镜头畸变及其校正技术

根据镜头畸变的数学模型,利用平面圆形网格模板的网格交点来求取镜头的径向、切向和薄棱镜畸变系数。利用网格交点及其附近的像素点信息,来拟合解析曲面并将所求曲面顶点坐标作为网格交点坐标,使像素位置坐标达到亚像素精度,事先求出镜头畸变系数,并在实际测量中对成像点加以校正的方法,可在不影响处理速度的前提下,提高了求解精度。     

一种基于模型采样的镀膜双镜头多光谱相机成像畸变校正方法

针对传统畸变校正方法无法完成镀膜双镜头多光谱相机影像畸变校正的问题,提出了一种基于数字畸变模型采样的单波段影像畸变校正方法。采用多片后方交会完成相机检校得到各类参数,建立起原始影像的数字畸变模型;对原始数字畸变模型采样得到各波段数字畸变模型;利用各波段数字畸变模型完成对应波段影像畸变的校正。实验结果表明,数字畸变模型采样法能够有效解决镀膜双镜头多光谱相机单波段影像畸变的校正问题,校正精度较高;通过进一步实验证明,经畸变校正后的单波段影像配准误差达到亚像素级,数据可用性高。     

航空变焦距镜头非线性畸变快速校正方法EI北大核心CSCD

针对航空变焦距镜头的非线性畸变随焦距变化而变化的问题,提出一种地面离线标定和机上在线校正相结合的快速校正方法。利用单参数除式模型校正镜头畸变,根据模板图像中共线点的投影不变性,采用变步长优化搜索方法求解出若干离散焦距下镜头的畸变系数和畸变中心坐标,分析了畸变参数随焦距变化的规律,建立了畸变参数与焦距之间的经验公式。在飞行实验中,将实际工作焦距值代入经验公式得到相应畸变参数对实景图像进行自动校正。对模板图像与实景图像的校正结果表明,该方法能有效校正变焦距镜头的非线性畸变,对3幅不同焦距下的720 pixel×576 pixel模板图像校正平均均方差约为2.68 pixel,平均校正时间约为4.82 s。该方法具有效率高,便于自动化实现和工程应用的优点。     

用射影不变性纠正鱼眼镜头畸变

计算机视觉通常采用针孔摄像机模型，但对于存在较大畸变的鱼眼镜头或广角镜头，必须首先纠正镜头畸变才能应用针孔模型。为了从同时存在透视变形和鱼眼畸变的图像中纠正鱼眼畸变，采用了以下射影不变性：空间中直线的投影为直线；平行直线束的投影平行或相交于一点；直线段投影的交比不变。首先，用待纠正的鱼眼镜头对一幅等间隔正交网格图成像，提取网格结点作为控制点，然后用射影不变性求解畸变模型，达到纠正畸变的目的。该算法巧妙地运用三次方程的Cardan解法。大大提高了速度。结果表明，运用该方法纠正鱼眼镜头畸变速度快、精度高。     

一种基于畸变等效曲面的图像畸变校正

介绍了一种图像畸变的等效曲面模型,用于修正由于光学系统或图像传感器产生的畸变。分析了球形畸变模型的原理,推导出基于球形模型的畸变校正公式,并给出了基于球形模型的畸变校正的具体实现方法,通过实验对这种畸变校正的效果进行了验证。这种曲面畸变模型的校正具有实现简单,不需求解方程组等优点,并可以推广到其他曲面,具有比较广泛的实用性。     

大视场短焦距镜头CCD摄像系统的畸变校正

从光学测量角度出发,结合计算机视觉中的摄像机标定方法,解决了大视场短焦距镜头CCD摄像系统的畸变校正问题。与摄像机标定不同,畸变校正中仅标定内部参数,外部参数作为已知条件。采用线性畸变模型,由最小二乘法解线性方程组得到摄像系统畸变模型的畸变系数。介绍了数字图像中像素间距和光学中心的标定方法。通过比较由标定参数得到的畸变图像和摄像机采集的畸变图像对实验标定精度进行评定,实验结果表明边缘视场(112°)的标定精度达到了0 75%。     

图像显示系统几何畸变的测量及校正

为了保证图像显示系统能够产生120°的大视场,在系统中使用了超广角耦合物镜,这样就不可避免地存在几何畸变。提出了一种基于点物成像原理,并采用数码相机和精密单轴转台进行畸变测量的方法,介绍了测量原理和测量过程,根据测量后得到的畸变规律,采用数字图像处理的方法对几何畸变进行了校正。校正后,图像显示系统的畸变小于0.4%,完全能够满足导弹景象匹配系统定位误差及定位概率的检测要求。     

