Quality-based registration refinement of airborne LiDAR and photogrammetric point clouds

Toschi, I.; Farella, Elisa Mariarosaria; Welponer, M.; Remondino, Fabio

doi:10.1016/j.isprsjprs.2020.12.005

Cited by 14 publications

(10 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The fusion of ALS and photogrammetry-derived DIM point clouds can only be performed efficiently if they are registered precisely to eliminate the geometric inconsistency between the two different types of data (Peng et al, 2019;Toschi et al, 2021;Yang et al, 2021). This registration process entails identifying the rigid or non-rigid transformation that best aligns the points in one point cloud with the corresponding points in another point cloud (Li et al, 2021).…”

Section: Point Cloud Fusionmentioning

confidence: 99%

“…When the photogrammetric imagery does not contain georeferencing information or ground control points (GCP), a few corresponding ALS point clouds were used as GCP information within the photogrammetric bundle block adjustment (BBA). This step was carried out because the quality of combining multi-source point cloud depends on inter-dataset registration and with regards to the LiDAR and DIM data processing chain, this registration can be categorized into four different groups (Toschi et al, 2021): (i) separate ALS stripe adjustment (SA) and bundle block adjustment (BBA), (ii) within the photogrammetric BBA use of ALS point clouds as ground control information (Yang et al, 2021), (iii) establish the transformation by using standard features (e.g., edges, boundaries, corners and 3D planar surfaces) (Peng et al, 2019), (iv) a single hybrid adjustment process with an integrated SA and BBA (Haala et al, 2020;Glira et al, 2019). Therefore, following the second category, a few corresponding ALS point clouds were used as GCP for DIM generation.…”

Section: Dim Datamentioning

confidence: 99%

“…This section demonstrates and evaluates the proposed point cloud fusion algorithm using airborne LiDAR and DIM data- sets. Due to the superior accuracy of LiDAR sensors over airborne photogrammetry (Zhang et al, 2019;Leslar, 2015;Toschi et al, 2021), LiDAR point clouds were used as the target points and DIM as the source. This choice ensured that the fusion algorithm could achieve a high level of precision and reliability in the resulting data.…”

Section: Experiments and Evaluationmentioning

confidence: 99%

See 2 more Smart Citations

Airborne Cross-Source Point Clouds Fusion by Slice-to-Slice Adjustment

Parvaz,

Teferle,

Nurunnabi

2024

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

View full text Add to dashboard Cite

Abstract. Point cloud fusion is a process plays pivotal role in geospatial data analysis that aims to integrate data from multiple sources to create a comprehensive and precise representation of the environment. Integrating point clouds acquired from cross-source or hybrid sensors presents unique challenges due to differences in geometric accuracy, precision, and the size of data gaps, along with variations in available attributes. Significant progress has been made in developing algorithms and methods to address these challenges, but the problems are not sufficiently resolved and remain one of the most challenging aspects of geospatial data processing. In this paper, we present a new approach for airborne cross-source point cloud fusion through a slice-to-slice adjustment. Our method generates cross-sectional slices and aligns them following some sequential steps. This approach enhances the accuracy and completeness of the fused point cloud, overcoming issues related to geometric disparities and data gaps. Experimental results demonstrate the effectiveness of our approach in improving registration accuracy, preserving geometric detail, and providing valuable insights for utilizing the potentials of both data sources.

show abstract

Section: Point Cloud Fusionmentioning

confidence: 99%

Section: Dim Datamentioning

confidence: 99%

Section: Experiments and Evaluationmentioning

confidence: 99%

See 1 more Smart Citation

Airborne Cross-Source Point Clouds Fusion by Slice-to-Slice Adjustment

Parvaz,

Teferle,

Nurunnabi

2024

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

View full text Add to dashboard Cite

show abstract

“…• Large scale urban point cloud (700 m x 700 m), derived with a hybrid aerial sensor over the city centre of Bordeaux (Toschi et al, 2021), including four classes (street, facades, roof, vegetation). • Large scale urban point cloud (220 m × 200 m), belonging to the Dublin dataset (Zolanvari et al, 2019), acquired with a LiDAR sensor (without colour information) and featuring five classes (street, vegetation, façade, window, roof).…”

Section: Considered Scenariosmentioning

confidence: 99%

Unsupervised Object-Based Clustering in Support of Supervised Point-Based 3d Point Cloud Classification

Grilli

Poux

Remondino

2021

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

Self Cite

View full text Add to dashboard Cite

Abstract. The number of approaches available for semantic segmentation of point clouds has grown exponentially in recent years. The availability of numerous annotated datasets has resulted in the emergence of deep learning approaches with increasingly promising outcomes. Even if successful, the implementation of such algorithms requires operators with a high level of expertise, large quantities of annotated data and high-performance computers. On the contrary, the purpose of this study is to develop a fast, light and user-friendly classification approach valid from urban to indoor or heritage scenarios. To this aim, an unsupervised object-based clustering approach is used to assist and improve a feature-based classification approach based on a standard machine learning predictive model. Results achieved over four different large scenarios demonstrate the possibility to develop a reliable, accurate and flexible approach based on a limited number of features and very few annotated data.

show abstract

“…gender scoring [6], statistical shape modeling [7], computer vision [8], multimedia applications [9], human-computer interactions [10], 3D deformation of the human spinal column detection [11], image face alignment [12], and 3D human body analysis [13], the non-rigid registration is a non-trivial and ill-defined problem with a high number of degrees-offreedom (DOFs). Accordingly, there are many challenges for preserving features of the source surface in the design and implementation of a non-rigid ICP registration algorithm [14], where the features account for salient geometric features which form compound higher-level descriptors. A salient geometric feature, or in short, a salient feature, consists of a cluster of descriptors that locally describe a nontrivial region of the surface [15] i.e.…”

Section: Introductionmentioning

confidence: 99%