一种补偿平移与旋转运动的快速电子稳像算法

针对视频图像序列的非稳特性,研究了电子稳像算法中灰度投影算法和块匹配算法各自的不足之处,提出了一种快速补偿视频图像序列间平移及旋转运动的稳定成像算法.该算法先采用灰度投影算法估计并补偿视频图像序列间的平移运动,再利用拉普拉斯变换在靠近图像的边缘区域选取几个具有明显特征的小块,运用块匹配算法进行匹配,计算并补偿其旋转运动量,以得到稳定的视频图像序列.通过理论分析和实验验证,表明这种稳像算法具有速度快、准确度高的特点.     

一种基于检测块动态选择的稳像算法

针对包含运动物体视频序列的稳像问题,提出了一种基于运动目标检测的数字稳像算法.该算法运用估计的全局运动参数补偿帧间全局运动,得到运动补偿前帧和当前帧的差值图像,并将差值图像分割成一系列的前景块和背景块,仅跟踪背景块来估计下一轮的全局运动参数.仿真结果表明,提出的稳像算法不仅计算复杂度低,而且估计的全局变换参数具有相当高的帧间变换保真度,可以用于占据半个像面的动目标的视频序列.     

一种基于特征块匹配的电子稳像方法

提出一种基于特征模块匹配的图像稳定的算法.该算法自适应地搜索图像特征模块进行后续稳像操作.在稳像过程中,提出一种图像特征匹配的快速算法,匹配得到局部运动矢量,然后对匹配结果进行一致性检验,最后采用可变窗口宽度的均值滤波法对图像序列运动轨迹滤波.试验表明,该算法有效地稳定了图像序列,并且满足实时性要求.     

基于两对互相垂直平行线的自定标方法

提出了一种利用图像中的互相垂直的两对平行线来实现相机的自定标方法.利用模约束可以提供一个方程,图像中的两对互相垂直的平行线可以提供两个线性方程的特性,该方法通过两幅图像就可以实现相机的自定标.所提出的相机定标方法在求解过程中是准线性的,实验数据结果表明,该定标方法具有鲁棒性好、收敛性好等优点.     

分数傅里叶域数字水印算法

基于离散分数傅里叶变换(DFRFT)快速算法发展了一种分数傅里叶谱域图像水印算法.该算法根据分数傅里叶变换谱具有空域和频域双域信息表达能力,分别对原始图像和所加水印信息进行不同级次的分数傅里叶变换,提取水印分数傅里叶谱的低频成分并直接将其叠加到原始图像的分数傅里叶谱中的对角像元上,然后再进行逆变换得到水印图像.在JPEG压缩、图像旋转及剪切等攻击方式下,对该水印算法进行了鲁棒性分析,数值实验表明该水印算法具有良好的抗攻击性和安全性.     

基于运动补偿和码率预分配的干涉多光谱图像压缩算法

提出了一种基于运动补偿的三维小波变换和基于码块预测的码率预分配的图像压缩算法.利用干涉多光谱图像成像推扫平移特性,在小波变换中使用运动补偿来减少帧间相关性,并对图像组中各个图像小波变换和量化后EBCOT编码码块的有效比特平面进行独立的熵估计.以图像估计熵总和指导整个图像组码率预分配,以解决平均分配码率对重建图像质量带来的影响.实验结果表明:该算法在8倍压缩时,图像序列的平均峰值信噪比比3D-SPIHT提高了0.85～1.25 dB,比单帧JPEG2000提高了1.91～4.25 dB, 算法复杂度低,易于硬件实现.     

基于双边滤波和双树复小波的图像去噪算法

提出了基于空域双边滤波和双树复小波变换的图像去噪算法.该算法使用双树复小波变换对含噪图像进行多尺度和多方向的分解,对各个高频方向子带使用带有方向窗的局部维纳滤波算法进行去噪.在重构过程中,对每一个尺度上重构得到的低通图像使用空域的双边滤波算法进一步的去除噪声.实验结果表明本文提出的图像去噪算法获得了明显的去噪性能改善.     

基于场景的红外焦平面阵列非均匀性评价方法

针对红外遥感图像非均匀性的定量评价问题,研究了基于场景的红外焦平面阵列响应非均匀性的定量评价方法。利用真实红外遥感图像序列量值分布的特点,针对小序列图像提出了基于序列排序和小波变换的标准图像构建算法;结合非均匀性评价指标,提出了利用构建标准图像的响应非均匀性对红外图像的非均匀性进行定量估计;分别利用黑体定标实验和统计方法对标准图像构建算法进行了验证,同时进行了不同校正算法的剩余非均匀性评价实验。结果表明,128frame序列构建得到的标准图像的非均匀性分别比同温黑体图像小0.92%,比统计标准图像大0.59%。该方法可用于定量估计红外遥感图像的非均匀性。     

基于双边滤波和双树复小波的图像去噪算法

提出了基于空域双边滤波和双树复小波变换的图像去噪算法.该算法使用双树复小波变换对含噪图像进行多尺度和多方向的分解,对各个高频方向子带使用带有方向窗的局部维纳滤波算法进行去噪.在重构过程中,对每一个尺度上重构得到的低通图像使用空域的双边滤波算法进一步的去除噪声.实验结果表明本文提出的图像去噪算法获得了明显的去噪性能改善.     

