Search results

Large size of aircraft assembly tooling structure and complex measurement environment exist. The laid enhanced reference points (ERS) are subject to a combination of nonuniform temperature fields and measurement errors, resulting in increased measurement registration errors. In view of the nonuniform temperature field and measurement errors affecting the ERS point registration problem, the purpose of this paper is to propose a neural network-based ERS point registration compensation method for large-size measurement fields under a nonuniform temperature field.

The approach is to collect ERS point information and temperature data, normalize the collected data to complete the data structure design and complete the construction of the neural network prediction model by data training. The data learning is performed to complete the prediction model construction, and the prediction model is used to complete the compensation analysis of ERS points. Finally, the algorithm is verified through experiments and engineering practice.

Experimental results show that the proposed neural network-based ERS point prediction and compensation method for nonuniform temperature fields effectively predicts ERS point deformation under nonuniform temperature fields compared with the conventional method. After the compensation analysis, the registration error is effectively reduced to improve registration accuracy. Reducing the combined effect of environmental nonuniform temperature field and measurement error has apparent advantages.

The method reduces the registration error caused by combining a nonuniform temperature field and measurement error. It can be used for aircraft assembly site prediction and registration error compensation analysis, which is essential to improve measurement accuracy further.

The size of the aircraft tooling structure is huge, and the ambient temperature is difficult to maintain a constant state. Aiming at the influence of current temperature, this paper aims to propose a compensation method for registration error of large-scale measurement fields based on multi-temperature sensors.

In this method, an enhanced reference points (ERS)–temperature regression model is constructed from ERS and temperature data. The ERS offsets compensation model is established by solving the offset through the regression model, and the ERS offset compensation analysis is carried out.

The experimental results show that the proposed registration error compensation algorithm has obvious advantages over traditional methods in reducing the influence of ambient temperature and improving the measurement accuracy by reducing the registration error.

This method reduces registration error caused by the influence of ambient temperature and is used for aircraft measurements in different temperature environments.

For the measurement of large-scale components in aircraft assembly, the evaluation of coordinate transformation parameters between the coordinate frames of individual measurement systems to the assembly frame is an essential task, which is usually completed by registration of the enhanced reference system (ERS) points. This paper aims to propose an analytical method to evaluate the uncertainties of transformation parameters considering both the measurement error and the deployment error of ERS points.

For each measuring station, the measured coordinates of ERS points are first roughly registered to the assembly coordinate system using the singular value decomposition method. Then, a linear transformation model considering the measurement error and deployment error of ERS points is developed, and the analytical solution of transformation parameters’ uncertainties is derived. Moreover, the covariance matrix of each ERS points in the transformation evaluation is calculated based on a new uncertainty ellipsoid model and variance-covariance propagation law.

For the transformation of both single and multiple measuring stations, the derived uncertainties of transformation parameters by the proposed analytical method are identical to that obtained by the state-of-the-art iterative method, but the solution process is simpler, and the computation expenses are much less.

The proposed uncertainty evaluation method would be useful for in-site measurement and optimization of the configuration of ERS points in the design of fixture and large assembly field. It could also be applied to other registration applications with errors on both sides of registration points.

Using a patented measuring technique, the perfecTest system has the ability to give a quantified and accurate measurement down to 25 micron (one mil) for any directional X‐axis and/or Y‐axis inner‐layer shift. Individual inner layers can be distinguished from each other and reviewed in either tabular or graphic format. Control limits are set within the software to provide pass/fail testing and all testing is performed in real time. With functions such as data storage, retrieval and SPC analysis, the system is a very powerful tool to monitor and manage inner‐layer registration problems and can lead the way to improved yield and reliability instead of creating more waste. Hole break‐out is the main cause of intermittent failure in PCBs. As the result of through‐hole plating, there is conductivity along the drill‐hole barrel. PCBs having such severe registration problems will pass in‐circuit and bare‐board tests. However, intermittent failure is often not detected until the product reaches the field and is in use.

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.

