基于视差信息的三维图像的计算机重构*

为了解决全景成像技术中观察者位于观察区域之外看到的图像会存在失真的问题,提出了一种基于视差信息的计算机重构3D视图技术.利用3D场景中的物体点经过不同微透镜在元素图像中记录的视差信息,根据光学路径分析,对重构视图中的失真部分用其它元素图像中存在的同一物体点的匹配像素进行替代,从而得到无失真的3D视图.该技术能够在更宽的观察区域内产生3D图像.     

基于多视差函数拟合的集成成像深度提取方法EI北大核心CSCD

集成成像系统采用微透镜阵列记录空间场景信息,只有一次成像过程,在获取场景深度时,不存在相机的参数设置和校准问题,简化了深度测量过程,具有很大优势。但是由于该系统采集的元素图像分辨率低,容易受记录噪声影响,元素图像之间匹配困难。基于集成成像系统能够计算重构出多幅视图的优势,充分利用多幅视图提供的冗余信息来改善立体匹配情况。采用颜色相似性的自适应窗口在参考视图与其他视图之间进行立体匹配,得到参考视图与其他视图之间的视差信息,利用视差与视图对之间的抽样距离满足正比函数关系的性质,对多个视差值进行函数拟合,减少了由于记录噪声等因素产生的误匹配,物体深度层次的分辨更加准确。     

基于多视差函数拟合的集成成像深度提取方法

集成成像系统采用微透镜阵列记录空间场景信息,只有一次成像过程,在获取场景深度时,不存在相机的参数设置和校准问题,简化了深度测量过程,具有很大优势。但是由于该系统采集的元素图像分辨率低,容易受记录噪声影响,元素图像之间匹配困难。基于集成成像系统能够计算重构出多幅视图的优势,充分利用多幅视图提供的冗余信息来改善立体匹配情况。采用颜色相似性的自适应窗口在参考视图与其他视图之间进行立体匹配,得到参考视图与其他视图之间的视差信息,利用视差与视图对之间的抽样距离满足正比函数关系的性质,对多个视差值进行函数拟合,减少了由于记录噪声等因素产生的误匹配,物体深度层次的分辨更加准确。     

利用匹配区域的纹理特征改善重构三维图像的视觉质量

研究了计算机重构三维图像时分辨率低的问题,提出一种改善3D计算机全景重构图像的视觉质量的方法,该方法利用3D空间的物体部分在每个元素图像中形成的匹配区域的纹理特征,从两个相邻的元素图像中的匹配区域提取出多个像素,经过加权计算重构出相应的图像区域.该方法与传统的计算机重构方法相比,提高了图像分辨率,同时也消除了从每个元素图像中提取多个像素直接重构图像时存在的"像素块"效应,改善了重构图像的视觉质量.     

基于立体像素匹配的图像重构技术研究

为了解决目前全景成像技术中分辨率低的问题,提出了一种新的基于3D场景立体像素光线映射的全景图像计算机重构技术.在全景成像技术中,3D场景的每个立体像素点经全景成像系统的编码系统分别映射在一定区域的多个体元素图像的不同像素点上.在计算机重构全景图像时,根据逆光学路径原理,提出了从立体像素映射到的体元素图像区域中提取对应立体像素的多个2D像素点来重构全景图像,使重构的全景图像最大分辨率可达到传统成像方法图像分辨率的N倍（N为映射区域面积）.提出的立体像素的匹配技术大大提高了重构的计算机全景图像分辨率.     

基于光场相机的全视差三维显示

针对光场相机采集的三维信息在显示时只能呈现非常窄的视差问题,对光场相机信息记录过程进行了分析,提出一种将光场数据转化为元素图像阵列(EIA),并对EIA中的物体进行重新采样以调整深度的全视差三维显示方法。为了验证该方法的有效性,使用光场相机采集真实三维物体,利用所提方法对光场图像进行处理后,将其在深度优先集成成像(DPII)系统中进行显示。实验结果表明,利用所提方法可以实现光场数据的全视差三维显示。     

