Search

Purpose Measurement uncertainty calculation is an important and complicated problem in digitised components inspection. In such inspections, a coordinate measuring machine (CMM) and laser scanner are usually used to get the surface point clouds of the component in different postures. Then, the point clouds are registered to construct fully connected point clouds of the component’s surfaces. However, in most cases, the measurement uncertainty is difficult to estimate after the scanned point cloud has been registered. This paper aims to propose a simplified method for calculating the uncertainty of point cloud measurements based on spatial feature registration. Design/methodology/approach In the proposed method, algorithmic models are used to calculate the point cloud measurement uncertainty based on noncontact measurements of the planes, lines and points of the component and spatial feature registration. Findings The measurement uncertainty based on spatial feature registration is related to the mutual position of registration features and the number of sensor commutation in the scanning process, but not to the spatial distribution of the measured feature. The results of experiments conducted verify the efficacy of the proposed method. Originality/value The proposed method provides an efficient algorithm for calculating the measurement uncertainty of registration point clouds based on part features, and therefore has important theoretical and practical significance in digitised components inspection.

A CAD-based approach to plan the laser scanning path of complex surfaces is presented. Firstly, in the laser scanning process the issues of the positional relationship between scanning direction and normal vectors of sampling points of the measured surface, sensor work area check and surface occlusion judgment are researched and a method for generating a single arbitrary surface scanning path is developed. Secondly, through the analysis of surface segmentation, scanning directions and displacements synthesis during the multi-surface scanning process, a scanning path planning method based on any complex multi-curved surface is established. Finally, through a simulation example, the feasibility of the method is verified.

In industrial production, the machine vision save a lot of manpower and material resources. And it has become an indispensable part of quality control solutions in automobile and parts manufacturing. In order to detect the integrity of automatic assembly for lamp parts, we first use the shape feature to determine whether the part shape is right through egde detection and template matching algorithms. And if the shape is right, we can further detect the color feature which are different in different sales regions using the color moments. As a result, the missing and mistake packing can be detected.

3D reconstruction as the basis of many applications,such as 3D printing, has become more and more importantfor many enterprises and researchersThe very important step in 3D reconstruction is the joining together of point cloud.This paper introduces the structures of a system to obtain three-dimensional point cloud data and a kind ofmethodsusing of the system to get point cloud data through the rotation and translation of the coordinate system, joining together the point cloud data.Experiment shows that this method has achieved good effect.

A novel method of measuring gap and surface differential was proposed to detect headlights contour in this paper. The method is based on the line structured laser sensor and robot system, and spline curve fitting was used to obtain the contour of the headlights and mold. After the fitting, the contour was resampled to get the shortest distance vector between the headlight contour curves and the mold contour curves, and through the decomposition of the vectors, the perpendicular and parallel components were obtained, and then the gap and surface differential measurement was completed.

Due to the impact of the lens optical system and the different manufacturing processes, there could be causing the vignetting when using a CCD camera to take pictures. This paper analyzes the causes of the lens vignetting and put forward a method for reducing this error to improve the accuracy of the intensity of illumination detection. Firstly, according to the analysis of the existing research on the camera lens, concluded that there are two factors that cause vignetting. Secondly, based on the conclusion, do large amounts of test in each case. Analyze for the influence of illumination test in each case. Finally, the article put forward a set of vignetting correction method to correct the image be collected in a fixed aperture. The result shows that the method can effectively reduce vignetting measuring error and has a good stability.

Camshift tracking algorithm is based on probability distribution of color , it is susceptible to be interfered by the same color in the background, which will lead to the failure of the target tracking. To overcome this problem it presented an improved Camshift tracking algorithm. It combined background subtraction method with three frame difference method to detect target, got rectangular characteristic parameters of the motion target area as the Camshift initialization parameters, replaced the general Camshift algorithm which is based on color feature. Experimental results show that Camshift algorithm combining the background subtraction method with three frame difference method can meet the requirements of the real-time and stability to a certain extent.

