Implicit reconstructions of thin leaf surfaces from large, noisy point clouds

Whebell, Riley M.; Moroney, Timothy J.; Turner, Ian; Pethiyagoda, Ravindra; McCue, Scott W.

doi:10.1016/j.apm.2021.05.014

Cited by 4 publications

(4 citation statements)

References 29 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We used the values of the radius of the curvature of the pipe relative to the size of its diameter as a correction factor that increases with each step for the location of each cloud point in the Cartesian coordinate system. Then, the reconstruction of the surface mesh [45], with respect to the coordinates of the point cloud, is assumed. The software implementation of this procedure was performed through a convolutional neural network with an integrated voxelization graph (LV-GCNN) [46], which allows reconstructing the surface without visible losses.…”

Section: Methodsmentioning

confidence: 99%

Machine Learning and 3D Reconstruction of Materials Surface for Nondestructive Inspection

Kartashov

Чернов

Alexandrov

et al. 2022

Sensors

View full text Add to dashboard Cite

During the steel pipeline installation, special attention is paid to the butt weld control performed by fusion welding. The operation of the currently popular automated X-ray and ultrasonic testing complexes is associated with high resource and monetary costs. In this regard, this work is devoted to the development of alternative and cost-effective means of preliminary quality control of the work performed based on the visual testing method. To achieve this goal, a hardware platform based on a single board Raspberry Pi4 minicomputer and a set of available modules and expansion cards is proposed, and software whose main functionality is implemented based on the systemic application of computer vision algorithms and machine learning methods. The YOLOv5 object detection algorithm and the random forest machine learning model were used as a defect detection and classification system. The mean average precision (mAP) of the trained YOLOv5 algorithm based on extracted weld contours is 86.9%. A copy of YOLOv5 trained on the images of control objects showed a mAP result of 96.8%. Random forest identifying of the defect precursor based on the point clouds of the weld surface achieved a mAP of 87.5%.

show abstract

Section: Methodsmentioning

confidence: 99%

Machine Learning and 3D Reconstruction of Materials Surface for Nondestructive Inspection

Kartashov

Чернов

Alexandrov

et al. 2022

Sensors

View full text Add to dashboard Cite

show abstract

“…Emphasizing the real contours and details of the trees, the Delaunay triangulation method is selected for street tree modeling [30]. The Delaunay triangulation method follows the criteria of the circumcircle being empty and maximizing the minimum internal angle [31], avoiding the generation of elongated triangles. The resulting street tree model not only significantly reduces data volume but also has better versatility in the standard obj model format.…”

Section: Street Trees Modelingmentioning

confidence: 99%

“…e Sens. 2023, 15, x FOR PEER REVIEW 6 of 20 launay triangulation method follows the criteria of the circumcircle being empty and maximizing the minimum internal angle [31], avoiding the generation of elongated triangles.…”

Section: Street Trees Modelingmentioning

confidence: 99%

See 1 more Smart Citation

Analysis of Light Obstruction from Street Lighting in Road Scenes

Ren,

Li,

Liu

et al. 2023

Remote Sensing

View full text Add to dashboard Cite

As urban greenery improves and the ecological environment is continuously optimized, road facilities are also impacted to varying degrees. For example, as vegetation grows, it causes varying degrees of obstruction to the lighting facilities on the roads. This article is based on vehicle-mounted LiDAR data and focuses on the point cloud data characteristics of different objects. Using appropriate modeling techniques, it accurately models road surfaces, green belts, streetlights, and other objects. On the Lumion platform, this system creates a 3D visualization of road scenes and examines the interplay between objects and lighting space, analyzing lit areas. Leveraging the precise 3D spatial relationships found in point clouds, it determines the effective illumination area on the ground from streetlights after object obstruction, comparing it to the theoretical illumination area. This not only visualizes the road scene but also quantifies the lighting obstruction rate. Furthermore, it assesses the lighting conditions in road scenes based on illuminance distribution, offering scientific insights and suggestions for enhancing road lighting.

show abstract