Search results

This paper aims to investigate a probabilistic mixture model for the nonrigid point set registration problem in the computer vision tasks. The equations to estimate the mixture model parameters and the constraint items are derived simultaneously in the proposed strategy.

The problem of point set registration is expressed as Laplace mixture model (LMM) instead of Gaussian mixture model. Three constraint items, namely, distance, the transformation and the correspondence, are introduced to improve the accuracy. The expectation-maximization (EM) algorithm is used to optimize the objection function and the transformation matrix and correspondence matrix are given concurrently.

Although amounts of the researchers study the nonrigid registration problem, the LMM is not considered for most of them. The nonrigid registration problem is considered in the LMM with the constraint items in this paper. Three experiments are performed to verify the effectiveness and robustness and demonstrate the validity.

The novel method to solve the nonrigid point set registration problem in the presence of the constraint items with EM algorithm is put forward in this work.

This paper aims to propose a robust method for non-rigid point set registration, using the Gaussian mixture model and accommodating non-rigid transformations. The posterior probabilities of the mixture model are determined through the proposed integrated feature divergence.

The method involves an alternating two-step framework, comprising correspondence estimation and subsequent transformation updating. For correspondence estimation, integrated feature divergences including both global and local features, are coupled with deterministic annealing to address the non-convexity problem of registration. For transformation updating, the expectation-maximization iteration scheme is introduced to iteratively refine correspondence and transformation estimation until convergence.

The experiments confirm that the proposed registration approach exhibits remarkable robustness on deformation, noise, outliers and occlusion for both 2D and 3D point clouds. Furthermore, the proposed method outperforms existing analogous algorithms in terms of time complexity. Application of stabilizing and securing intermodal containers loaded on ships is performed. The results demonstrate that the proposed registration framework exhibits excellent adaptability for real-scan point clouds, and achieves comparatively superior alignments in a shorter time.

The integrated feature divergence, involving both global and local information of points, is proven to be an effective indicator for measuring the reliability of point correspondences. This inclusion prevents premature convergence, resulting in more robust registration results for our proposed method. Simultaneously, the total operating time is reduced due to a lower number of iterations.

In this work, the authors aim to provide a set of convenient methods for generating training data, and then develop a deep learning method based on point clouds to estimate the pose of target for robot grasping.

This work presents a deep learning method PointSimGrasp on point clouds for robot grasping. In PointSimGrasp, a point cloud emulator is introduced to generate training data and a pose estimation algorithm, which, based on deep learning, is designed. After trained with the emulation data set, the pose estimation algorithm could estimate the pose of target.

In experiment part, an experimental platform is built, which contains a six-axis industrial robot, a binocular structured-light sensor and a base platform with adjustable inclination. A data set that contains three subsets is set up on the experimental platform. After trained with the emulation data set, the PointSimGrasp is tested on the experimental data set, and an average translation error of about 2–3 mm and an average rotation error of about 2–5 degrees are obtained.

The contributions are as follows: first, a deep learning method on point clouds is proposed to estimate 6D pose of target; second, a convenient training method for pose estimation algorithm is presented and a point cloud emulator is introduced to generate training data; finally, an experimental platform is built, and the PointSimGrasp is tested on the platform.

Hull block assembly is a vital task in ship construction. It is necessary to obtain the actual poses of the assembly features to guide further block alignment. Traditional methods use single-point measurement, which is time-consuming and may lead to loss of key information. Thus, large-scale scanning is introduced for data acquisition, and this paper aims to provide a precise and robust method for retrieving poses based on point set registration.

The main problem of point registration is to find the correct transformation between the model and the scene. In this paper, a vote framework based on a new point pair feature is used to calculate the transformation. First, a special edge indicator for multiplate objects is proposed to determine the edges. Subsequently, pair features with an edge description are noted for every point. Finally, a voting scheme based on agglomerative clustering is implemented to determine the optimal transformation.

The proposed method not only improves registration efficiency but also maintains high accuracy compared to several commonly used approaches. In particular, for objects composed of plates, the results of pose estimation are more promising because of the compact pair feature. The multiple ship longitudinal localization experiment validates the effectiveness in real scan applications.

The proposed edge description performs a better detection for the edges of multiplate objects. The pair feature incorporating the edge indicator is more discriminative than the original template, resulting in better robustness to outliers, noise and occlusions.

This paper aims to automatically derive a 2D parametric model of the main characteristic lines of a car from images, blueprints or hand-made sketches of its side view. Then this model can be used for the further computer-aided design manipulation starting from images of the side view of a car.

The method combines different image edge detection techniques and edge removal processes with optimization techniques according to local and global constraints specific of the single curves to automatically construct a precise parametric model of the main character lines of a car from images. First, process the car image to compute the most important curves and then warp a car template model to match its feature points and curves with the ones detected in the image.

