Implementation of opto-digital stereo object tracking system

In this paper, a new opto-digital stereo object tracking system using the variable window mask and the optical binary phase extraction joint transform correlator (BPEJTC) is proposed. At the first step, with the distance information from the stereo camera to the tracking object easily acquired by the structural elements of a stereo vision system, the area of the tracking object can be digitally extracted by using the variable window mask. And, at the second step, by carrying out the optical BPEJTC between this reference image obtained from the variable window mask and the stereo input image, the coordinates of the tracking object's location can be acquired, and then with these values, the convergence angle and the pan/tilt of the stereo tracking camera can be finally controlled. From some experimental results, the proposed system is found to be able to effectively extract the area of the target object from the input image having the background noises by using the variable window mask. And, with the location values of the tracking object obtained by using the optical BPEJTC, the convergence angle and the pan/tilt of the stereo cameras can be controlled. Finally, a feasibility of real-time implementation of the adaptive stereo object tracking system using the proposed algorithm is also suggested.     

A calibration method for stereo vision sensor with large FOV based on 1D targets

Large FOV (field of view) stereo vision sensor is of great importance in the measurement of large free-form surface. Before using it, the intrinsic and structure parameters of cameras should be calibrated. Traditional methods are mainly based on planar or 3D targets, which are usually expensive and difficult to manufacture especially for large dimension ones. Compared to that the method proposed in this paper is based on 1D (one dimensional) targets, which are easy to operate and with high efficiency. First two 1D targets with multiple feature points are placed randomly, and the cameras acquire multiple images of the targets from different angles of view. With the fixed angle between vectors defined by the two 1D targets we can establish the objective function with intrinsic parameters, which can be later solved by the optimization method. Then the stereo vision sensor with two calibrated cameras is set up, which acquire multiple images of another 1D target with two feature points in unrestrained motion. The initial values of the structure parameters are estimated by the linear method for the known distance between two feature points on the 1D target, while the optimal ones and intrinsic parameters of the stereo vision sensor are estimated with non-linear optimization method by establishing the minimizing function involving all the parameters. The experimental results show that the measurement precision of the stereo vision sensor is 0.046 mm with the working distance of about 3500 mm and the measurement scale of about 4000 mm×3000 mm. The method in this paper is proved suitable for calibration of stereo vision sensor of large-scale measurement field for its easy operation and high efficiency.     

Development of a vision-based lane detection system considering configuration aspects

Vision-based lane-sensing systems require accurate and robust sensing performance in lane detection. Besides, there exists trade-off between the computational burden and processor cost, which should be considered for implementing the systems in passenger cars. In this paper, a stereo vision-based lane detection system considering sensor configuration aspects such as field of view (FOV), span pixels, resolution, etc is developed. An inverse perspective mapping method is formulated based on the relative correspondence between the left and right cameras so that the 3D road geometry can be reconstructed in a robust manner. The selection rule of the sensor configuration and specifications is investigated for a standard highway. Based on the selected sensor configurations, it is shown that sensing region range on the camera image coordinate can be determined for the best lane-sensing performance. The proposed system is implemented on a passenger car and its real-time sensing performance is verified experimentally.     

Application of stereo vision to the study of mixed-mode crack-tip deformations

A stereo vision is applied to evaluate crack-tip parameters for fracture specimens subjected to a mixed-mode loading (tension and shear). By using a special loading device, the applied remote loading is oriented at an angle with respect to the axis of the crack. At each loading angle, the calibrated vision system was rotated so that the axis of the crack is parallel to the horizontal or vertical axis of the image frame. At a load close to the crack initiation, the displacement field around the crack-tip region of the fracture specimen was measured relative to a specimen coordinate system located at the crack tip of the fracture specimen. During the experiment, the fracture specimen was subjected to rigid body translation and rotation. Hence, the displacement fields are affected by the rigid body translation and rotation. Using the experimentally determined displacements and the analytically determined displacements with several higher order terms being included, the stress intensity factors and the amount of rigid body translation and rotation were calculated through a least-squares fit. The effect of the rigid body motion on the measured displacements was then eliminated using the computed rigid body translation and rotation. Experimental results indicate that a K_I and K_II dominant region is observed in the corrected displacement fields.     

逆向工程中基于共轴立体视觉的曲面边界测量

针对逆向工程中引导性曲面边界信息的快速获取问题,系统地研究了共轴立体视觉测量方法,建立该方法的数学模型,详细分析了摄像机焦距、基线距等系统结构参数及被测点空间位置对测量精度的影响,通过数学分析确定摄像机基线距的最佳取值范围,研究共轴立体视觉测量系统特殊的极线几何关系.提出基于共轴立体视觉的曲面边界快速测量方法,利用三坐标测量机的精密机械系统及精确的空间定位能力,用单个摄像机以两次共轴定位摄取图像的方式实现共轴立体视觉测量功能,然后利用共轴立体视觉外极线相互平行且通过各自像平面主点的特殊极线几何关系简化同源像点匹配过程,从而快速获取被测曲面的边界信息.实验结果表明:用基于三坐标测量机的单摄像机共轴立体视觉测量方法获取的曲面边界平均误差为0.268 mm,基本满足逆向工程中对引导性曲面边界的测量精度要求.     

