靶场光测多站图像多目标判读技术

靶场光测图像处理的一个突出难题是如何分辨图像上出现的大量小目标。利用单站序列图像上目标的运动关联、目标在多站图像上像点之间的对极几何约束关系,提出了靶场多站图像多目标判读的新技术,采用单站序列图像目标卡尔曼运动轨迹预测及跟踪识别技术,以及多站图像对极几何约束同名目标匹配技术,突破了传统方法仅进行单站图像判读的局限。仿真结果证明其能有效解决多站、多目标的判读问题,已成功应用于我们研制的多套靶场图像判读系统中。     

一种基于惯量椭圆位姿补偿的扩展目标精确定位方法

针对空间扩展目标成像纹理特征少,背景单一,以及目标旋转、平移、形变等运动特征所导致的目标定位跟踪困难的问题,研究了基于惯量椭圆位姿补偿的目标定位跟踪方法(TTBPC)。通过惯量椭圆测姿法获取目标准确俯仰角姿态,将目标位置、姿态补偿到统一的坐标系下,结合最近距离特征点匹配原则,最终实现对空间扩展目标的精确跟踪定位。通过仿真图像和实际图像对TTBPC方法的有效性进行了验证,平移目标定位误差小于0.26pixel,旋转目标定位误差小于0.28pixel,实际空间扩展目标定位误差小于0.3pixel,这表明该方法应用于空间扩展目标定位的有效性和准确性。     

单站光测图像确定空间目标三维姿态

充分利用空间目标的几何先验知识,提出了一种基于特征点提取的单站经纬仪等光测设备确定空间目标三维姿态的迭代算法,避免了多站图像的立体匹配,并在此基础上用vc++进行了模拟仿真.实验表明:该方法准确度较高,结果稳定,应用价值高.     

靶场目标三维姿态测量结果自动修正方法

中轴线法是靶场试验中常用的目标三维姿态测量方法,但该方法求出的俯仰角和偏航角所描述的目标姿态可能与目标实际情况相异,必须进行人工修正。提出一种自动对姿态判读结果进行修正的方法,将目标共线方程与中轴线方程联立求解,可快速地求出角度修正量。实验证明该方法可对中轴线法求解结果进行自动修正,得到正确的目标三维姿态角,提高了姿态判读的速度和可靠性。     

一种无人值守设备的红外弱小多目标自动判读方法

提出了一种实用的适用于无人值守设备的红外弱小多目标自动判读方法。通过判别图像序列段累积信号量的变化进行目标段自动定位,对定位后的目标段运用轨迹预测关联的方法进行目标识别。给出了目标图像序列的处理结果,证明该自动判读方法可以在保证目标识别率的同时大大缩短图像判读时间,能够满足光电测量中快速提供测量结果的需求。     

基于光学图像的目标姿态判读处理

 采用摄像测量方法获取目标的飞行姿态是目前靶场主要测量手段之一。目标姿态可以用目标轴线在空间的方位来描述,即姿态的判读处理就是要提取线的空间位置。在进行图像判读处理时除采用传统的基于点的判读处理手段外,还应根据不同的目标类型采用不同的判读处理方法,文中给出了基于首尾两点的判读处理、基于线的判读处理和基于模型的三种判读处理方法,有效解决了不同类型目标的判读处理问题。     

基于灭点理论的目标姿态角单站测量

提出了一种基于灭点理论的目标姿态角单站测量方法,利用单台光测设备的参数信息和目标在图像上的灭点位置计算目标特征直线的方向,从而获得目标的姿态信息。同时给出了灭点精度评价指标,可作为姿态测量精度的评价标准和加权融合时的权重因子。通过模拟实验和实物实验验证了方法的可行性和准确性。该方法与中轴线法测量偏差均值为0.2,表明所给出的方法对类柱状目标的姿态测量具有精度高、易于实现的特点。     

基于单站地基望远镜的空间目标姿态估计方法

为实现从单站光测图像中估计出已知3D模型的空间目标姿态,利用Vega Prime提出了一种采用仿真图像进行相关度局部最优搜索的姿态估计方法,该方法无需建立2D-3D特征投影关系和大量的模型匹配库。首先,对输入图像进行图像预处理,获得目标原始图像。然后,利用Vega Prime加载目标3D模型生成仿真匹配图像,并进行图像预处理获得目标匹配图像,计算两幅图像相关度。最后,更新3D目标模型姿态,直至仿真匹配图像与目标原始图像的相关度值取得局部最优,输出目标模型姿态。仿真实验结果表明,采用本文所提方法的观测仿真图像姿态平均估计误差为3.85°,仿真原图可实现姿态准确估计,表明该方法是一种空间目标姿态估计的有效方法。     

