FM
Fancong Meng
info
Please Note
<p>This page displays the records of the person named above and is not linked to a unique person identifier. This record may need to be merged to a profile.</p>
2 records found
1
Mobile laser scanning (MLS) systems are often used to efficiently acquire reference data covering a large-scale scene. The terrestrial laser scanner (TLS) can easily collect high point density data of local scene. Localization of static TLS scans in mobile mapping point clouds can afford detailed geographic information for many specific tasks especially in autonomous driving and robotics. However, large-scale MLS reference data often have a huge amount of data and many similar scene data; significant differences may exist between MLS and TLS data. To overcome these challenges, this paper presents a novel deep neural network-based localization method in urban environment, divided by place recognition and pose refinement. Firstly, simple, reliable primitives, cylinder-like features were extracted to describe the global features of a local urban scene. Then, a probabilistic framework is applied to estimate a similarity between TLS and MLS data, under a stable decision-making strategy. Based on the results of a place recognition, we design a patch-based convolution neural network (CNN) (point-based CNN is used as kernel) for pose refinement. The input data unit is the batch consisting of several patches. One patch goes through three main blocks: feature extraction block (FEB), the patch correspondence search block and the pose estimation block. Finally, a global refinement was proposed to tune the predicted transformation parameters to realize localization. The research aim is to find the most similar scene of MLS reference data compared with the local TLS scan, and accurately estimate the transformation matrix between them. To evaluate the performance, comprehensive experiments were carried out. The experiments demonstrate that the proposed method has good performance in terms of efficiency, i.e., the runtime of processing a million points is 5 s, robustness, i.e., the success rate of place recognition is 100% in the experiments, accuracy, i.e., the mean rotation and translation error is (0.24 deg, 0.88 m) and (0.03 deg, 0.06 m) on TU Delft campus and Shanghai urban datasets, respectively, and outperformed some commonly used methods (e.g., iterative closest point (ICP), coherent point drift (CPD), random sample consensus (RANSAC)-based method).
...
Mobile laser scanning (MLS) systems are often used to efficiently acquire reference data covering a large-scale scene. The terrestrial laser scanner (TLS) can easily collect high point density data of local scene. Localization of static TLS scans in mobile mapping point clouds can afford detailed geographic information for many specific tasks especially in autonomous driving and robotics. However, large-scale MLS reference data often have a huge amount of data and many similar scene data; significant differences may exist between MLS and TLS data. To overcome these challenges, this paper presents a novel deep neural network-based localization method in urban environment, divided by place recognition and pose refinement. Firstly, simple, reliable primitives, cylinder-like features were extracted to describe the global features of a local urban scene. Then, a probabilistic framework is applied to estimate a similarity between TLS and MLS data, under a stable decision-making strategy. Based on the results of a place recognition, we design a patch-based convolution neural network (CNN) (point-based CNN is used as kernel) for pose refinement. The input data unit is the batch consisting of several patches. One patch goes through three main blocks: feature extraction block (FEB), the patch correspondence search block and the pose estimation block. Finally, a global refinement was proposed to tune the predicted transformation parameters to realize localization. The research aim is to find the most similar scene of MLS reference data compared with the local TLS scan, and accurately estimate the transformation matrix between them. To evaluate the performance, comprehensive experiments were carried out. The experiments demonstrate that the proposed method has good performance in terms of efficiency, i.e., the runtime of processing a million points is 5 s, robustness, i.e., the success rate of place recognition is 100% in the experiments, accuracy, i.e., the mean rotation and translation error is (0.24 deg, 0.88 m) and (0.03 deg, 0.06 m) on TU Delft campus and Shanghai urban datasets, respectively, and outperformed some commonly used methods (e.g., iterative closest point (ICP), coherent point drift (CPD), random sample consensus (RANSAC)-based method).
Heritage documentation is implemented by digitally recording historical artifacts for the conservation and protection of these cultural heritage objects. As efficient spatial data acquisition tools, laser scanners have been widely used to collect highly accurate three-dimensional (3D) point clouds without damaging the original structure and the environment. To ensure the integrity and quality of the collected data, field inspection (i.e., on-spot checking the data quality) should be carried out to determine the need for additional measurements (i.e., extra laser scanning for areas with quality issues such as data missing and quality degradation). To facilitate inspection of all collected point clouds, especially checking the quality issues in overlaps between adjacent scans, all scans should be registered together. Thus, a point cloud registration method that is able to register scans fast and robustly is required. To fulfill the aim, this study proposes an efficient probabilistic registration for free-form cultural heritage objects by integrating the proposed principal direction descriptor and curve constraints. We developed a novel shape descriptor based on a local frame of principal directions. Within the frame, its density and distance feature images were generated to describe the shape of the local surface. We then embedded the descriptor into a probabilistic framework to reject ambiguous matches. Spatial curves were integrated as constraints to delimit the solution space. Finally, a multi-view registration was used to refine the position and orientation of each scan for the field inspection. Comprehensive experiments show that the proposed method was able to perform well in terms of rotation error, translation error, robustness, and runtime and outperformed some commonly used approaches.
...
Heritage documentation is implemented by digitally recording historical artifacts for the conservation and protection of these cultural heritage objects. As efficient spatial data acquisition tools, laser scanners have been widely used to collect highly accurate three-dimensional (3D) point clouds without damaging the original structure and the environment. To ensure the integrity and quality of the collected data, field inspection (i.e., on-spot checking the data quality) should be carried out to determine the need for additional measurements (i.e., extra laser scanning for areas with quality issues such as data missing and quality degradation). To facilitate inspection of all collected point clouds, especially checking the quality issues in overlaps between adjacent scans, all scans should be registered together. Thus, a point cloud registration method that is able to register scans fast and robustly is required. To fulfill the aim, this study proposes an efficient probabilistic registration for free-form cultural heritage objects by integrating the proposed principal direction descriptor and curve constraints. We developed a novel shape descriptor based on a local frame of principal directions. Within the frame, its density and distance feature images were generated to describe the shape of the local surface. We then embedded the descriptor into a probabilistic framework to reject ambiguous matches. Spatial curves were integrated as constraints to delimit the solution space. Finally, a multi-view registration was used to refine the position and orientation of each scan for the field inspection. Comprehensive experiments show that the proposed method was able to perform well in terms of rotation error, translation error, robustness, and runtime and outperformed some commonly used approaches.