书籍扫描图像畸变模型的可视化描述及应用

利用设备无关显示技术在计算机屏幕上显示一条仿真曲线、一个矩形和一个箭头,可视化地描述在扫描时书籍页边的变形情况、页面上的均匀条带及成像系统对称中心的位置。通过调整仿真曲线,使之较好地反映页边实际变形,并使矩形的大小和位置及箭头的位置合适,可对存在严重畸变的书籍扫描图像进行较好的校正,而无需用户精确地测量这些随具体情况而变化的参数。校正后,用OCR(Optical Character Recognition)软件进行文字识别时,总的平均纠错率为94.5%,页平均纠错率最高达99.6%。用TWAIN(Technology Without An Interesting Name)标准直接从扫描仪读取图像数据,并用椭圆拟合方法方便地测量扫描仪的成像畸变系数。     

PMD光学元件面形测量中的镜头畸变校正研究

相位测量偏折术(PMD)是近几年在光学测量领域内普遍使用的一种非接触式的高精度测量方法,该方法需要CCD相机拍摄经被测光学元件反射的在显示屏上显示的条纹图,而CCD自身存在的镜头畸变会对测量精度产生一定的影响。为避免这一影响,提出了在梯形畸变和镜头畸变同时存在的情况下保留梯形形状而只校正镜头畸变的矢量Zernike多项式校正方法。该方法首先利用光轴与被拍摄面的交点及相机和被拍摄面的相对位置来求取与光轴垂直的辅助面上的标准图,然后利用矢量Zernike多项式拟合标准图与畸变图的坐标得到二者的映射关系,接着运用得到的映射关系对畸变图进行校正。实验结果表明:提出的畸变校正方法可以有效地降低测量误差,提高测量精度。     

基于畸变率的图像几何校正

大视场成像光学系统中的畸变会降低图像质量，必须预以校正。提出一种新的校正方法，即根据畸变率的定义推导出畸变校正公式。根据公式，在镜头畸变率已知的情况下可以很容易地校正畸变。对于畸变率未知的情况，给出了建立畸变模型的方法，通过畸变模型可近似计算畸变率。得出通过控制畸变模型中某一个形状的参数可以控制畸变量大小的结论。提出的方法已经在实际工程中采用。实践证明，这种模型可以满足大多数镜头的畸变校正要求。     

Compensation for reconstructed image distortion using camera model with lens distortion in incoherent digital holography

Optical Review - In incoherent digital holography (IDH), the reconstructed image may be distorted owing to lens aberration, which degrades image quality. A conventional camera lens is designed such...     

Real-time distortion correction of fish-eye lens based on Bayer image signal

The imaging system based on a fish-eye lens generally has to correct the distortion of fish-eye images. The distortion correction based on the Bayer image signal is valuable, such as reducing the computation burden of image signal processing chips and providing a new imaging system structure of fish-eye lens. In this paper, a distortion correction method of fish-eye lens based on the Bayer image signal is proposed. Firstly, a distortion correction method that focuses on vertical straight lines and processing delay is proposed. Secondly, according to the correlation among color channels of the Bayer image, a novel Hermite interpolation method appropriate for Bayer image signal is proposed. Finally, a prototype system of fish-eye-lens-based imaging is established and the real-time field-programmable gate array (FPGA) implementation of the proposed method is demonstrated. The experiment demonstrates that the proposed distortion correction is not only characteristic of real-time processing and the smaller computation amount, but also applicable to embedded hardware.     

Line-based camera calibration with lens distortion correction from a single image

Camera calibration is a fundamental and important step in many machine vision applications. For some practical situations, computing camera parameters from merely a single image is becoming increasingly feasible and significant. However, the existing single view based calibration methods have various disadvantages such as ignoring lens distortion, requiring some prior knowledge or special calibration environment, and so on. To address these issues, we propose a line-based camera calibration method with lens distortion correction from a single image using three squares with unknown length. Initially, the radial distortion coefficients are obtained through a non-linear optimization process which is isolated from the pin-hole model calibration, and the detected distorted lines of all the squares are corrected simultaneously. Subsequently, the corresponding lines used for homography estimation are normalized to avoid the specific unstable case, and the intrinsic parameters are calculated from the sufficient restrictions provided by vectors of homography matrix. To evaluate the performance of the proposed method, both simulative and real experiments have been carried out and the results show that the proposed method is robust under general conditions and it achieves comparable measurement accuracy in contrast with the traditional multiple view based calibration method using 2D chessboard target.     

Using the camera pin-hole model restrictions to calibrate the lens distortion model

Camera calibration required the computation of camera pin-hole and lens distortion models. The lens distortion is estimated alone or together with the pin-hole model, by using some existing lens distortion non-metric or self-calibration methods. If both models are computed together, then the models are adjusted to training data, but not to real camera. This is because both pin-hole and lens distortion models are coupled. If they are computed separately, difficulties arise since calibration of lens distortion alone is an unstable process. To improve existing camera calibration methods, this paper proposes a metric calibration method to compute lens distortion separately from the pin-hole model. This method is solved under stable conditions, independently of the computed lens distortion model, since pin-hole and distortion models are computed separately. Images of a planar template are used. First, using distorted control points extracted from images, a set of undistorted points which fits in the pin-hole model are computed. Second, with distorted and undistorted control points, lens distortion is calibrated by using a metric calibration process.     