基于机器视觉的三维灰度重构技术

基于双目立体视觉三维重构原理，采用主动扫描实现特征匹配的三维灰度重构技术，通过特征结构光扫描物体表面，由经过预标定的两套成像传感器拍摄其图像，经过图像处理程序提取出特征点，完成特征匹配，再计算出物体表面三维轮廓，同时将成像传感器中的灰度信息映射到相应的特征点，从而实现特征点的三维信息和颜色信息的重构和匹配。该技术为三维彩色逼真场景的重构奠定了基础。     

应用于多视点视频编码的基于深度特征的图像区域分割与区域视差估计

基于多视点视图深度特征,提出一种通过简单块匹配运算划分多视点视图区域并估计区域视差的算法.首先基于深度对象的概念确定图像中具有不同深度的区域数量以及这些区域对应的区域视差,再根据误差最小化准则初步确定每个图像块所属区域.当区域中图像块数量小于某个阈值时,采用区域合并算法将该区域中的每个图像块合并到与它的视差最为接近的其它图像区域,通过迭代形成最终的有效图像区域划分.实验表明,该算法能够以图像块为基本单元有效地划分各深度层区域,并准确估计对应的区域视差.     

利用空间光调制器实现计算全息三维显示

针对目前三维物体计算全息算法数据量大、计算速度慢及共轭像影响再现效果问题,提出一种全息体视图计算方法。根据人眼双目视差立体视觉原理,由摄像机获取三维物体不同角度的二维序列视差图像,通过计算全息算法得到多视角全息图,合成三维物体全息体视图。在迭代傅里叶变换算法的基础上,采用预设初始相位并增加反馈因子的方法,提高相息图的计算效率。基于液晶空间光调制器构建光学系统,对计算的全息体视图进行了光学再现。结果表明:该方法有效地排除了共轭像的干扰,相息图的迭代计算效率提高30%以上,再现图像与目标图像的结构相似度大于0.85。     

一种基于图像融合的无参考立体图像质量评价方法

提出了一种基于图像融合的立体图像质量评价方法。通过对立体图像的左右视图进行图像融合生成一幅彩色图像,融合算法采用主成分分析(PCA),使用归一化互相关(NCC)视差图算法,生成了对应的视差图;对融合图像和视差图分别进行归一化亮度系数和谱能量参数的提取,作为支持向量回归(SVR)的输入数据,在经过充分的训练后对立体图像的质量评分进行预测。在LIVE 3D立体图像数据库上的实验结果表明,提出的算法优于最新的无参考立体图像质量评价方法,与人类的主观评价具有较好的一致性。     

Three-dimensional free view reconstruction in axially distributed image sensing

In this Letter, we propose a three-dimensional(3D) free view reconstruction technique in axially distributed image sensing(ADS). In typical integral imaging, free view reconstructed images can be obtained by tilting all elemental images or tilting the reconstruction plane due to large lateral perspectives for 3D objects. In conventional ADS, the reconstructed images at only a front view can be generated since the sensor is moved along with its optical axis so that it has small lateral perspectives for 3D objects. However, the reconstructed 3D images at any viewing point may be obtained because the virtual viewing camera may capture these slightly different perspectives for 3D objects. Therefore, in this Letter, we employ the virtual viewing camera to visualize the 3D images at the arbitrary viewing point. To support our proposed method, we show the experimental results.     

Optical 3D watermark based digital image watermarking for telemedicine

Region of interest (ROI) of a medical image is an area including important diagnostic information and must be stored without any distortion. This algorithm for application of watermarking technique for non-ROI of the medical image preserving ROI. The paper presents a 3D watermark based medical image watermarking scheme. In this paper, a 3D watermark object is first decomposed into 2D elemental image array (EIA) by a lenslet array, and then the 2D elemental image array data is embedded into the host image. The watermark extraction process is an inverse process of embedding. The extracted EIA through the computational integral imaging reconstruction (CIIR) technique, the 3D watermark can be reconstructed. Because the EIA is composed of a number of elemental images possesses their own perspectives of a 3D watermark object. Even though the embedded watermark data badly damaged, the 3D virtual watermark can be successfully reconstructed. Furthermore, using CAT with various rule number parameters, it is possible to get many channels for embedding. So our method can recover the weak point having only one transform plane in traditional watermarking methods. The effectiveness of the proposed watermarking scheme is demonstrated with the aid of experimental results.     

