Conference Paper

Calibration of a Structure Light based Windshield Inspection System

Authors:
To read the full-text of this research, you can request a copy directly from the authors.

Abstract

Three dimensional optic measurement system's accuracy is highly related with the field of inspection. Increasing of field inspection costs increasing camera / projector pixel area on the test surface. Small surface changes within one pixel area cannot be directly detected, which will lower the system accuracy. A pixel-to-pixel strategy is developed to solve this problem. Increasing field of inspection also costs a longer standoff distance. The random image noise from the environment, uncertainties functions by lens distortion and resolution variation are all amplified. Therefore, a more complicated calibration model for each pixel is proposed to calibrate the system. In traditional structured light vision systems, a single sensor usually detects around 10,000 - 50,000 mm2, and the 3D vision sensor in this paper needs to detect around 2,400,000 mm2. Larger detection range gives more challenge to finish the calibration tasks. This paper proposes a clear calibration procedure to a large field of inspection structured light system. Last the comparison with the CMM measured results is used to prove that the calibration tasks have been successfully achieved.

No full-text available

Request Full-text Paper PDF

To read the full-text of this research,
you can request a copy directly from the authors.

... We have the relation that = because the magnitude is zero in the z direction. Let denote the coordinate of the i-th virtual dot in the world frame; then, it can be represented by: (15) where is the transformation matrix from virtual screen to world frame. The view vector in the world coordinate frame corresponding to the i-th virtual dot can be identified due to camera calibration. ...
... The unit of the transmission distortion is degrees (in the color bar). Last, the different methods for 3D shape inspection are compared, including: CMM, diffuse object inspection system developed in our lab [15], transparent object inspection system proposed in the paper. ...
Article
Full-text available
In this paper, an automotive glass inspection system based on fringe patterns is proposed. This system can inspect the 3D shape, reflection normal and transmission distortion of the automotive glass simultaneously. A stereo vision and an iterative algorithm based on the reflection law of the specular surface are utilized to obtain the 3D shape and reflection normal of the inspected part. Then, the bottom reflection of the glass is analyzed such that the line-shifting fringe pattern, based on step wave, is adopted. Additionally, the transmission distortion is defined as an angle between the incoming ray and the outgoing ray for the inspected automotive glass. The method proposed is simpler and more flexible than traditional methods. Finally, the system accuracy has been examined, with the results demonstrating its high accurate performance.
Article
Full-text available
In this paper a new calibration method is proposed for projector using an uncalibrated camera. The dif-ficulty to get a set of 2D-3D points in order to com-pute the projection matrix between the projector and the world has no easy solution. Methods are either inaccu-rate or inflexible. We propose an easy and automatic way to calibrate a projector. This method is a combina-tion of pattern detection and visual servoing. The idea is to project a calibration pattern and superimpose it on a known printed pattern. We tested and demonstrated the feasibility and accuracy of our approach.
Conference Paper
Full-text available
This paper presents new calibration methods for a robotic area sensing system developed for industrial manufacturing inspection. A pixel-to-pixel calibration scheme is introduced to obtain standoff and baseline distance of the developed area sensor. In the three-dimensional (3D) space, the area sensor also has an offset angle in a plane parallel to the reference plane. Calibration of this offset angle improves the measurement accuracy. This pixel-to-pixel scheme is particularly useful for calibrating a 3D optical area sensor with off-the-shelf lenses. An exploding vector method was also developed to transform the measured depth map to a 3D point cloud. A robotic area sensing system was developed by integrating an area sensor prototype and a robot sensor planning system. Calibration of the integrated system includes robot calibration, part calibration, and robot hand-eye calibration. Calibration error of the robot system is related to the area sensor positioning accuracy in the robotic system. Experimental results show a successful practice of the developed robotic area sensing system. Measurement performance of the developed system can be improved if advanced lenses are adopted.
Conference Paper
Full-text available
Obtaining automatic 3D profile of objects is one of the most important issues in computer vision. With this information, a large number of applications become feasible: from visual inspection of industrial parts to 3D reconstruction of the environment for mobile robots. In order to achieve 3D data, range finders can be used. Coded structured light approach is one of the most widely used techniques to retrieve 3D information of an unknown surface. An overview of the existing techniques as well as a new classification of patterns for structured light sensors is presented. This kind of systems belong to the group of active triangulation method, which are based on projecting a light pattern and imaging the illuminated scene from one or more points of view. Since the patterns are coded, correspondences between points of the image(s) and points of the projected pattern can be easily found. Once correspondences are found, a classical triangulation strategy between camera(s) and projector device leads to the reconstruction of the surface. Advantages and constraints of the different patterns are discussed.
