A Survey of Calibration Methods for Traditional Cameras Based on Line Structure Light

During the measurement procedure of line structured light 3D scanning system, there is a key about how to calibrate the whole measuring system. In order to solve the calibration problem in the line structured light 3D measured system such as requirements for high accuracy calibration models, complicate calibration procedures and so on. This paper presented a calibration method combining the vertical calibration board with 3D scanning measurement system, this method is simple, easy to operate and the assistant adjustment equipment is unnecessary. Experiment showed that this method can attain relative accuracy about 0.51% which indicates the rationality of this method. Introduction At present, with the development of science and technology, the intelligent non-contact 3D measurement methods have gained wide attention , the line structured light scanning system is applying to precise measurement, based on computer vision technology with reliable, inexpressive and anti-interference. Its processing principle is the deformed of line laser stripe due to the height adjustment of the target object. Capturing the distortion of light pattern, and obtaining the dimensional information of the object surface based on the position relationship between the laser and camera coordinate system. Therefore, the 3D measurement system parameter calibration is one of the key loop of entire detection tasks. Normally, the calibrate methods of structure sensor is drawing method, toothed target calibration method, mechanical adjustment method and so forth. The first two methods both require precise expensive auxiliary equipments, and not suitable for field calibration, besides that, the mechanical adjustment for the factors involved in too many legal, making measurement precision is lower, besides it was proposed based on the calibration reference coplanar method, but this method is to create a three-dimensional world coordinate system in the plane of reference, each time the reference position moves, the positional relationship between the world coordinate system and camera coordinate system must be readjusted [1]. In response to these problems, a vertical scale grid Othello plate is designed as a calibration target system, and improving the traditional calibration methods. Result shows that the improved method can obtain the plane equation of camera coordinate system only single measurement, and the calibration of coordinate system can also be obtained easily. This method doesn’t require expensive precision equipment , simple operation and is suitable for field calibration. Parameter Calibration of Line Structured Light Scanning System Calibration of line structured light 3D scanning system includes two loops, first is the camera spatial location model parameter calibration and the second is the calibration of line structured light plane. (1)Calibration of camera model parameter Camera parameters can be obtained based on the principle of coplanar, non-coplanar and self-calibration. Accounting for this, besides taking into account the simplicity and accuracy of line structured light plane. Article designs the vertical calibration plate as shown in figure 1, and the International Conference on Intelligent Systems Research and Mechatronics Engineering (ISRME 2015) © 2015. The authors Published by Atlantis Press 65 corner of vertical calibration plate also can be extracted as the calibration points needed in experiment, as shown in figure 2. Fig.1 .Vertical calibration chessboard captured by camera Fig.2. Corner extraction of vertical by calibration board Tsai camera calibration based on the radial constraint and Zhang Zhengyou calibration based on planar target are mainly methods of camera calibration[2]. Considering the camera in experiment is general, only the first-order radial distortion is considered in the process of camera calibration, this paper chooses the Tsai calibration method. Firstly most of camera model parameters are solved by the radial collimator constraints, then the effective focal length, nonlinear distortion factor and other parameters are obtained so that the final completion of calibration of the camera[3,4]. (2)Parameter calibration of line structured light plane The design of line structured light scanning system is driven by a stepping motor line lasers, cameras and a rotating platform composed of the entire scanning system platform shown in figure 3.firstly the laser projects the line structure light beam to the target, the deformation line structured light strip can be obtained and captured due to the height modulate of measured object. Among the deformation stripes the location information between line laser and camera and depth data of the measured surface is included, and the three-dimensional information of target object can be obtained based on the cross-ratio invariance principle, after scanning the whole surface using stepper motor the entire three-dimensional contour data can be obtained completely[5]. Fig. 3. 3D scanning prototype used in this experiment

Camera calibration is the premise and important step of servo control. In the process of optimization of camera calibration parameters, this paper establishes a nonlinear model including the camera internal parameters, the external parameters and distortion coefficients, each parameter is optimized by iterative steps, it avoids the coupling among all parameters, and the efficiency of calibration is obviously enhanced. It is easier to get pose information from 2D image signal to 3D. the optimization model is established with the minimum re-projection error. Under the Matlab simulation, the result shows that the non-linear model calibration algorithm improves the calibration precision compared with the traditional calibration algorithm. Servoing control of robot visual is more accurately with the proposed method.

The 3D measurement system based on line-structured light uses a camera to capture laser stripes due to changing in the shape of an object, and uses the acquired pixel coordinates for 3D reconstruction. System calibration is an important step in 3D measurement. The current camera calibration algorithm research mainly focuses on improving the algorithm itself, and there is less research on the influence of external factors. This paper proposes a coplanar hybrid calibration algorithm based on the error screening model by combining the error screening model, mathematical model and neural network model. It is mainly divided into two steps. The first step is to use the radial array constraint calibration algorithm based on the error screening model to solve the camera’s internal and external parameters. The second step uses the camera internal and external parameters obtained in the first step to convert the pixel coordinates into real three-dimensional coordinates, and compares the calculated three-dimensional coordinates with the actual coordinates. Using machine learning to establish a compensation network, get a compensation function, and use the resulting 3D world coordinates to perform point cloud stitching. Experiments show that compared with the traditional calibration algorithm, the calibration algorithm has a small error and reduces the calibration error by about 6.5%.