从单站光测图像确定空间目标三维姿态

基于在物距与焦距之比远大于 1时摄像系统成像模型可用平行投影近似的事实 ,充分利用空间目标的几何先验知识 ,提出了一种从单站经纬仪等光测设备获得的图像确定火箭等空间目标三维姿态的方法 ,避免了多站图像的立体匹配。仿真实验和实际应用结果表明 :方法精度高 ,结果稳定     

基于中轴线像长匹配的靶场单站姿态测量方法

目前,靶场姿态测量以多台套交会测量为主,对于单站姿态测量尚没有较好的解决方案。为了解决该问题,以投影轴对称目标为例,提出了一种基于中轴线像长匹配的单站姿态测量方法。将透视投影拓展为2种等效形式,将体现目标姿态状态的中轴线向像面透视投影,可获得中轴线的投影像长或在等效物面的等效物长,根据目标中轴线的先验长度、相机内外参数及成像信息,经像长匹配即可获取目标的偏航角和俯仰角;实际工程试验验证了该算法的可行性,偏航角精度为1.7°,俯仰角精度约1°,满足靶场单站测姿需求;并对姿态测量模型关键因素进行了误差分析。该方法可适用于非投影轴对称目标。     

用于相机标定的球靶标投影误差分析与校正

由于球体具有轮廓连续性好等优点,在摄像机标定,尤其是多相机标定方面获得了广泛的应用。利用球作为标定靶标可以弥补平面靶标在多相机标定中出现视角过大时畸变太大甚至于观测不到的不足,但是空间球经透视投影后成像一般并非标准圆,而是一个椭圆。椭圆几何中心与球心真实成像中心并不一致,从而影响了标定精度。造成球心成像误差的因素主要有两个,即球的相对大小及相对于相机的位置。通过分析空间球成像模型,仿真研究了各因素对球心成像误差影响的大小,寻找球心的透视投影像点与其成像椭圆几何中心之间的误差变化规律,并建立了两者之间的误差校正模型,最后通过实验验证了该校正模型的可行性和有效性。通过校正,球心投影像点定位精度可达到亚像素级。     

二元微透镜的位相平衡设计及叠加积分衍射分析

本文用位相平衡设计法设计了优化的二元光学微透镜，这是对Ｄａｍｍａｎｎ阶梯波带片设计方法的一种改进。文中用叠加积分作了二元微透镜的衍射分析，并与Ｄａｍｍａｎｎ法作了比较，计算结果说明本方法的衍射效率及象质都优于Ｄａｍｍａｎｎ法。且能设计更大相对口径（Ｄ／ｆ）。具有任意物象距的二元微透镜。     

用黑斑的椭圆性质在二值化图像中探测眼睛

在人脸探测过程中 ,双眼的探测非常重要 ,这是因为人的眼睛是人脸上最容易探测到的特征部位。通过双眼探测还可以对人脸进行定位 ,使人脸的尺寸归一化。探讨了一种基于二值化图像中黑斑几何形状分析的双眼探测方法。因为人眼的形状比较接近椭圆 ,故在研究黑斑的几何性质时 ,重点研究黑斑的椭圆性质。所谓黑斑的椭圆性质是指按照黑斑的面积和二阶矩计算出的椭圆参数。椭圆的主要参数有椭圆的面积、质心、长短轴的长度和方向角 (椭圆长轴与水平方向的夹角 )。为了描述黑斑与椭圆的近似程度 ,定义了椭圆硬度的概念。根据眼睛黑斑椭圆参数的测试数据建立了探测眼睛的数学模型 ,并给出了根据此模型得到的双眼探测的实验结果     

Measurement of diameter of cylindrical openings using a disk beam probe

The diameter of cylindrical openings is conventionally measured with the mechanical and contact method. In this paper, we propose a contactless optical approach to measure dimensions of inner profile by a disk beam probe. The measurement is carried out illuminating the inner cylindrical opening with a light sheet generated by the disk beam probe, acquiring the light-section image of the object with an image sensor through a widefield lens, then numerically extracting the diameter. During measurement, the probe is aligned inside the openings with high freedom, which causes the optical section to manifest a variety of ellipses besides a circle accordingly. The measurement method copes with extracting the diameter of openings from various ellipse optical sections. The high freedom of probe alignment allows optical diameter measurement to be flexible and simple. The dimensions of cylindrical openings of manufactured parts were practically measured with the disk beam probe, and the results were confirmed with those measured with the vernier caliper.     