Article
Full-text available
The authors describe a general-purpose, representation-independent method for the accurate and computationally efficient registration of 3-D shapes including free-form curves and surfaces. The method handles the full six degrees of freedom and is based on the iterative closest point (ICP) algorithm, which requires only a procedure to find the closest point on a geometric entity to a given point. The ICP algorithm always converges monotonically to the nearest local minimum of a mean-square distance metric, and the rate of convergence is rapid during the first few iterations. Therefore, given an adequate set of initial rotations and translations for a particular class of objects with a certain level of `shape complexity', one can globally minimize the mean-square distance metric over all six degrees of freedom by testing each initial registration. One important application of this method is to register sensed data from unfixtured rigid objects with an ideal geometric model, prior to shape inspection. Experimental results show the capabilities of the registration algorithm on point sets, curves, and surfaces
Article
Full-text available
This paper describes the measurement principle, the mathematical model, the calibration procedure, and the error model of a range sensor based on the Coded-Light Approach. The sensor consists of a CCD camera and a stripe projector. 3-D measurements are realized by the intersection of camera rays with the planes generated by the stripe projection. Using projective geometry renders camera and projector model linear. A calibration procedure based on general least squares estimation is proposed. The estimation model consists of the analytic and stochastic model of the range sensor, which yields estimations of the model parameters as well as their accuracies. With introduction of unknown 3-D check points in the estimation model, the accuracy of the range measurements is obtained as the main result of this work. The confidence interval of the range measurements is represented as an error ellipsoid. It turned out that the longest axis of the error ellipsoid is almost parallel to ...
Article
A new technique for three-dimensional (3D) camera calibration for machine vision metrology using off-the-shelf TV cameras and lenses is described. The two-stage technique is aimed at efficient computation of camera external position and orientation relative to object reference coordinate system as well as the effective focal length, radial lens distortion, and image scanning parameters. The two-stage technique has advantage in terms of accuracy, speed, and versatility over existing state of the art. A critical review of the state of the art is given in the beginning. A theoretical framework is established, supported by comprehensive proof in five appendixes, and may pave the way for future research on 3D robotics vision. Test results using real data are described. Both accuracy and speed are reported. The experimental results are analyzed and compared with theoretical prediction. Recent effort indicates that with slight modification, the two-stage calibration can be done in real time.
Article
Tool condition monitoring has gained considerable importance in the manufacturing industry over the preceding two decades, as it significantly influences the process economy and the machined part quality. Recent advances in the field of image processing technology have led to the development of various in-cycle vision sensors that can provide a direct and indirect estimate of the tool condition. These sensors are characterised by their measurement flexibility, high spatial resolution and good accuracy. This paper provides a review of the basic principle, the instrumentation and the various processing schemes involved in the development of these sensors.
Article
In this paper, we present a detailed model and analysis of several error sources and thier effects on measuring three-dimensional (3D) surface properties using the structured lighting technique. The analysis is based on a general system configuration and identifies three types of error surces—system modeling error, image processing error and experimental error. Absolute and relative error bounds in obtaining 3D surface orientation and curvature measurements using structured lighting are derived in terms of the system parameters and likely error sources. In addition to the quantization error, other likely error sources in system modeling and experimental setup are also considered. Even though our analysis is on structured lighting, the results are readily applicable to other triangulation-based techniques such as stereopsis. Finally, our analysis focuses on error in inferring surface orientation and principal surface curvature. Such analyses, to our knowledge, have never been attempted before.
Conference Paper