Occlusion removal technique for improved recognition of partially occluded 3D objects in computational integral imaging

In this paper, we propose an occlusion removal technique for improved recognition of 3D objects that are partially occluded in computational integral imaging (CII). In the reconstruction process of a 3D object which is partially occluded by other objects, occlusion degrades the resolution of reconstructed 3D images and thus this affects negatively the recognition of a 3D object in CII. To overcome this problem, we introduce a method to eliminate occluding objects in elemental image array (EIA) and the proposed method is applied to 3D object recognition by use of CII. To our best knowledge, this is the first time to remove occlusion in CII. In our method, we apply the elemental image to sub-image (ES) transform to EIA obtained by a pickup process and those sub-images are employed for occlusion removal. After the transformation, we correlate those sub-images with a reference sub-image to locate occluding objects and then we eliminate the objects. The inverse ES transform provides a modified EIA. Actually, the modified EIA is considered to be an EIA without the object that occludes the object to be reconstructed. This can provide a substantial gain in terms of the image quality of 3D objects and in terms of recognition performance. To verify the usefulness of the proposed technique, some experimental results are carried out and the results are presented.     

Three-dimensional polarimetric integral imaging

A three-dimensional (3D) polarimetric image sensing and display technique based on integral imaging is proposed. Three-dimensional polarization distribution of reflected light from a 3D object can be measured as elemental image arrays by a rotating linear polarizer. After the measurement of the polarization of the 3D object, the 3D polarimetric object can be reconstructed optically by displaying the polarization-selected elemental images in spatial light modulators with two quarter-wave plates. Experimental demonstration of 3D polarimetric imaging of a 3D object attached to two orthogonal linear polarizers is presented. To the best of our knowledge, this is the first report on 3D polarimetric sensing imaging and 3D optical reconstruction by integral imaging.     

基于传像束的小视场双目结构光三维形貌恢复

提出基于传像束传像和双目结构光相结合的三维测量系统,实现对无法直接成像场景,如爆轰、冲击等环境下的小视场物体三维面形测量。测量时,结构光直接投影到待测物体表面,变形条纹图像首先被成像到传像束前端面,再被传递到传像束后端面,并最终被成像到相机像平面,再由三频相位展开技术计算出对应相位分布后,即可根据双目匹配计算出待测物体高度。实验表明,传像束系统能完成高精度的微小物体三维形貌重建,且由于传像束的引入,增加了测量系统的灵活性。     

A novel three-dimensional digital watermarking scheme basing on integral imaging

A novel integral imaging-based three-dimensional (3D) digital watermarking scheme is presented. In the proposed method, an elemental image array (EIA) obtained by recording the rays coming from a 3D object through a pinhole array in the integral imaging system is employed as a new 3D watermark. The EIA is composed of a number of small elemental images having their own perspectives of a 3D object, and from this recorded EIA various depth-dependent 3D object images can be reconstructed by using the computational integral imaging reconstruction (CIIR) technique. This 3D property of the EIA watermark can make a robust reconstruction of the watermark image available even though there are some data losses in the embedded watermark by attacks. To show the robustness of the proposed scheme against attacks, some experiments are carried out and the results are discussed.     

Resolution-enhanced 3D image correlator using computationally reconstructed integral images

In this paper, a novel resolution-enhanced three-dimensional (3D) image correlator using the computationally reconstructed integral images is proposed in order to extract target object’s 3D location data in a scene. Elemental images of the reference and target objects are picked up by lenslet arrays and using these elemental images, reference and target plane images are reconstructed on the output plane by means of a modified computational integral imaging reconstruction technique. Then, through cross-correlations between the reconstructed reference and the target plane images, 3D location data of the target object can be extracted from the correlation outputs. With the purpose of showing the feasibility of the proposed method, some computational and optical experiments on the target objects in space are carried out and the results are presented.