Yonghuai Liu

Learning and diagnosis in manufacturing processes through an executable Bayesian network

In this paper we present a novel approach to modelling a manufacturing process that allows one to learn about causal mechanisms of manufacturing defects through a Process Modelling and Executable Bayesian Network (PMEBN). The method combines probabilistic reasoning with time dependent parameters which are of crucial interest to quality control in manufacturing environments. We demonstrate the concept through a case study of a caravan manufacturing line using inspection data.

Correspondenceless motion estimation from range images

Estimation of rigid-body motion parameters in computer vision is normally performed from image correspondences between two coordinate frames. A large number of methods and algorithms have been proposed based on that such correspondences are known. Unfortunately, the establishment of correspondences is often time-consuming and, in many cases, impossible. In this paper, we propose a novel correspondenceless motion estimation algorithm based on the cross matrix. For a comparative study, we also implemented a correspondenceless motion estimation algorithm based on the scatter matrix. Experimental results have demonstrated that our method is more accurate and robust than the scatter matrix-based algorithm.

Using geometric properties of correspondence vectors for the registration of free-form shapes

The registration of free-form shapes by the iterative closest point algorithm (ICP) has attracted much attention from the computer vision and image processing community since it was first proposed in 1992. Many methods, mainly based on incorporating invariants described in a single coordinate frame have been devised to improve the accuracy and efficiency of the algorithm. In this paper, a novel method to improve image registration is proposed based on rigid constraints derived from geometric properties of correspondence vectors synthesised into a singe coordinate frame. False matches, which occur in almost every iteration of the ICP algorithm are eliminated through properties of the motion. For an accurate estimation of the geometric parameters of the motion, the Monte Carlo method is used in conjunction with a median filter. Experimental results based on both synthetic data and real images show that the improved method can effectively eliminate false matches, is accurate, robust, and efficient for the registration of free-form shapes with small motions.

Motion parameter constraints analysis from a single image

Motion parameter estimation is a fundamental problem in image processing and image understanding. A large number of algorithms have been proposed based on a number of different geometrical considerations, such as perspective or epipolar geometries. However, proposed motion estimation algorithms do not explicitly use the distance between feature points and angle information as rigid constraints to calibration. In this paper, we present a new geometric analysis of correspondence data and derive explicit expressions for rigid constraints that are then used to estimate motion parameters. We then present a novel, efficient motion parameter estimation algorithm based on a coarse to fine strategy from a single image data. For a comparative study of performance, we also extended to the 3D-2D case a well known 2D-2D motion estimation algorithm based on the epipolar geometry. Experimental results demonstrate that the coarse to fine strategy is appropriate for the problem and that the algorithm generally performs better than the extended epipolar geometry based algorithm.

A novel 3D-2D computer vision algorithm for automatic inspection of filter components

This paper describes investigation results on the design of a real time, computer vision based system for automatic inspection of filter components in a manufacturing line. The problem involves reasoning about an object’s 3D structure from 2D images. In computer vision, this is normally referred to as a 3D-2D problem. In this paper, we first present a geometrical analysis of image correspondence vectors synthesised into a single coordinate frame. The analysis is based on geometrical considerations that are fundamentally different from analytical, perspective, or epipolar geometries. The camera setup stems from the geometrical implications of such analysis and from the given background knowledge of the task within the context of the production line. We then describe a novel geometrical algorithm to estimate parameters of interest that include depth estimation and the position and orientation of the camera in world coordinate frame. The algorithm provides the closed form solution to all estimated parameters making full use of distance between feature vectors and angle information. For a comparative study of algorithm performance, we also developed an algorithm based on epipolar geometry. Experimental results show that the geometrical algorithm performs significantly better than the algorithm based on epipolar geometry.