基于相位相关的柱面全景图像自动拼接算法

提出了一种基于相位相关法的柱面全景图像自动拼接算法.首先推导了柱面投影公式、柱面平移量和平面平移量之间的对应公式;然后利用改进的相位相关法计算出归一化相位相关度的峰值及峰值点坐标,利用峰值作为相关度准则判断两幅图像是否存在重叠部分,利用最大相关度求交自动确定头尾图像,利用峰值点坐标判断两幅相邻网像的位置关系;最后将环绕拍摄的序列图像投影到柱面坐标系下,根据计算得到的变换参数进行图像拼接并融合得到柱面全景图像.实验结果表明,该算法能有效地实现顺序混乱的序列图像的自动排序和柱面全景图像拼接,具有较高的稳健性和拼接精度.     

TDI CCD立体相机内方位元素测试技术研究

依据星载单镜头TDICCD立体相机立体成像及测绘要求,结合普通测绘相机内方位元素,提出了星载单镜头立体相机内方位元素由焦距、主点位置、立体成像角(前后视交角)、图像畸变等元素组成。给出了立体相机内方位元素测试的方法,针对立体相机成像特点给出了计算内方位元素的算法,对其进行了实际测量和验证,其测量结果符合设计要求。     

Robust geometric,phase and colour structured light projection system calibration

This paper introduces a new comprehensive procedure for both geometric and colour calibration of structured light system. In order to perform both geometric and colour calibration procedure, a new calibration artifact is proposed. The intrinsic and extrinsic parameters of projector and camera are estimated by using an extended pinhole camera model with a tangential and radial distortion. Camera image plane coordinates are obtained by extracting features from images of a calibration artifact. Projector image plane coordinates are calculated on the basis of continuous phase maps obtained from a fringe pattern phase reconstruction procedure. In order to stereo calibrate camera-projector system, pairs of corresponding image plane points are calculated with subpixel accuracy. In addition, one of three pattern views is used in colour calibration. RGB values of a colour field pattern detected by camera and their reference values are compared. This comparison leads to derivation of a colour transformation matrix. The performance of the proposed method is tested by measuring plane, sphere and distance reference. Also 360 degrees complex object 3D model from a set of measurements is obtained. Residual mean errors for all tests performed are calculated.     

单目移动机器人相对位姿估计方法

针对单目机器人在运动过程中，相邻关键帧之间基线过短造成的相机相对位姿无法恢复的问题，提出了一种利用已知特定参照物迭代优化本质矩阵，从而快速分解出机器人相对位姿的方法。该方法首先利用已知特定参照信息获取本质矩阵估计值，然后通过相邻关键帧之间匹配的特征点组和本质矩阵估计值对本质矩阵迭代优化，最后通过本质矩阵的约束关系求得平移向量估计值，利用基于李群的姿态表示得到旋转矩阵估计值，并进一步优化得到唯一解。该方法避免了由于基线过短而无法恢复相机相对位姿的弊端。实验结果表明，该方法可快速解算并优化机器人相对位姿关系，且相对位移估计绝对误差小于0.03 m，优于传统方法。     

中心折反射全方位摄像头的标定方法研究

提出一种适用于不同类型中心折反射全方位摄像头的标定方法。该方法不需标定模板的运动信息或限定全方位摄像头的类型,而只用通过摄像头观测分布在不同位置的标定模板,对所采集的标定模板图像提取角点,将包含内外部参数的非线性图像投影函数用泰勒级数展开表示,应用Levenberg-Marquardt算法计算摄像头外部参数,而通过对超定方程组的伪逆求解计算内部参数。整个标定过程简单,快速而且自动完成。利用某一中心折反射全方位摄像头对提出的方法进行了实验,实验结果表明,该方法可获得很好的标定效果。     

二次JPEG压缩对噪音相关性的影响分析

为了探索二次JPEG压缩对数字图像与数码相机的噪音相关性的影响,提高利用噪音相关性鉴别数字图像来源的正确性,降低鉴别过程的复杂性,利用高品质的JPEG格式图像进行JPEG压缩实验,从噪音相关性的分布、噪音相关性的大小、鉴别阈值和错误拒绝率四个角度分析了JPEG压缩对噪音相关性的影响.实验结果表明,JPEG压缩会改变噪音相关性的分布特性,噪音相关性与品质因素是二次曲线关系,当品质因素小于90时,鉴别阈值取平均值,错误拒绝率不会产生明显的误差,从而找到了一种利用噪音相关性鉴别二次JPEG压缩图像来源的新捷径.     