For wafer defects detection system in a certain resolution, CCD camera cannot get the whole wafer image at one-time, which has an influence on the subsequent defect feature extraction. This paper puts forward an image registration method based on geometry features. Firstly, the information of the image edge can be extracted by using an improved edge detection operator, then using Hough transform to extract the horizontal and vertical lines of the information of the image edge. Secondly, using the correlation of linear characteristic to define the registration standards of the image. A new method, simplex-simulated annealing algorithm, is presented to optimize the registration coefficient of the image. Finally the method is tested and evaluated by the matching effect, the results show that it can effectively achieve the automatic wafer image registration and has a good stability.

This paper put forward a 3D reconstruction method of the headlight contours based on laser scanning technology and robotics. Firstly, according to the present three-dimensional measurement techniques, the article put forward a set of headlight contour detection method based on the analytic geometry model and the line laser source scanning principle. It establishes a 3D scanning model and coordinate transformation model for 3D reconstruction of the headlight contour. Secondly, according to the demanding accuracy it structures the 3D reconstruction system. Finally it realizes the 3D reconstruction of the headlight contour based on the method, and the result is tested and evaluated matching effect, the result shows that can effectively realize the 3D reconstruction of headlight contour and the method has a good stability.

As the rapid progress and development of DSP technology, a new way to solve the video signal processing was accepted. At present, binocular measurement instrument are mostly based on image capture card and PC on the market, expensive and bulky. For this problem, binocular vision measurement system based on DSP has been used. The design of elementary program structure about video processing which based on DSP/BIOS was finished and verified for its real-time. The results of the experiment show that the driver can be well used in the system.

For more spacious place where the human bodies are sparse and scattered, it is feasible and convenient to use algorithm of frames subtraction or background elimination which can extract the human bodies. But for the places where the human bodies are intensive such as on the buses, it is difficult to segment the bodies using this algorithm. Since the heads are more scattered than the bodies, an algorithm based on several regional growths and feature extraction is studied to detect the heads in this paper. If the clothes gray is similar to the head gray, morphologic operation is used to extract the head again. Also this paper shows the flow of the algorithm in detail. Many images taken by the bus camera were used for the experiment. The result shows that the bus passengers could be detected exactly which verified the effectiveness of the algorithm.

The beam direction is one of the important items of headlamp detection, but lots of internal detection machines can not give it an accurate detection. This article analyses the factors which affect the detection result, then gives a new accurate solution. It uses an automobile position fixing system and data correction algorithm. Firstly, the beam direction angle and the automobile's laying deviation angle can be got by using computer vision technique and image processing technique. Secondly, the beam direction angle can be corrected according to the automobile's laying deviation angle. This new method can efficiently find accurate optical axis direction of low beam and high beam. So it tells people whether the automobile works in good condition. It solves the detection problems of improper parking distance and improper laying position, decreases detection error and increases precision. What's more, it solves the internal actuality of inaccurate detection.

According to the features of human head which is taken by camera shooting from the top down, this paper puts forward a new and rapid algorithm for detecting human head. First, the image is binarized by threshold, then searches diameter of circular-like head area in binary image. If the diameter is in conformity with the feature of human head, then identifies the possible head further in other directions. This algorithm avoids complex concepts and mathematical formulas. Its time-space complexity is much lower than other common methods used in detecting circular-like targets. This paper compares this algorithm with Hough Transform (HT) which is typical in detecting circular-like targets. It is several tens of times faster than HT, and the memory this algorithm needs is several tens of times less than HT. At the end of this paper, practice proves its good effect.

The corners are the key points of light shape of automotive low-beam headlamp. This paper explains, in detail, the principle of detecting the corners of light shape of automotive low-beam headlamp with Inverse Hough Transform. There are three major steps. Firstly, gradient method is used to obtain the dividing line between brightness and darkness in the light shape of low-beam headlight. Secondly, the straight lines are detected by Hough Transform. Thirdly, the cross-points of straight lines are obtained by Inverse Hough Transform. Then according to the reasonability of cross-points, unreasonable cross-points are removed, thus the corners of light shape are obtained. Experiments are used to prove that this method can resist noise and work on real time. It realizes the transformation from manual detection to automatic detection in detecting optical axes of vehicle lamp.

The fusion techniques have been developed quickly in recent years and become an important remote sensing research topic. This paper systematically discusses the technique used in pixel level image fusion including IHS transform, YIQ transform, HLS transform, HSV transform, PCA transform, HPF transform and wavelet transform image fusion method. Then a fusion experiment of IKONOS image is made to compares the different merging methods from spectral quality and the spatial quality in order to choose the suitable method for the high resolution image.

