Abstract:Three-dimensional (3D) line-laser scanning is a widely used 3D reconstruction technique in the industry. As a key procedure of 3D line-laser scanning, centerline extraction of laser stripes directly determines the accuracy of reconstructed 3D models. Because of the noise inside laser stripes, centerline extraction methods based on the gray distribution may provide biased results. In order to address this problem, a Voronoi-diagram-based method (VM) for centerline extraction, which can extract centerlines accurately under severe noises, is proposed. To solve the emerging problems when the Voronoi diagram is applied to line-laser stripes, a fast pruning algorithm based on the distribution of graph centrality is proposed, and two centerline extension algorithms based on least square fitting are developed. The experiments are performed on synthetic images and a line-laser 3D scanner to evaluate the method's accuracy, robustness, and efficiency. The VM method is proved to have better accuracy and robustness than the traditional method. Simulation experiments show that the VM can extract centerlines from noisy images with an average accuracy of 0.35 pixels. Also, 3D reconstruction experiments of a phi 20-mm standard sphere demonstrate an average accuracy of 0.0282 mm. With four-thread acceleration, the proposed method can process images with a resolution of 2448 x 2048 pixels in 0.5 s. The accuracy and speed of the proposed method can be adjusted by changing the parameter related to the density of contour points, which makes this method flexible and widely applicable in applications with different requirements.

Study of path extraction algorithm based on improved boundary distance transformation

The Distance Transform Method in Virtual Endscope

New Fast Algorithm for Extracting Center Path

3-D centerline extraction for virtual cardiovascular endoscopy

An Automatic Centerline Extraction Algorithm Based on Distance from Boundary Field

A study on extracting centerline in virtual endoscopy

Improved algorithm for blood vessel centerline extraction

Fast Centerline Extraction Algorithm for Colon Detection

A 3-D Center Path Finding Algorithm Base on Two Distance Fields

Method Based on Edge Constraint and Fast Marching for Road Centerline Extraction from Very High-Resolution Remote Sensing Images

Centerline Extraction Based on Cross-section Algorithm and Snake Model

Centerline Extraction of 2D Vessels Based on Wavelet Magnitude

An Improved Extraction of Boundary Characteristic from Scattered Data

Automatic Centerline Extraction For Virtual Colonoscopy

Vascular Cross-Section Extraction Algorithm Based on Centerline

A novel segmentation algorithm for complex 3D mesh model in computer vision.

A Novel Type of Boundary Extraction Method and Its Statistical Improvement for Unorganized Point Clouds Based on Concurrent Delaunay Triangular Meshes

Coronary Centerline Extraction Based on Ostium Detection and Model-Guided Directional Minimal Path

An Effective Road Centerline Extraction Method From VHR

A Voronoi-Diagram-based Method for Centerline Extraction in 3D Industrial Line-Laser Reconstruction Using a Graph-Centrality-based Pruning Algorithm

Model-guided directional minimal path for fully automatic extraction of coronary centerlines from cardiac CTA.