Geometrical pose and structural estimation from a single image for automatic inspection of filter components

This paper describes research on the application of machine vision techniques to a real time automatic inspection task of air filter components in a manufacturing line. A novel calibration algorithm is proposed based on a special camera setup where defective items would show a large calibration error. The algorithm makes full use of rigid constraints derived from the analysis of geometrical properties of reflected correspondence vectors which have been synthesized into a single coordinate frame and provides a closed form solution to the estimation of all parameters. For a comparative study of performance, we also developed another algorithm based on this special camera setup using epipolar geometry. A number of experiments using synthetic data have shown that the proposed algorithm is generally more accurate and robust than the epipolar geometry based algorithm and that the geometric properties of reflected correspondence vectors provide effective constraints to the calibration of rigid body transformations.

Rigidity analysis from range image data for automatic inspection of filter components

Recent developments in computer vision and pattern recognition have enabled the development of sophisticated vision-based quality control systems for automatic inspection. We present a new geometrical analysis of rigid body transformation parameters from properties of reflected correspondence vectors. Based on this analysis, we propose a novel algorithm to calibrate transformation parameters of 3D objects in a fast production line of filter components. The algorithm performs a rigidity analysis from range images acquired from two cameras, making full use of distance and angle information providing a closed form solution to all parameters of interest. The method is used in conjunction with a decision support system where components falling outside specified thresholds are to be rejected. For a comparative study of algorithm performance, we also implemented a well known procedure for rigidity analysis based on quaternions. Experimental results demonstrate that our novel algorithm has a number of advantages over the quaternion method and that its performance is superior or similar to the quaternion method.

Using Rigid Constraints to Analyse Motion Parameters from Two Sets of 3D Corresponding Point Pattern

Given a rigid body motion expressed by a set of 2D or 3D correspondences, the distance between feature points and angular information can be used as rigid constraints to calibrate motion parameters. In this paper, we first present a novel geometrical analysis of properties of reflected correspondence vectors. The analysis provides explicit expressions for distance between feature points and angle measurement synthesised into a single coordinate frame providing the closed form solutions to all motion parameters of interest. A novel calibration algorithm is proposed and compared with an algorithm based on the least squares method. The algorithm demonstrates the importance of the geometrical properties of reflected correspondence vectors to motion parameter estimation and that it is generally more accurate than algorithms based on the least squares method.

A new correspondenceless geometric algorithm for motion estimation from range images

A number of techniques for motion estimation from range images have been proposed over the past decade. A feature of existing techniques is that they do not explicitly use distance and angle information as rigid constraints to estimation. We have investigated the implications of this to the performance of algorithms and, in this paper, we propose a new correspondenceless algorithm for motion estimation from range images making full use of rigid constraints and projection information. The algorithm provides a closed form solution to all parameters of interest based on geometrical properties of correspondence vectors that have been synthesised into a single coordinate frame. For a comparative study of performance, we also implemented another correspondenceless motion estimation algorithm based on the eigenstructure analysis of the scatter matrix. Experimental results based on both synthetic and real images show that the proposed algorithm is accurate and robust and that it compares favourably with the scatter matrix algorithm.

A two-stage approach to 3D scene structural analysis from a single two-dimensional image

Structural analysis of a 3D scene from 2D image pixel positions is an important and unresolved issue in image processing and pattern recognition. Current algorithms for 3D structural analysis do not explicitly use geometrical properties of distance between feature points and angle information as constraints to calibrate transformation parameters. We propose a novel two-stage approach to analyse the structure of a 3D scene based on the explicit expressions of distance and angle measurement. In the first stage, we transform the 3D-2D problem into a linear problem and propose a robust algorithm based on the total least squares method to estimate the rotation matrix and translation vector. In the second stage, we propose a coarse-to-fine geometric algorithm to refine the estimation. The proposed two-stage approach provides the closed form solutions to structural parameters through the analysis of image correspondence vectors synthesised into a single coordinate frame. For a comparative study of performance, we also extended a well known 2D-2D epipolar geometry algorithm proposed by R.Y. Tsai and T.S. Huang (1984) to solve 3D-2D problems. Experimental results show that the proposed approach and geometric algorithms are in general more accurate, stable, and efficient than the extended Tsai and Xuang algorithm.