The kernel fractional-stop nonlinear discriminant analysis (KF-NDA) method not only extends the fractional-step linear discriminant analysis (F-LDA) method to a nonlinear version, but also further improves the generalization ability of traditional kernel nonlinear discriminant analysis (K-NDA). On the other hand, the Gabor transformed face images exhibit strong characteristics of spatial locality, scale and orientation selectivity, similar to those displayed by Gabor wavelets. Such characteristics produce salient local features that are most suitable for face recognition (FR). Hence, the augmented Gabor feature vector (AGFV) derived from a set of downsampled Gabor wavelet representations of face images is robust to the various of face images and simultaneously exhibits the more discriminatory information. Based on the AGFV and the KF-NDA, a robust feature extraction framework, i.e., the Gabor KF-NDA (GKF-NDA), is proposed for FR. In this framework, the KF-NDA method is directly applied to extract the robust nonlinear feature from the AGFV. Experimental results tested on the popular databases show that the GKF-NDA is more effective than oilier existing FR approaches.

Recently, a direct fractional-step linear discriminant analysis (DF-LDA) algorithm was proposed and successfully applied to face recognition (FR). However, the classification performance of DFT-LDA is degraded by the limitations of the direct linear discriminant analysis (D-LDA) used in DF-LDA. We describe a novel DF-LDA to solve this problem, and based on this novel DF-LDA, a novel Gabor DF-LDA (GDF-LDA), which directly applies the novel DF-LDA to the high-dimensional augmented Gabor feature vectors (AGFV) derived from the Gabor wavelet representation of face images, has been proposed for FR. The GDF-LDA not only is robust to facial variations, but also overcomes the limitations of the previous DF-LDA. The comparative results on the ORL database show that the GDF-LDA is more effective than existing FR methods.

On the one hand, the support vector machine (SVM) has the high performance in tackling small sample size and high-dimensional data, and has the good generalization ability too. On the other hand, Gabor wavelet exhibits strong characteristics of spatial locality, scale, and orientation selectivity, and the Gabor representations of face images can produce salient local features that are most suitable for face recognition. This paper proposes a new face recognition method based on independent Gabor features (IGF) and SVM. The proposed method has four steps as follows: 1) an augmented Gabor feature vector (AGFV) is derived from a set of downsampled Gabor wavelet representations of face images; 2) an IGF is obtained by applying the independent component analysis (ICA) to the AGFV; 3) To decrease the computational complexity and improve the recognition rate, Genetic Algorithms (GA) is used to select the optimal IGF set for classification; 4) the SVM is used to classify the optimal IGF. The experiments tested on the Yale database show that this method is very effective.

Face recognition problem is challenging because face images can vary considerably in terms of facial expressions, lighting conditions and so on. This paper introduces a novel face recognition using support vector machines with the robust feature extracted by kernel principal component analysis (KPCA), which is robust to facial variations. This method derives firstly an augmented Gabor-face vector based on the Gabor wavelet transformation of face images using different orientation and scale local feature, which is robust to changes in facial expression and pose. KPCA is used to extract the feature of the augmented Gabor-face vector so that the principal components is computed within the space spanned by high-order correlation of input of augmented Gabor-face vector and produce a good performance. Finally, the support vector machine (SVM), which has high generalization capabilities and high performance in tackling small sample size in the pattern recognition task, is used to classify the feature. The comparative experiments in the ORL face database show that this algorithm is more effective than the previous methods.

This paper introduces a novel Gabor Generalized Foley-Sammon Transform (GGFST) method for face recognition (FR). The GGFST method can directly apply the generalized Foley-Sammon transform (GFST) method that has the best separable ability in a global sense to the high-dimensional augmented Gabor feature vectors derived from the Gabor wavelet representation of face images. This method has three novelties: 1) the GGFST method is robust to facial variations; 2) the GGFST method can overcome the limitations of traditional FR approaches by incorporating some middle methods as the preprocessing steps for dimension reduction so as to discard some significant discriminatory information; and 3) the GGFST method has the best separable ability in a global sense. The comparative experiments on the ORL database show that the GGFST method is more effective than the previous methods.