On-axis diffuse directional screen based on a transmission Gabor hologram

Holographic single-component diffuse screens based on on-axis transmission holograms are studied theoretically and experimentally. The configurations of extra holographically reconstructed images of the diffuse screen (conjugate image, halo, and zeroth-order spot) are analyzed. A so-called centered hologram is suggested in which the conjugate image of the diffuse screen spatially coincides with the main image, thereby eliminating distortions inserted by the conjugate image. Also, hologram recording using a Mach-Zender interferometer is suggested and implemented. Such an approach makes it possible to eliminate the shadow of an object on the hologram. It is noted that, in recording a planar diffuse screen, the presence of the halo does not distort the image of an object projected through the screen: only the light intensity is partially lost. Analysis of the images projected shows that the luminance of the zeroth-order spot, while relatively low, should be diminished in one way or another.     

Design of object surveillance system based on enhanced fish-eye lens

A new method is proposed for the object surveillance system based on the enhanced fish-eye lens and the high speed digital signal processor （DSP）. The improved fish-eye lens images an ellipse picture on the charge-coupled device （CCD） surface, which increases both the utilization rate of the 4：3 rectangular CCD and the imaging resolution, and remains the view angle of 183° The algorithm of auto-adapted renewal background subtraction （ARBS） is also explored to extract the object from the monitoring image. The experimental result shows that the ARBS algorithm has high anti-jamming ability and high resolution, leading to excellent object detecting ability from the enhanced elliptical fish-eye image under varies environments. This system has potential applications in different security monitoring fields due to its wide monitoring space, simple structure, working stability, and reliability.     

Improvement of the optical image reconstruction based on multiplexed quantum ghost images

Ghost imaging allows one to obtain information on an object from the spatial correlation function between photons propagating through or reflected from the object and photons of the reference arm. In this case, detection in the object arm is performed over the entire aperture of the beam and, therefore, it does not give information on the object. The reference beam does not interact with the object, but is recorded with a scanning point detector or a CCD array permitting the measurement of the spatial correlation function of photons in two arms. The use of multimode entangled quantum light beams by illuminating the object by one beam and orienting other beams to reference arms makes it possible to obtain simultaneously several ghost images (GIs). Cross correlations of multiplexed GIs (MGIs) are determined by eighth-order field correlation functions. A special algorithm is developed for calculating higher-order correlations of Bose operators. The presence of GI cross correlations is used for improving the quality of the reconstructed object’s image by their processing using the measurement reduction method. An example of the computer simulation of the image reconstruction by MGIs formed in the field of four-frequency entangled quantum states is considered. It is found that in this case the reduced GI has a signal-to-noise ratio several times higher than that of GIs.     

Phase-shifting digital holography with burst imaging method

The application of a burst imaging method to digital holography (DH) with multiple image capturing is firstly demonstrated. Nobody has discussed this approach explicitly, although the burst imaging method is not so uncommon. The burst imaging method intermittently captures several images with a high repetition rate followed by a relatively long rest time. If the light wave from an object has a much smaller complex amplitude change in the frame interval of an image sensor than the precision of the DH system, the wave can be regarded as being in a static state. As a result, this solves a fatal problem in DH for an object that moves or deforms while capturing multiple interference images. In this research, the burst imaging method was applied to phase-shifting DH, and it is simply implemented by a high-speed image sensor and a continuous phase shifting which is good for high-speed phase shifting. The measurement error was analytically solved for the movement speed of an object. Finally, the proposed method was experimentally used to observe an evaporating droplet of ethanol on a glass substrate.

     

近红外波段硅基金属光子晶体平板的亚波长成像特性

利用时域有限差分方法,理论上研究了由正方形金属嵌入背景材料硅中组成的二维正方晶格和三角晶格光子晶体的亚波长成像特性。采用Drude模型描述金属银的色散特性,在近红外波段该模型可以很好地描述金属的实际介电常数。通过结构参数的设计,在上述两种结构中实现了波长在1550nm附近的亚波长成像,并且发现金属对入射光的吸收使得成像位置处的光强稍有降低,但是对于光子晶体亚波长成像的质量并无影响。相对于通常的硅基空气孔型光子晶体亚波长成像器件,该种硅基金属型全固态光子晶体结构更加稳定,因而可以更好地在复杂全光集成回路中加以实际应用。     

兰姆波结构健康监测中的概率损伤成像

应用概率成像方法对兰姆波结构健康监测中板的损伤进行识别。根据兰姆波损伤散射信号的传播时间以及传感器网络中一对发射-接收传感器的空间位置来确定一个椭圆轨迹,该轨迹显示了损伤可能出现的位置。将监测区域均匀网格化,计算各网格节点到椭圆轨迹的最短距离,将此距离映射为损伤出现的概率。采用灰度等级对所有网格节点处的概率值进行量化,则由每一个发射接收传感器对就确定了一个灰度图像。为了突出损伤,应用图像融合技术对传感器网络中所有发射接收传感器对所构成的灰度图像进行融合。对铝板中横穿孔损伤的实验结果显示了该方法能够有效地确定损伤的位置和范围,有助于推动兰姆波结构健康监测的实用化。