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.     

基于元素图像阵列采样的集成成像重构方法

高质量的集成成像系统需要采集和存储大量图像数据。提出了一种新的计算机重构方法,可以减少集成成像系统显示端在计算机重构时所需要使用的数据量。传统的计算机重构方法要利用每个元素图像,而此方法对CCD相机采集的元素图像阵列进行周期采样,仅利用被采样的元素图像即可重构出多视点图像。这种方法充分利用集成成像中邻近的元素图像中具有匹配像素的特点,将未被采样的元素图像信息由邻近的被采样的元素图像中的匹配像素替代,从而减少了重构过程中使用的数据量,这给不同性能的集成成像显示端带来很大的灵活性,降低了对存储空间和传输带宽的要求。     

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.     

一种自由曲面光学视觉重构方法研究

通过对物体摄像来提取表面 3D信息正成为一种新兴的反求建模方法。传统的视觉算法是基于特征对应的 ,要求物体表面具有几何特征或色彩纹理 ,不适于无纹理自由曲面的重构。将计算机视觉方法与CAGD技术紧密结合 ,对无纹理自由曲面提出了一种新的视觉重构方法。首先将物体表面分割为N边域曲面片 ,经由边缘线提取、匹配 ,重构出曲面片的边界曲线 ;然后通过反射模型、图像光亮度分析求出曲面片的跨界切矢 ;最后综合上述信息 ,构造出光滑的曲面模型     

基于广义正交迭代算法的立体视觉定位

提出了一种新的基于广义正交迭代算法的立体视觉定位.该算法通过提取CenSurE局部特征和相应的U-SURF描述符,采用SAD方法进行子像素立体匹配,并利用U-SURF描述符匹配进行前后帧图像特征跟踪.在RANSAC框架下对匹配点进行3D-3D运动估计获得了运动参量的初始值.由于3D-3D运动估计使3D点集间欧式距离误差...     

基于GoogLeNet和空间谱变换的高光谱图像超分辨率方法

为了提高高光谱图像的空间分辨率,提出了一种基于GoogLeNet和空间谱变换的高光谱图像超分辨率(SR)方法.设计出遥感图像的光谱SR框架,对图像中不同反射光谱进行提取;采用GoogLeNet的稀疏编码对粗像素光谱进行放大,并投影到高分辨率字典上,将潜在SR表示进行反转,以获得超分辨光谱;为了提高图像重构的保真度,利用...     

零件的角点提取及匹配定位

本文基于三坐标测量机(CMM)设计了一套视觉检测系统,该系统能够对零件实际空间特征信息进行比较全面地提取。针对位于CMM平台上带有角点的零件,利用Harris算子提取从CMM三个不同方位获取的零件图像的角点。对于Harris算子提取到的角点,本文提出一种八链码搜索法和SUSAN区域法相结合的伪角点剔除方法,最后基于立体视觉的原理,提出"距离空间图"匹配算法,将以上3幅图像一一建立匹配关系。实验中多次改变零件在CMM中姿态时,多次实验数据表明本文的角点提取精度与真实角点间仅存在1~2像素的偏差,零件的定位误差为1~3 mm。通过实验验证,角点匹配和定位的稳定性和精度满足要求,具有一定的抗干扰性和实用性。     

彩色栅线动态三维测量中自适应相关匹配技术

彩色栅线结构光三维测量可仅靠左右摄像机单幅图像完成像素级相位匹配,能够实现运动或者变形物体型面重构,但其测量精度有待进一步提高.在依据相位进行粗匹配的基础上,采用立体视觉数字相关技术在左右图像局部区域进行更细致的匹配.计算相关系数时,依据彩色栅线各个周期的色彩变化特性白适应调整不同颜色通道的权重,依据不同视点表面变形特性自适应调整左右相关窗口的尺寸.在极线方向上双线性插值进行亚像素匹配.实验结果证明自适应数字相关匹配技术在保证彩色栅线动态检测,分辨率高优点的同时,可有效改善测量精度至0.1 mm,扩大测量适用范围.     

基于亚像素区域加权能量特征的多尺度图像融合算法

对矩形和圆形区域中各像素进行亚像素划分,确定各亚像素的权值,得到基于哑像素的综合加权区域能量.融合箅法首先对源图像进行金字塔分解,然后对金字塔的高频细节分量使用基于哑像素加权区域能量特征的融合规则取大,对低频粗糙分量取平均.得到融合图像的塔形分解,最后重构融合图像.仿真结果表明,新算法融合效果较常规的区域能量特征作为融合规则的多分辨率图像融合算法效果更好,从清晰度和熵的评价来看,提高了融合图像的品质.     