In this paper we introduce a new point cloud registration strategy for 3D surface inspection, named object-oriented method. Traditionally, the registration methods use a fitting based algorithm, which is a recursive process and too time consuming to do quality inspection. Moreover, the fitting based algorithms require that the measured data and the CAD model have same shapes. Photogrammetry and ICP-based fitting algorithms are commonly used in this area. However, both of them cannot be applied directly to online point cloud registration for quality inspection. Since it requires: 1.) no markers on the object surface; 2.) a very limited inspection period of time for each part on an assembly line; and 3.) the measured work piece can not be assumed as same as the CAD model. The online registration is currently demanded by 3D surface inspection in automotive glasses manufacturing industry. The object-oriented registration method is developed based on matching the invariant geometric features between two data sets instead of the artificial markers and the recursive fitting algorithms. Therefore the registration speed is much higher. A datum filter is developed to extract the invariant features from the measured point cloud, based on the information that the CAD model carries. The curvature information and moment values are used as characteristic features to design the filter. The registration matrix is calculated by the invariance parts in both data sets. The new online registration method has been applied to the quality inspection in the windshield manufacturing industry. Experimental results are presented in this paper to demonstrate the effectiveness of the developed strategy.
Article
This paper describes a general purpose, representation independent method for the accurate and computationally efficient registration of 3-D shapes including free-form curves and surfaces. The method handles the full six-degrees of freedom and is based on the iterative closest point (ICP) algorithm, which requires only a procedure to find the closest point on a geometric entity to a given point. The ICP algorithm always converges monotonically to the nearest local minimum of a mean-square distance metric, and experience shows that the rate of convergence is rapid during the first few iterations. Therefore, given an adequate set of initial rotations and translations for a particular class of objects with a certain level of 'shape complexity', one can globally minimize the mean-square distance metric over all six degrees of freedom by testing each initial registration. For examples, a given 'model' shape and a sensed 'data' shape that represents a major portion of the model shape can be registered in minutes by testing one initial translation and a relatively small set of rotations to allow for the given level of model complexity. One important application of this method is to register sensed data from unfixtured rigid objects with an ideal geometric model prior to shape inspection. The described method is also useful for deciding fundamental issues such as the congruence (shape equivalence) of different geometric representations as well as for estimating the motion between point sets where the correspondences are not known. Experimental results show the capabilities of the registration algorithm on point sets, curves, and surfaces.
Article
In this paper we address the problem of geometric video projector calibration using a markerless planar surface (wall) and a partially calibrated camera. Instead of using control points to infer the camera-wall orientation, we find such relation by efficiently sampling the hemisphere of possible orientations. This process is so fast that even the focal of the camera can be estimated during the sampling process. Hence, physical grids and full knowledge of camera parameters are no longer necessary to calibrate a video projector.
Article
We investigate the problem of dynamic calibration for our structured light system. First, a method is presented to estimate the rotation matrix and translation vector between the camera and the projector using plane-based homography. Then an approach is introduced to analyze theoretically the error sensitivity in the estimated pose parameters with respect to noise in the projection points. This algorithm is simple and easy to implement. Finally, some numerical simulations and real data experiments are carried out to validate our method.
Conference Paper
This paper introduces a fully automated dimensional inspection (ADI) system using a digital area sensor. Instead of measuring a part surface point-by-point using a Coordinate Measuring Machine (CMM), an ADI system can measure a part surface patch-by-patch. Therefore, the inspection time can be significantly reduced. Many existing sensor planners were developed for one or two cameras. In the proposed ADI system, the sensor planner is developed for a digital area sensor, which is made by a CCD camera and a digital projector. In application, sensor viewpoints are estimated automatically. On each viewpoint, thousands of points on part surfaces can be measured in only several seconds.
Article
In this paper, the procedure developed to calibrate a whole-field optical profilometer and the evaluation of the measurement performance of the system are presented. The sensor is based on the projection of structured light and on active triangulation. The dependence of the measurements on the geometric parameters of the system is shown, as well as the criterion to calibrate the system. From the extensive set or experiments carried out to evaluate the measurement performance, good linearity has been observed, and an overall mean value of the measurement error equal to 40 μm, with a variability of about ±35 μm has been estimated
Homography estimation and heteroscedastic noise-a first order perturbation analysis
  • P Chen
  • D Suter