The paper provides method to construct parametric model from an image using maximum cover ratio to the edge points obtained by state-of-the-art edge detection algorithms. A feature points’ organization mechanism produces quadric curves to express feature curves of a product.

The robustness of the presented method depends on the completeness of edge detection results and the accuracy of some key points’ registration result, so if the image is not good, the result cannot be trusted. Only side-view is considered in this paper. Additional limits in the process regard the side view verification: pictures of the front or rear view can be wrongly classified as lateral ones when they contain round lights.

This program enables designers to convert the image to geometric parametric model directly.

The method is applicable to shaded pictures, sketches and blue prints of the side view of a car. It can process a database of car images in a batch mode or a specific picture on user demand. The method classifies the cars to different categories: SUV/Wagon/Hatchback, sedan, city and coupe. The authors obtain good results for every category.

The purpose of this paper is to find automatic post‐processing scheme to give textures and motion data to three dimensional (3D) body scan data.

Semi‐implicit particle‐based method was applied to post‐processing of 3D body scan data. The template avatar mesh was draped onto the target scan data and the texture/motion data were transferred to regenerated body. Automatic body feature detection was used to correlate the template body with the target body.

Using semi‐implicit particle method, there are advantages in both computational stability and accuracy. The calculation is done in a few minutes and even data with many holes could be used.

There are several researches for body feature detection and scan body regeneration but this paper aims for fully automatic method which needs no human intervention. The semi‐implicit particle method, which is popularly used for cloth simulation, is applied to body data regeneration. The conventional 3D body scan data, which had no colors and motions can be given textures and motions with this approach. And even the face can be freely interchanged with the use of external face generation software.

Examines the fifthteenth published year of the ITCRR. Runs the whole gamut of textile innovation, research and testing, some of which investigates hitherto untouched aspects. Subjects discussed include cotton fabric processing, asbestos substitutes, textile adjuncts to cardiovascular surgery, wet textile processes, hand evaluation, nanotechnology, thermoplastic composites, robotic ironing, protective clothing (agricultural and industrial), ecological aspects of fibre properties – to name but a few! There would appear to be no limit to the future potential for textile applications.

With the rapid development of the 3C industry, the problem of automated operation of 3C wire is becoming increasingly prominent. However, the 3C wire has high flexibility, and its deformation is difficult to model and control. How to realize the automation operation of flexible wire in 3C products is still an important issue that restricts the development of the 3C industry. Therefore, this paper designs a system that aims to improve the automation level of the 3C industry.

This paper designed a visual servo control system. Based on the perception of the flexible wire, a Jacobi matrix is used to relate the deformation of the wire to the action of the robot end; by building and optimizing the Jacobi matrix, the robot can control the flexible wire.

By using the visual servo control system, the shape and deformation of the flexible wire are perceived, and based on this, the robot can control the deformation of the flexible wire well. The experimental environment was built to evaluate the accuracy and stability of the system for controlling the deformation of the flexible wire.

An image-based visual servo system is proposed to operate the flexible wire, including the vision system, visual controller and joint velocity controller. It is a scheme suitable for flexible wire operation, which has helped to automate flexible wire-related industries. Its core is to correlate the motion of the robot end with the deformation of the flexible wire through the Jacobian matrix.

Additive manufacturing (AM) has shown its capability in producing complex geometries. Due to the additive nature, the in situ layer-wise inspection of geometric accuracy is essential to making AM reach its full potential. This paper aims to propose a novel automated in-plane alignment and error quantification framework to distinguish the fabrication, measurement and alignment errors in AM.

In this work, a multi-resolution framework based on wavelet decomposition is proposed to automatically align two-dimensional point clouds via a polar coordinate representation and then to differentiate errors from different sources based on a randomized complete block design approach. In addition, a two-stage optimization model is proposed to find the best configuration of the multi-resolution framework.

The proposed framework can not only distinguish errors attributed to different sources but also evaluate the performance and consistency of alignment results under different levels of details.

A sample part with different featured layers, including a simple free-form layer, a defective layer and a layer with internal features, is used to illustrate the effectiveness and efficiency of the proposed framework. The proposed alignment method outperforms the widely used iterative closest point algorithm.

This work fills a research gap of state-of-the-art studies by automatically quantifying different types of error inherent in manufacturing, measuring and part alignment.

Gives a bibliographical review of the finite element meshing and remeshing from the theoretical as well as practical points of view. Topics such as adaptive techniques for meshing and remeshing, parallel processing in the finite element modelling, etc. are also included. The bibliography at the end of this paper contains 1,727 references to papers, conference proceedings and theses/dissertations dealing with presented subjects that were published between 1990 and 2001.