Camera calibration plays an important role in the field of machine vision and photogrammetry, and among the practical calibration methods, the one proposed by Zhang ZhengYou is higher accuracy and easily operated. However, this method needs to move the camera (or the planar target) to get three or more target images at different locations, and it is better to uniformly fill the calibration target in whole measurement volume to improve the calibrate precision. But manual movement and placement of the targets frequently will increase the difficulty in guaranteeing the uniform distribution of target. In view of this situation and according to the linear imaging model of the camera, a new camera calibration method based on the virtual planar targets is proposed in this paper. A liquid crystal display was used as a target plane, and the 2D target graphics were displayed on this LCD screen. Using TSai's camera calibration method to get initial parameters, a serial of images of the virtual planar targets in different positions were captured with keeping the display position unchanged and are used to calculate the internal and external parameters of the camera by classic Zhang's camera calibration method, and the new internal and external parameters would again guide the movement of virtual target. After several iterations, camera parameters can be obtained with high precision. The presented method is flexible and easy to operate, and it has been applied to calibrate different cameras and an actual 3D shape measurement system in our Lab. The comparison results of the transverse coordinates in plane calculated by this method and by Zhang's camera calibration method shows that this proposed method is quite accurate and reliable.

CCD camera calibration technique is an important part of image processing. The calibration accuracy of CCD camera parameters determines the precision of the image processing system. With the development of industry, the demand for image processing precision is increasing, so the precision and robustness of CCD camera calibration technique are required to be higher. In order to further improving camera calibration accuracy and simplifying calibration process, in this paper, we proposed a new method of camera calibration on the basis of the traditional camera calibration technique. Different from the traditional calibration method, this new calibration technique uses a “black box system” to build a link between the ideal image points and actual ones instead of setting up linear CCD camera imaging model (i.e. pinhole imaging model) and nonlinear imaging model about camera distortion. The actual pixel coordinates of the camera image are the input values, and the reprojection points obtained by the world coordinates are the output values. In order to verify the validity of this method, we carry out experiments based on the improved calibration method of zhang zhengyou. 1 And finally we verified the feasibility of this method by comparing the experimental results respectively obtained though carrying out the calibration method this paper proposed and the traditional calibration method by camera calibration toolbox.

We propose a novel method for automatic camera calibration and foot-head homology estimation by observing persons standing at several positions in the camera field of view. We demonstrate that human body can be considered as a calibration target thus avoiding special calibration objects or manually established fiducial points. First, by assuming roughly parallel human poses we derive a new constraint which allows to formulate the calibration of internal and external camera parameters as a Quadratic Eigenvalue Problem. Secondly, we couple the calibration with an improved effective integral contour based human detector and use 3D projected models to capture a large variety of person and camera mutual positions. The resulting camera auto-calibration method is very robust and efficient, and thus well suited for surveillance applications where the camera calibration process cannot use special calibration targets and must be simple.

Camera calibration is an essential research field with a high potential for emerging algorithms and calibration targets. In this work, we present a virtual experiments approach in Gazebo simulator for exhaustive camera calibration methods and calibration targets evaluation. Key steps of the camera calibration workflow were adapted to the Gazebo simulation approach. We proposed a virtual camera calibration evaluation pipeline that includes camera modeling and calibration target’s pose generation in a viewing frustum. Experiments in the Gazebo demonstrated virtual environment feasibility for camera calibration evaluation while comparing checkerboard and circle grid targets, and allowed to achieve an acceleration of more than 30 times compared to the real-time. Experimental results exhibited a need for additional calibration steps incorporation, such as outliers rejection and optimal calibration target poses generation.

In the vision 3-D measurement field, camera calibration results directly affect the accuracy of measurement. So camera calibration method is hot spot in this field. This paper proposes a new efficient calibration method with binocular camera. First, the method uses an accurate extraction algorithm to extract the ellipse center of calibration target image. Second, the paper presents a automatic matching algorithm based RANSAC (RANSAC SAMPLE CONSENSUS), the automatic matching algorithm is simple, fast, and can complete space calibration feature points matching its image point at once. At the beginning of matching, sort calibration points in the target according to a certain sequence. The sequence of feature points extracted in the calibration image does not correspond with one of the feature points in calibration target, because of the camera distortion. But obtain the correct matching points by the RANSAC algorithm. According to the camera calibration model, use the correct matching points and obtain correct intrinsic and extrinsic parameters of each camera and the relative parameters of two cameras. Finally, through experimental verification, the results show that the calibration method is fast and robust, and has higher calibration accuracy.