Real-time camera pose estimation based on volleyball court view

The use of technology in sports has increased in recent years. One of the most influential of these technologies is referee support systems. Team sports such as volleyball require accurate and robust tracking systems that do not affect either the players or the court. This paper introduces the application of intrinsic and extrinsic camera calibration in a 12-camera volleyball referee system. Intrinsic parameters are calculated by using the classic pinhole model and Zhang’s method. To perform extrinsic calibration in real time, the volleyball court is treated as a global calibration artifact. Calibration keypoints are defined as court-line intersections. In addition, a new keypoint detection algorithm is proposed. It enables achievement of an accurate camera pose in regard to the court. With all 12 cameras calibrated in a common coordinate system, a dynamic camera stereo pair creation is possible. Therefore, with known ball 2D image coordinates, the 3D real ball coordinates can be reconstructed and the ball trajectory can be estimated. The performance of the proposed method is tested on a synthetic data set, including 3Ds Max rendering and real data scenarios. The mean camera pose error calculated for data biased with keypoint detection errors is approximately equal to 0.013% of the measurement volume. For the real data experiment with a human hand phantom, it is possible to determine the presence of the human phantom on the basis of the ball reflection attitude.     

A novel method of camera pose estimation by parabolic motion

Camera pose estimation is a basic and crucial problem in computer vision, accordingly a novel method is proposed for pose estimation based on parabolic motion in our paper. Firstly, the intersection of lines which are the image plane projection of free-falling trajectories in different locations is computed. According to the properties of vanishing point, the intersection is defined as the vanishing point in gravity direction. Secondly, the image plane projected curve of parabolic trajectory is obtained by Sampson Approximation. Finally, the camera pose is estimated by employing the projective geometry properties of vanishing point and vanishing line implicated in the projected parabola, provided that the intrinsic parameters of camera are specified. The absolute Euclidean distance of translation is obtained innovatively with the known frame frequency. Numerical simulation as well as practical experiment in this paper demonstrates the correctness and feasibility of our method, with the known frame frequency, as the experiment show that compared with the traditional checkerboard method the mean errors of rotation axis, rotation angle and translation are respectively 0.017 rad, 0.007 rad and 11.650 mm by our method. It can generally satisfy the accuracy requirements of camera.     

Research on camera calibration using new optimization strategy

An important task for stereo vision is camera calibration, whose goal is to obtain the intrinsic and extrinsic parameters of each camera. This paper proposes a new accurate calibration method with multilevel process of camera parameters. In order to improve the calibration accuracy, a sub-pixel corner detection method is presented. We start with several views of a planar calibration to obtain some intrinsic camera parameters and to build an accurate model with lens distortion on a planar calibration target. Flexibly making use of geometry imaging theory, our algorithm obtains all the parameters through logical organization of solving order, accordingly avoids obtaining possible local optimized problem when solving the non-linear equation, gets over the relativity influence of every unknown parameters of traditional calibration way, and makes the error distributed among the constraint relation of parameters. Experiments with real images are carried out to verify the image correction effect and numerical robustness of our results. Compared with classical calibration techniques, that use expensive equipment and complicated mathematical computation, the proposed technique, which was verified by experiment, achieves high accuracy and reliable parameters.     

基于直线的射影不变性和极线约束标定摄像机参数

提出一种用图像对标定摄像机参数的新方法。传统自标定方法需要先由图像特征计算基础矩阵,该方法则是从图像中提取一组未知长度和长度比的平行直线,以图像中心坐标为主点初值,利用直线的射影不变性先求得摄像机焦距、旋转矩阵和平移向量的初值;再由图像对固有的极线约束来优化摄像机的参数,从而得到高精度的主点、焦距、旋转矩阵和平移向量。使用该方法对图像目标进行三维重建,仿真实验表明：该方法对场景要求较少,原理简单,方法易行,标定摄像机参数精度较高。     

基于无穷单应的大视场摄像机标定方法

在大视场摄像机标定中,常常会出现由于场景过于单一而很难达到自标定所需要的场景约束和运动约束条件、立体标定所需要的强立体条件或者平面靶板标定所需要的绝对共面条件,如指向高空区域的摄像机标定任务就很难满足上述要求,因而大视场摄像机标定需要较为弹性的标定算法。提出一种基于无穷单应的大视场摄像机标定方法,该方法最少只需要4个非共线控制点和摄像机粗略的位置即可求解无穷单应,并且提出一种坐标变换方法以保证线性求解和优化无穷单应时的稳定性。从无穷单应中分解得到摄像机参数初始值,通过Levenberg-Marquardt(LM)优化算法最终实现摄像机的标定。在优化过程中,通过假设图像中心为主点和采用一阶径向畸变模型,相对增加了优化过程中的剩余自由度,能够实现4个像点为观测值的参数优化。相比于强立体或共面的条件,此方法所需条件很容易满足。仿真和实际实验验证了此方法的正确性和高精度,以及重复测量实验的灵活易实施。