 99%

“…Blomley et al (2016) extracted geometric features from the LiDAR point clouds based on multiple scales and different neighbourhood types. Then, Random Forest (RF) classifier was applied to classify the terrain into ground, buildings, cars, trees and low vegetation.…”

mentioning

confidence: 99%

See 2 more Smart Citations

Land/Water Discrimination and Land Cover Classification Using Multispectral Airborne LiDAR Data

Morsy¹

2023

Preprint

View full text Add to dashboard Cite

Multispectral airborne Light Detection And Ranging (LiDAR) systems are currently available. Optech Titan is an example of these systems, which acquires LiDAR point clouds at three independent wavelengths (1550, 1064 and 532 nm) from Earth’s surface. This dissertation aims to use the radiometric information (i.e., intensity) of the Optech Titan LiDAR data along with the geometric information (e.g., height) for land/water discrimination in coastal zones and land cover classification of urban areas. A set of point features based on elevation, intensity, and geometry was extracted and evaluated for land/water discrimination in coastal zones. In addition, an automated land/water discrimination approach based on seeded region growing algorithm was presented. Two data subsets were tested at Lake Ontario and Tobermory Harbour in Ontario, Canada. The elevation and geometry-based features achieved average overall accuracies of 72.8% - 83.3% and 69.9% -74.4%, respectively, while the intensity-based features achieved an average overall accuracy of 59.0% - 63.4%. The region growing method achieved an average overall accuracy of more than 99%, and the automation of this method is restricted by having double returns from water bodies at the 532 nm wavelength. A hierarchal point-based classification approach was presented for land cover classification of urban areas. The collected point clouds at the three wavelengths were first merged and three intensity values were estimated for each LiDAR point, followed by three-level classification approach. First, a ground filtering method was applied to separate non-ground from ground points. Second, three normalized difference vegetation indices (NDVIs) were computed, followed by NDVIs’ histograms construction. A multivariate Gaussian decomposition (MVGD) was then used to divide those histograms into buildings or trees from non-ground and roads or grass from ground points. Third, classes such as power lines, swimming pools and different types of trees were labeled based on their spectral characteristics. Three data subsets were tested representing different complexity of urban areas in Oshawa, Ontario, Canada. It is shown that the presented approach has achieved an overall accuracy up to 93.0%, which increased to more than 99% by considering the spatial coherence of the LiDAR point clouds.

show abstract

Section: Evaluation Of Classification Resultsmentioning

confidence: 99%

Section: Acknowledgmentsmentioning

confidence: 99%

See 1 more Smart Citation

Land/Water Discrimination and Land Cover Classification Using Multispectral Airborne LiDAR Data

Morsy¹

2023

Preprint

View full text Add to dashboard Cite

show abstract

“…For the experiment, a commonly structured workflow was used (e.g. Weinmann et al, 2014;Blomley et al, 2016), consisting of four components: neighbourhood definition (Section 2.1), feature extraction (Section 2.2), selection of relevant features (Section 2.3) and classification (Section 2.4).…”

Section: Methodsmentioning

confidence: 99%

Semantic Classification of Sandstone Landscape Point Cloud Based on Neighbourhood Features

Tomková

Lysák

Potůčková

2020

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

View full text Add to dashboard Cite

Abstract. The technology of airborne laser scanning enables fast and accurate gathering spatial data containing also echoes from the terrain below the vegetation canopy that is beneficial for topographic mapping of wooded sandstone landscapes in Czechia, Poland, and Germany. The challengeable task is to determine the ground points in the point cloud because commonly used filtration methods do not successfully distinguish between vegetation and rock pillars and faces. In this paper, we replace filtration with classification approach using the features derived from characteristics of points within a neighbourhood of optimized sizes, such as eigenvalue-based features and echo ratio. Random Forest classifier is trained and tested on the manually labelled dataset with a density of almost 650 points/m2 from the Adršpach-Teplice Rocks. The overall accuracy reaches 87% but recall and precision of non-ground points are unsatisfactory. Misclassified non-ground points are located also within trees, thus we do not consider the result as suitable for DTM processing without further processing.

show abstract

“…(iii) The scale of the neighbourhood selected for each point depends on the properties of the neighbourhood. The scale can be selected based on eigenentropy (de Blomley et al, 2016) (Blomley et al, 2016). This strategy allows to generalise from a specific data set, but still suffers from the disability to capture context from different scales, like for example a car has tires (scale less than 1m) and appears almost always on the road (scale of several meters).…”

Section: Related Workmentioning

confidence: 99%

Feature-Extraction From All-Scale Neighborhoods With Applications to Semantic Segmentation of Point Clouds

Leichter

Werner

Sester

2020

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

View full text Add to dashboard Cite

Abstract. Feature extraction from a range of scales is crucial for successful classification of objects of different size in 3D point clouds with varying point density. 3D point clouds have high relevance in application areas such as terrain modelling, building modelling or autonomous driving. A large amount of such data is available but also that these data is subject to investigation in the context of different tasks like segmentation, classification, simultaneous localisation and mapping and others. In this paper, we introduce a novel multiscale approach to recover neighbourhood in unstructured 3D point clouds. Unlike the typical strategy of defining one single scale for the whole dataset or use a single optimised scale for every point, we consider an interval of scales. In this initial work our primary goal is to evaluate the information gain through the usage of the multiscale neighbourhood definition for the calculation of shape features, which are used for point classification. Therefore, we show and discuss empirical results from the application of classical classification models to multiscale features. The unstructured nature of 3D point cloud makes it necessary to recover neighbourhood information before meaningful features can be extracted. This paper proposes the extraction of geometrical features from a range of neighbourhood with different scales, i.e. neighborhood ranges. We investigate the utilisation of the large set of features in combination with feature aggregation/selection algorithms and classical machine learning techniques. We show that the all-scale-approach outperform single scale approaches as well as the approach with an optimised per point selected scale.

show abstract

Classification of Airborne Laser Scanning Data Using Geometric Multi-Scale Features and Different Neighbourhood Types

Cited by 34 publications

References 39 publications

Land/Water Discrimination and Land Cover Classification Using Multispectral Airborne LiDAR Data

Land/Water Discrimination and Land Cover Classification Using Multispectral Airborne LiDAR Data

Semantic Classification of Sandstone Landscape Point Cloud Based on Neighbourhood Features

Feature-Extraction From All-Scale Neighborhoods With Applications to Semantic Segmentation of Point Clouds

Contact Info

Product

Resources

About