Camera calibration using synthetic random speckle pattern and digital image correlation

Large Sky Area Multi-object Fiber Spectroscopic Telescope – LAMOST, with a 1.75m-diameter focal plane on which 4000 optical fibers are arranged, is one of major scientific projects in China. During the surveying process of LAMOST, the optical imaging system makes the astrometric objects be imaged in the focal plane, and the optical fiber positioning system controls the 4000 fibers to be aligned with these objects and obtain their spectrum. In order to correct the positioning error of these optical fibers, the CCD camera is used to detect these fibers’ position in the way of close-range photogrammetry. As we all know, the calibration quality of the CCD camera is one of the most important factors for detection precision. However, the camera calibration has two following problems in the field work of LAMOST. First, the camera parameters are not stable due to the changes of on-site work environment and the vibration during movement. So, the CCD camera must be on-line calibrated. Second, a large-size high-precision calibration target is needed to calibrate the camera, for the focal plane is very big. Making such a calibration target, it is very difficult and costly. Meanwhile, the large calibration target is hard to be fixed on LAMOST because of the space constraint. In this paper, an improved bundle adjustment self-calibration method is proposed to solve the two problems above. The results of experiment indicate that this novel calibration method needs only a few control points while the traditional calibration methods need much more control points to get the same accuracy. So the method could realize the on-line high-precision calibration of CCD camera for LAMOST.

During the manufacturing process of heavy forgings, simple contact measuring techniques are still used to check the dimensions, therefore an optical measuring system is in demand. In this paper, a camera calibration method for the passive measuring system, which is being developed in collaboration with an industrial partner, is proposed. Our approach is based on space resection and works with robust coded targets, which are distributed in the field of view. The coordinates of targets are measured using TRITOP (GOM) measuring system. This solution allows to build a large calibration field, without a need of large calibration objects. The camera calibration works in 2 steps - at first, the intrinsic parameters of the camera, including lens distortion, are calibrated. These parameters are considered as stable, due to the use of special camera covers. Multi-image version of the calibration method and dense field of calibration targets are used. The second step is performed from every image and employs a single-image extrinsic camera parameters calibration method. Only a few coded calibration targets, mounted on stable objects in the scene, are required. The calibration method was tested in industrial conditions. The method showed great results, the average reprojection error was under 0.1 px. The effect of thermally affected zone on the calibration process is discussed.

Accurate and efficient geometric calibration is important to precisely acquire geometric information in thermal cameras. Typically, the calibration procedure involves obtaining defined control points from the images of a calibration target and utilizing them to estimate the geometric calibration parameters based on a camera model. Calibration targets include checkerboards, circle grids, Hermann grids, and specially designed patterns. Thermal cameras typically have lower resolution than visible cameras and the control points are localized by creating contrast through heating. These differences make the calibration process of thermal cameras more complex. Conventional thermal camera calibration methods require that the entire calibration target is captured and that all the control points from each image are identified. Due to these limitations, geometric calibration of a thermal camera traditionally requires advanced image processing or substantial manual intervention.This work presents a calibration method using a ChArUco board as the calibration target. The proposed method allows the calibration to be performed with partial-view images and does not require that all the control points from each image are extracted, thereby allowing the calibration to be performed without requiring complex image processing. The proposed calibration method was successfully implemented to estimate the calibration parameters of two different thermal cameras; in both cases achieving an overall mean reprojection error below 0.4 pixels. The effectiveness of the approach was further demonstrated by performing the calibration with a lower mean reprojection error than the MATLAB camera calibrator app. Furthermore, the effectiveness of using different heat sources to create contrast was evaluated.

A novel algorithm based on nonlinear optimization for parameters calibration of wheeled robot mobile chasses

In the field of three-dimensional (3-D) metrology based on fringe projection profilometry (FPP), accurate camera calibration is an essential task and a primary requirement. In order to improve the accuracy of camera calibration, the calibration board or calibration target needs to be manufactured with high accuracy, and the marker points in calibration image require to be positioned with high accuracy. This paper presents an improved camera calibration method by simultaneously optimizing the camera parameters and target geometry. Specifically, a set of regularly distributed target markers with rich coded concentric ring pattern is first displayed on a liquid crystal display (LCD) screen. Then, the sub-pixel edges of all coded bands radial straight lines are automatically located at several positions of the LCD screen. Finally, the sub-pixel edge point set is mapped into parameter space to form a line set, and the intersection of the lines is defined as the center pixel coordinates of each target point to complete the camera calibration. The simulation and experimental results verify that the proposed camera calibration method is feasible and easy to operate, which can essentially eliminate the perspective transformation error to improve the accuracy of camera parameters and target geometry.

Traditional multi-camera calibration is usually on one side of the calibration board. If there are several cameras distributing on both sides of the calibration board, each camera needs to be calibrated on each side. Thus, the error accumulated from each calibration will have a non-ignorable impact on the results. In this paper, we presents a method that using transparent glass board to replace ordinary calibration board based on Zhang Zhengyou’s calibration method. We also introduce two new parameters (depth and refractive index of glass). Therefore, calibration of several cameras can be completed by one experiment. The calibration of camera in the front of the transparent glass board uses traditional calibration method and LM iterative algorithm to obtain the internal and external camera parameters. The calibration of camera on back side uses the traditional calibration model and the light refraction model, and the results without considering the refraction condition are served as the initial values. Th...