Correcting non-linear lens distortion in cameras without using a model

Camera lens distortion calibration is the first step in resolving any metric application with a camera. To date, lens distortion was corrected using some existing lens distortion non-metric or self-calibration methods. Using a lens distortion model means defining a global rule to correct the entire image. This global rule does not take into account particular lens distortion effects not represented by the model. Moreover, to calibrate the model, only some features of the scene such as straight lines, circles or vanishing points are used. Since only the feature of the scene used to calibrate the model is guaranteed by the distortion rectification, it is certain that the model will not be precise. The result is an approximation of the real image distortion.To improve the lens distortion rectification, a method without using a model is proposed. Using a set of control points distributed across the entire image, they are corrected to assure all the restrictions of the scene. With both sets of points, the points detected in the image and the undistorted ones, image local transformations are defined considering only nearby control points. Rather than calibrating a global model, local functions are characterized. The distortion correction is defined by a rectification surface composed of local surface patches each influenced by nearby control points. This method is more sensitive to local deformations and allows the image to be corrected in accordance with its distortion.     

Continuous image distortion by astrophysical thick lenses

Image distortion due to weak gravitational lensing is examined using a non-perturbative method of integrating the geodesic deviation and optical scalar equations along the null geodesics connecting the observer to a distant source. The method we develop continuously changes the shape of the pencil of rays from the source to the observer with no reference to lens planes in astrophysically relevant scenarios. We compare the projected area and the ratio of semi-major to semi-minor axes of the observed elliptical image shape for circular sources from the continuous, thick-lens method with the commonly assumed thin-lens approximation. We find that for truncated singular isothermal sphere and NFW models of realistic galaxy clusters, the commonly used thin-lens approximation is accurate to better than 1 part in 10⁴ in predicting the image area and axes ratios. For asymmetric thick lenses consisting of two massive clusters separated along the line of sight in redshift up to Δz = 0.2, we find that modeling the image distortion as two clusters in a single lens plane does not produce relative errors in image area or axes ratio more than 0.5%.     

用于光纤传像束的大视场低畸变物镜设计

在大截面传像束前置光学系统设计中,采用负-正型式的像方远心光路结构,在像差校正过程中引入偶次非球面和弯月形厚透镜,可较好地解决镜头轴外像差校正与像面照度均匀性问题,同时使镜头结构紧凑、小型化。通过理论计算和优化,给出工作在660 nm,焦距f=1.22 mm,相对孔径D/f=1∶3,视场角2为60的镜头设计实例。设计结果表明:该镜头各视场在40lp/mm空间频率处的MTF值超过0.8,全视场畸变小于0.05%,场曲低于50m,像面照度均匀,像质优良,满足像方远心光路要求,可用于实际观察和测量。     

Memory-efficient algorithm for lens distortion compensation

Distortion introduced by a lens in a measurement system based on an image sensor usually must be compensated. The memory used for distortion compensation by a lookup table is proportional to the image sensor size. To reduce the memory usage, a compression algorithm is proposed and implemented.     

Improving motion estimation by accounting for local image distortion

Cardiac elastography is a useful diagnostic technique for detection of heart function abnormalities, based on analysis of echocardiograms. The analysis of the regional heart motion allows assessing the extent of myocardial ischemia and infarction. In this paper, a new two-stage algorithm for cardiac motion estimation is proposed, where the data is taken from a sequence of 2D echocardiograms. The method combines the advantages of block-matching and optical flow techniques. The first stage employs a standard block-matching algorithm (sum of absolute differences) to provide a displacement estimate with accuracy of up to one pixel. At the second stage, this estimate is corrected by estimating the parameters of a local image transform within a test window. The parameters of the image transform are estimated in the least-square sense. In order to account for typical heart motions, like contraction/expansion, translation and rotation, a local affine model is assumed within the test window. The accuracy of the new algorithm is evaluated using a sequence of 500 grayscale B-mode images, which are generated as distorted, but known copies of an original ROI, taken from a real echocardiogram. The accuracy of the motion estimation is expressed in terms of errors: maximum absolute error, root-mean-square error, average error and standard deviation. The errors of the proposed algorithm are compared with these of the known block-matching technique with cross-correlation and interpolation in the sub-pixel space. Statistical analysis of the errors shows that the proposed algorithm provides more accurate estimates of the heart motion than the cross-correlation technique with interpolation in the sub-pixel space.