Image Registration for Multimodal Inspection of Mechanical Parts

Inspection Data

Image registration is the process of aligning the corresponding features of images in the same coordinate system. Multimodal registration has been widely used in medical imaging and geographic imaging. However, it has not been broadly applied in the inspection imaging of mechanical parts. Multimodal registration can improve inspection accuracy and quality by combining complementary inspection data from different inspection methods, or “modalities”. The research focus of this work is to develop a computational algorithm to register a CMM point cloud with a CT image in the 2-D (planar) domain. Dealing with outliers is the major concern for achieving required registration accuracy. Targeting solving this problem, a new registration metric is proposed in this work, which makes application of the traditional ICP (Iterative Closest Point) algorithm robust, by optimizing the search for closest points.

Application of the 'Iterative Closest Point' Algorithm for Determination of Rotation Axis of a Turntable

Solid State Phenomena ◽

10.4028/www.scientific.net/ssp.199.273 ◽

2013 ◽

Vol 199 ◽

pp. 273-278

Author(s):

Ireneusz Wróbel

Keyword(s):

Reverse Engineering ◽

Point Cloud ◽

Rotation Axis ◽

Rapid Development ◽

Point Clouds ◽

Iterative Closest Point ◽

Mechanical Parts ◽

Scanning Process ◽

Reverse engineering [ is a field of technology which has been under rapid development for several recent years. Optic scanners are basic devices used as reverse engineering tools. Point cloud describes the shape of a scanned object. Automatic turntable is a device which enables a scanning process from different viewing angles. In the paper, the algorithm is described which has been used for determination of rotation axis of a turntable. The obtained axis constitutes the base for an aggregation of particular point clouds into single resultant common cloud describing the shape of the scanned object. Usability of this algorithm for precise scanning of mechanical parts was validated, precision of shape replication was also evaluated.

ITERATIVE CLOSEST POINT ALGORITHM FOR ACCURATE REGISTRATION OF COARSELY REGISTERED POINT CLOUDS WITH CITYGML MODELS

ISPRS Annals of Photogrammetry Remote Sensing and Spatial Information Sciences ◽

10.5194/isprs-annals-iv-2-w5-201-2019 ◽

2019 ◽

Vol IV-2/W5 ◽

pp. 201-208 ◽

Cited By ~ 1

Author(s):

S. Goebbels ◽

R. Pohle-Fröhlich ◽

P. Pricken

Keyword(s):

Point Cloud ◽

Point Clouds ◽

Iteration Step ◽

Iterative Closest Point ◽

Target Point ◽

Standard Tool ◽

Test Scenarios ◽

Plane Mode ◽

City Models

<p><strong>Abstract.</strong> The Iterative Closest Point algorithm (ICP) is a standard tool for registration of a source to a target point cloud. In this paper, ICP in point-to-plane mode is adopted to city models that are defined in CityGML. With this new point-to-model version of the algorithm, a coarsely registered photogrammetric point cloud can be matched with buildings’ polygons to provide, e.g., a basis for automated 3D facade modeling. In each iteration step, source points are projected to these polygons to find correspondences. Then an optimization problem is solved to find an affine transformation that maps source points to their correspondences as close as possible. Whereas standard ICP variants do not perform scaling, our algorithm is capable of isotropic scaling. This is necessary because photogrammetric point clouds obtained by the structure from motion algorithm typically are scaled randomly. Two test scenarios indicate that the presented algorithm is faster than ICP in point-to-plane mode on sampled city models.</p>

Registration method for three-dimensional point cloud in rough and fine registrations based on principal component analysis and iterative closest point algorithm

Traitement du signal ◽

10.3166/ts.34.57-75 ◽

2017 ◽

Vol 34 (1-2) ◽

pp. 57-75

Author(s):

Heqiang TIAN ◽

Xiaoqing DANG ◽

Jihu WANG ◽

Dongmei WU

Keyword(s):

Principal Component Analysis ◽

Point Cloud ◽

Three Dimensional ◽

Principal Component ◽

Component Analysis ◽

Iterative Closest Point ◽

Registration Method ◽

MPCR-Net: Multiple Partial Point Clouds Registration Network Using a Global Template

10.20944/preprints202109.0180.v1 ◽

2021 ◽

Author(s):

Shijie Su ◽

Chao Wang ◽

Ke Chen ◽

Jian Zhang ◽

Yang Hui

Keyword(s):

Point Cloud ◽

Cloud Model ◽

Ideal Point ◽

Point Clouds ◽

3D Measurement ◽

Registration Accuracy ◽

Point Cloud Registration ◽

Point Groups ◽

Rigid Body Transformation

With the advancement of photoelectric technology and computer image processing technology, the visual measurement method based on point clouds is gradually applied to the 3D measurement of large workpieces. Point cloud registration is a key step in 3D measurement, and its registration accuracy directly affects the accuracy of 3D measurements. In this study, we designed a novel MPCR-Net for multiple partial point cloud registration networks. First, an ideal point cloud was extracted from the CAD model of the workpiece and was used as the global template. Next, a deep neural network was used to search for the corresponding point groups between each partial point cloud and the global template point cloud. Then, the rigid body transformation matrix was learned according to these correspondence point groups to realize the registration of each partial point cloud. Finally, the iterative closest point algorithm was used to optimize the registration results to obtain a final point cloud model of the workpiece. We conducted point cloud registration experiments on untrained models and actual workpieces, and by comparing them with existing point cloud registration methods, we verified that the MPCR-Net could improve the accuracy and robustness of the 3D point cloud registration.