立体视觉测量中的图像匹配策略研究

立体视觉测量广泛应用于机器人导向和各种柔性制造系统中 ,其中图像特征的准确匹配始终是一个难题。推导了共极性线方程 ,采取相关匹配和特征匹配相结合的方法来提高图像匹配的速度和精度 ,利用建模估计得到匹配像点的精确坐标 ,使三维测量精度达到目标距离的 1/10 0 0 0左右     

An optimized in vivo multiple-baseline stereo imaging system for skin wrinkles

Recent developments in computer vision have attempted to mimic human visual physiology. One application of this technology is the evaluation of skin wrinkles. We have developed a non-convergence stereo system that can be calibrated in vivo and can be controlled to the level of microns for baseline. We are able to obtain more accurate 3D information by calibrating nonlinear interrelations between the disparity of object and depth information.     

基于脉冲耦合神经网络的弹簧卡箍缺陷检测

传统的弹簧卡箍缺陷多为产后人工全检,存在漏检与缺陷率上升等现象,这不但会使成本上升、也对人力资源提出了考验。为此实现自动实时在线全检就成为急需解决的课题,设计了基于机器视觉的弹簧卡箍在线自动检测系统,该系统安装在弹簧卡箍流水线两侧,搭建特定光源,通过激光传感器外部触发工业相机对其表面进行图像捕获,送上位机进行缺陷判定与定位,最后通过RS485将判定结果送下位机来控制剔除机制。实验结果显示：该系统采用改进的脉冲神经网络(PCNN)能准确提取目标缺陷区域并对缺陷进行判定,可在0.348 s每个零件的速度下,检测出弹簧卡箍表面大于10像素的缺陷。通过对不同弹簧卡箍进行检测验证实验,证明了PCNN算法对缺陷分割的准确性和有效性。     

双通道回声抵消系统中改进算法的定点化实现

本文研究了双通道回声抵消系统的三个核心模块:频域多延时自适应滤波器算法(MDF),双端说话检测算法(DTD),以及残留回声抑制算法(RES)。针对频域算法的特点提出了改进的双端说话检测算法和基于维纳滤波与谱减法的残留回声抑制算法,然后以上述三种算法为核心模块实现双通道回声抵消系统,并对系统定点化以便在定点DSP处理器上实时实现,分析并解决了定点化的精度问题所带来的影响。     

带钢表面质量在线检测系统研究与设计

为了提高现有带钢表面质量检测技术在缺陷检测精度与识别率上存在的问题,设计了基于机器视觉技术的带钢表面质量自动检测系统,从系统整体构成、视觉传感系统、软件开发及检测与分类算法等方面进行了深入地研究,实现了包括图像的采集、传输、缺陷的实时检测和定位,缺陷的分类以及缺陷的存储与报警等功能。实验结果表明,本系统可以对带钢表面常见的边裂、氧化、结疤等几十种不同类型的缺陷进行精确地检测。与现有的缺陷检测技术相比,本系统中设计的算法在检测精度和实时吞吐量上都具有很大优势。     

三维机器视觉中图象匹配的几种摄象机几何模型和误差分析

三维视觉中的一个主要问题是从不同的角度来确定物体的几何位置,而物体的位置则可由两幅实体图象的视差来确定.本文重点讨论图象匹配中几种摄象机的几何模型和误差分析.     

Precise perimeter measurement for 3D object with a binocular stereo vision measurement system

A novel measurement scheme for a three dimensional (3D) object's surface boundary perimeter is proposed. This scheme consists of three steps. First, a binocular stereo vision measurement system with two CCD cameras is devised to obtain the two images of a detected object's 3D surface boundary. Second, two B-spline active contours are applied to converge to the object's contour edges accurately in the two CCD images to perform the stereo matching. Finally, for the reconstructed 3D active contour, its true contour length is computed as the detected object's true boundary perimeter. An experiment on a bent surface's perimeter measurement indicates that this scheme's measurement repetition error decreases to 0.6%.     