Light Detection and Ranging (LiDAR) offers a fast and effective way to acquire DSM and extract ground objects such as building, trees and so on. However, it is difficult to extract sharp and precise building boundary from LiDAR data, because its ground sample distance (GSD) is often worse than that of high resolution image. Recently, fusion of LiDAR and high resolution image becomes a promising approach to extract precise boundary. To find the correct and precise boundary, the aim of this paper is to present a series of novel algorithms to improve the quality.

To find the correct and precise boundary, this paper presents a series of novel algorithms to improve the quality. At first, a progressive algorithm is presented to register LiDAR data and images; second, a modified adaptive TIN algorithm is presented to filter ground point, where a region growth method is applied in the adaptive TIN algorithm; third, a novel criterion based on the density, connectivity and distribution of point cluster is developed to distinguish trees point; fourth, a novel method based on the height difference between neighbor points is employed to extract coarse boundaries; at last, a knowledge based rule is put forward to identify correct building boundary from parallel edges.

Thorough experiments, it is conducted that: the registration results are accurate and reliable; filtered ground points has good quality, without missing or redundancy; all tree clusters bigger than one grid are detected, and points of walls and edges are eliminated with the new criterion; detected edges exactly locate at real building boundaries, and statistics show the detection correctness is 98 percent, and the detection completeness is 95 percent.

All results prove that precise boundary can be extracted with fusion of LiDAR and high resolution image.

Veincheck is a biometric system which uses back‐of‐hand vein patterns to verify an individual's identity. A biometric system can be loosely defined as a means of verifying an individual's identity by analysis of a physiological feature or a characteristic action. There are a wide variety of biometric systems under development ranging from signature to fingerprint to voice. Biometric systems are often compared on the basis of their False Accept Rates (FAR) and False Reject Rates (FRR). These rates are also referred to as Type 1 and Type 2 errors respectively. The acceptable values and the relative importance for these error rates varies from application to application. It is generally agreed that error rates of much less than 1 per cent are needed before deploying a biometric system for use by the general public.

Tracking physical resources at the construction site can generate information to support effective decision-making and building production control. However, the methods for conventional tracking usually offer low reliability. This study aims to propose the integrated Smart Twins 4.0 to track and manage metallic formworks used in cast-in-place concrete wall systems using internet of things (IoT) (operationalized by radio frequency identification [RFID]) and building information modeling (BIM), focusing on increasing quality and productivity.

Design science research is the research approach, including an exploratory study to map the constructive system, the integrated system development, an on-site pilot implementation in a residential project and a performance evaluation based on acquired data and the perception of the project’s production team.

In all rounds of requests, Smart Twins 4.0 registered and presented the status from the formworks and the work progress of buildings in complete correspondence with the physical progress providing information to support decision-making during operation. Moreover, analyses of the system infrastructure and implementation details can drive researchers regarding future IoT and BIM implementation in real construction sites.

The primary contribution is the system proposal, centralized into a mobile app that contains a Web-based virtual model to receive data in real time during construction phases and solve a real problem. The paper describes Smart Twins 4.0 development and its requirements for tracking physical resources considering theoretical and practical previous research regarding RFID, IoT and BIM.

The modern manufacturing industry has put forward higher requirements for the assembly accuracy of components with the development of the industrial technology. For precision assembly, the traditional assembly process study based on tolerance has had difficulty in meeting these requirements. Hence, the distribution of the form errors must be considered. The registration between the two mating surfaces with form errors determines the parts’ assembly position, and is the basis for the prediction and control of the assembly accuracy. This study aims to provide a new surfaces registration method which takes form errors into consideration.

This study presents a new registration approach based on the minimum potential energy. A unique set of contact points on mating surfaces that meet the actual conditions can be obtained and the spatial position of the assembled part is calculated.

The experimental results show that the calculated values are in good agreement with the experimental values. Furthermore, the root mean square error is within 2%, which proves the validity and accuracy of the approach.

This paper provides an effective and new method for precision assembly which takes form errors into consideration. The method can give the optimal solution of the contact points, which is more consistent with the actual assembly situation and provides a basis for predicting assembly accuracy.