Proceedings of the International Conference on Evolutionary Computation Theory and Applications ◽

THREE-DIMENSIONAL POINT-CLOUD REGISTRATION USING A GENETIC ALGORITHM AND THE ITERATIVE CLOSEST POINT ALGORITHM

10.5220/0003718405470552 ◽

2011 ◽

Keyword(s):

Genetic Algorithm ◽

Point Cloud ◽

Three Dimensional ◽

Iterative Closest Point ◽

Point Cloud Registration ◽

MPCR-Net: Multiple Partial Point Clouds Registration Network Using a Global Template

Applied Sciences ◽

10.3390/app112210535 ◽

2021 ◽

Vol 11 (22) ◽

pp. 10535

Author(s):

Shijie Su ◽

Chao Wang ◽

Ke Chen ◽

Jian Zhang ◽

Hui Yang

Keyword(s):

Point Cloud ◽

Cloud Model ◽

Ideal Point ◽

Point Clouds ◽

3D Measurement ◽

Registration Accuracy ◽

Point Cloud Registration ◽

Point Groups ◽

Rigid Body Transformation

With advancements in photoelectric technology and computer image processing technology, the visual measurement method based on point clouds is gradually being applied to the 3D measurement of large workpieces. Point cloud registration is a key step in 3D measurement, and its registration accuracy directly affects the accuracy of 3D measurements. In this study, we designed a novel MPCR-Net for multiple partial point cloud registration networks. First, an ideal point cloud was extracted from the CAD model of the workpiece and used as the global template. Next, a deep neural network was used to search for the corresponding point groups between each partial point cloud and the global template point cloud. Then, the rigid body transformation matrix was learned according to these correspondence point groups to realize the registration of each partial point cloud. Finally, the iterative closest point algorithm was used to optimize the registration results to obtain the final point cloud model of the workpiece. We conducted point cloud registration experiments on untrained models and actual workpieces, and by comparing them with existing point cloud registration methods, we verified that the MPCR-Net could improve the accuracy and robustness of the 3D point cloud registration.

The edge-driven dual-bootstrap iterative closest point algorithm for multimodal retinal image registration

10.1117/12.770840 ◽

2008 ◽

Cited By ~ 1

Author(s):

Chia-Ling Tsai ◽

Chun-Yi Li ◽

Gehua Yang

Keyword(s):

Image Registration ◽

Retinal Image ◽

Iterative Closest Point ◽

Iterative closest point registration for fast point feature histogram features of a volume density optimization algorithm

Measurement and Control ◽

10.1177/0020294019878869 ◽

2020 ◽

Vol 53 (1-2) ◽

pp. 29-39 ◽

Cited By ~ 3

Author(s):

Lu-shen Wu ◽

Guo-lin Wang ◽

Yun Hu

Keyword(s):

Coordinate System ◽

Optimization Algorithm ◽

Point Cloud ◽

High Speed ◽

Local Coordinate System ◽

Volume Density ◽

Iterative Closest Point ◽

Registration Accuracy ◽

Fine Registration ◽

Point Feature

Motivated by the high speed but insufficient precision of the existing fast point feature histogram algorithm, a new fast point feature histogram registration algorithm based on density optimization is proposed. In this method, a 44-section blank feature histogram is first established, and then a principal component analysis is implemented to calculate the normal of each point in the point cloud. By translating the coordinate system in the established local coordinate system, the normal angle of each point pair and its weighted neighborhood are obtained, and then a fast point feature histogram with 33 sections is established. The reciprocal of the volume density for the central point and its weighted neighborhood are calculated simultaneously. The whole reciprocal space is divided into 11 sections. Thus, a density fast point feature histogram with 44 sections is obtained. On inputting the testing models, the initial pose of the point cloud is adjusted using the traditional fast point feature histogram and the proposed algorithms, respectively. Then, the iterative closest point algorithm is incorporated to complete the fine registration test. Compared with the traditional fine registration test algorithm, the proposed optimization algorithm can obtain 44 feature parameters under the condition of a constant time complexity. Moreover, the proposed optimization algorithm can reduce the standard deviation by 8.6% after registration. This demonstrates that the proposed method encapsulates abundant information and can achieve a high registration accuracy.