一种双目立体视觉系统的误差分析方法

基于摄像机透视成像的针孔模型,分析了立体视觉中摄像机标定和三维重建过程的主要误差来源。基于各主要误差源的模型分析,建立了双目视觉系统3D测量误差与摄像机参数、基线长度、测量距离等因素之间的关系式。如已知相关参数,可以估算出双目视觉系统的3D测量精度,或根据3D测量精度要求,初步确定摄像机的内部各项参数和基线长度、测量距离等参数。     

Sphere-based calibration method for trinocular vision sensor

A new method to calibrate a trinocular vision sensor is proposed and two main tasks are finished in this paper, i.e. to determine the transformation matrix between each two cameras and the trifocal tensor of the trinocular vision sensor. A flexible sphere target with several spherical circles is designed. As the isotropy of a sphere, trifocal tensor of the three cameras can be determined exactly from the feature on the sphere target. Then the fundamental matrix between each two cameras can be obtained. Easily, compatible rotation matrix and translation matrix can be deduced base on the singular value decomposition of the fundamental matrix. In our proposed calibration method, image points are not requested one-to-one correspondence. When image points locates in the same feature are obtained, the transformation matrix between each two cameras with the trifocal tensor of trinocular vision sensor can be determined. Experiment results show that the proposed calibration method can obtain precise results, including measurement and matching results. The root mean square error of distance is 0.026 mm with regard to the view field of about 200×200 mm and the feature matching of three images is strict. As a sphere projection is not concerned with its orientation, the calibration method is robust and with an easy operation. Moreover, our calibration method also provides a new approach to obtain the trifocal tensor.     

无人机视频运动目标实时检测及跟踪

对无人机拍摄视频中的地面运动目标提出一种实时检测跟踪算法。该算法利用特征点的对应关系将图像对配准,再对配准图像进行变化检测,根据变化和运动信息检测目标并消除虚警,将检测与跟踪相结合,对目标跟踪失效的情况,能重新正确定位目标,从而获取目标的完整运动轨迹。采用无人机拍摄的地面车辆图像的测试结果表明,算法能有效检测跟踪运动目标,并能达到25 f/s以上的实时处理速度。     

成像型激光探测系统动态范围扩展技术研究

 针对成像型激光探测系统动态范围不足的问题,以现有广角远心鱼眼透镜和CMOS摄像器件为硬件基础,提出通过图像处理的方法扩展系统动态范围。系统采用帧相减技术滤除背景干扰和噪声,进而进行激光光斑的识别和定位。在此基础上,针对接收激光能量的不同,提出采用不同方法扩展动态范围。当激光能量较低时,激光信号可能淹没在噪声中无法准确判读和精确定位,通过滤波技术排除噪声,提高图像信噪比,从而提高系统探测灵敏度,即扩展探测下限。对于探测上限,即当强激光入射时,由于鱼眼光学系统内会产生反射干扰光,有可能在成像面形成不规则的干扰光斑,通过光斑匹配模式识别的方法排除杂光干扰,提高识别上限。实验结果表明,该方法使系统动态范围从为70 dB提高到89 dB,提高近30%,有效扩大了探测系统的作用范围。     

光电成像探测中目标方位测量方法

为了在光电成像探测目标的同时给出目标相对探测器的方位信息,探讨了一种基于单目视觉的运动目标方位测量方法。根据投射成像原理,推算出二维目标像点坐标与目标的空间三维位置之间的映射关系,建立了单目视觉测量目标方位的数学模型。结合帧间差分法和KLT方法检测静止背景下和变化背景下的运动目标,并对目标特征点进行亚像素定位。试验和仿真结果分析表明,该方法能够有效提取目标的特征点,可对目标较为准确地定位,并通过单目视觉测量出目标方位信息,误差控制在8%的范围内。     

基于双目视觉的边缘检测算法

提出了一种基于立体视觉的物体边缘检测的方法。先对立体图像对进行基于图割的立体匹配方法求取场景的视差图，然后再用Canny的边缘检测方法对视差图进行边缘检测。立体视觉方法有效解决了单目视觉检测方法中的一些难点，利用了物体在空间的深度信息，对复杂背景下的物体和具有复杂纹理物体的边缘检测有很高的鲁棒性。实验结果表明该边缘检测方法优于传统的单目视觉边缘检测方法。     

高灵敏度空间点目标探测系统设计

本文面向空间点目标探测,设计了基于高灵敏度CMOS传感器的空间点目标探测系统。首先对CMOS传感器图像进行降噪,提高传感器的探测灵敏度;其次,采用DSP+FPGA嵌入式架构,设计了基于星图匹配信息构建的点目标探测算法,并详细介绍了算法原理和步骤。最后,采用电子星图模拟器对该探测系统进行测试。结果表明:该嵌入式系统具备1 024×1 024@20p格式视频的实时处理能力,可以探测6等星。当信噪比大于6,视轴指向误差小于1°时,对于不同运动速度、不同尺寸点目标均能准确探测,识别正确率接近100%。综合而言,该空间点目标探测方法的计算精度高、适应性强、可靠性高,能够应用于空间点目标的有效探测。     

中药口服液中可见异物检测系统研究

针对目前生产线上中药口服液中可见异物检测系统存在的速度慢、精度低等问题,设计了一种可同时检测多个药瓶的基于机器视觉的异物快速检测系统。介绍了成像系统和机械系统。提出了一种基于边缘轮廓的图像配准算法,有效地消除了药瓶晃动的影响。提出了一种逐点阈值分割图像的二值化方法,能得到质量较好的二值化图像。实验结果表明,该系统能快速、有效地检测出可见异物,能较好地满足在线检测需求。