Modified computational integral imaging-based double image encryption using fractional Fourier transform

In this paper, we propose an image encryption technique to simultaneously encrypt double or multiple images into one encrypted image using computational integral imaging (CII) and fractional Fourier transform (FrFT). In the encryption, each of the input plane images are located at different positions along a pickup plane, and simultaneously recorded in the form of an elemental image array (EIA) through a lenslet array. The recorded EIA to be encrypted is multiplied by FrFT with two different fractional orders. In order to mitigate the drawbacks of occlusion noise in computational integral imaging reconstruction (CIIR), the plane images can be reconstructed using a modified CIIR technique. To further improve the solution of the reconstructed plane images, a block matching algorithm is also introduced. Numerical simulation results verify the feasibility and effectiveness of the proposed method.     

基于自适应正则化的超分辨率重建方法

传统正则化超分辨重建得到的图像往往存在过度平滑或伪信息残留的问题,结合超分辨重建模型对重建图像伪信息的产生进行了分析,针对传统方法的不足提出了基于图像区域信息自适应的正则化方法,通过图像的区域信息将图像划分为平滑区与非平滑区域,对不同区域选用不同的先验模型进行约束。同时考虑人眼的视觉感知特性,结合区域信息实现正则化参数的自适应选取。实验结果表明该方法在抑制重建图像伪信息的同时能有效保护细节,效果要优于传统方法与单一的先验模型约束,对于红外与可见光图像重建效果的提升提供了一定的理论参考。     

Depth-controlled reconstruction of 3D integral image using synthesized intermediate sub-images

In this paper, we propose a method that controls the depth of the three-dimensional (3D) object existing over the depth-of-focus in integral imaging. The depth control method is performed only in a computer by synthesizing the intermediate sub-images between original sub-images obtained by transforming the captured elemental images. In the reconstruction process, we can obtain reconstructed 3D images with the better image quality within depth-of-focus than that reconstructed over the depth-of-focus. To demonstrate the feasibility of our method, optical and computational experiments are carried out and its results are presented.     

Computational integral imaging reconstruction method of 3D images using pixel-to-pixel mapping and image interpolation

In this paper, we propose a novel computational integral imaging reconstruction (CIIR) method to improve the visual quality of the reconstructed images using a pixel-to-pixel mapping and an interpolation technique. Since an elemental image is magnified inversely through the corresponding pinhole and mapped on the reconstruction output plane based on pinhole-array model in the conventional CIIR method, the visual quality of reconstructed output image (ROI) degrades due to the interference problem between adjacent pixels during the superposition of the magnified elemental images. To avoid this problem, the proposed CIIR method generates dot-pattern ROIs using a pixel-to-pixel mapping and substitutes interpolated values for the empty pixels within the dot-pattern ROIs using an interpolation technique. The interpolated ROIs provides a much improved visual quality compared with the conventional method because of the exact regeneration of pixel positions sampled in the pickup process without interference between pixels. Moreover, it can enable us to reduce a computational cost by eliminating the magnification process used in the conventional CIIR. To confirm the feasibility of the proposed system, some experiments are carried out and the results are presented.     

基于极线局部校正的特征匹配方法

针对利用多视图像进行目标三维测量、结构重建时在极线约束下对直接区域灰度相关进行同名特征匹配常常失效的问题,提出了一种基于极线局部校正的特征匹配算法。介绍了极线约束匹配的原理,分析了相关方法在极线约束匹配中的缺陷以及在多种像机位姿配置下的图像特征间的约束关系,在此基础上提出了一种极线局部区域校正的方法,对待匹配区域进行校正使自动相关匹配能有效执行,结合最小二乘匹配得到了高精度的匹配结果。实验结果证明了新算法的有效性,大大提高了自动匹配的可靠性、速度和精度。     

Light energy matching method in high-resolution image reconstruction

Many approaches for high-resolution image reconstruction have been proposed in some literatures. One of the most commonly ways is to reconstruct a high-resolution image from a number of rotated and translated images with low resolution. In this process, the exposure difference among original images will decrease the quality of the reconstructed image. In order to remove the influence of the exposure difference, a light energy matching method is proposed in this paper. The theoretical analysis is illustrated in details. Experimental results show that the theoretical analysis is correct and the proposed method is valid.     

A 3D image encryption technique using computer-generated integral imaging and cellular automata transform

A novel three-dimensional (3D) image encryption approach by using the computer-generated integral imaging and cellular automata transform (CAT) is proposed, in which, the two-dimensional (2D) elemental image array (EIA) digitally recorded by light rays coming from the 3D image is mapped inversely through the virtual pinhole array according to the ray-tracing theory. Next, the encrypted image is generated by using the 2D CAT scrambling transform for the 2D EIA. The reconstructed process is carried out by using the modified computational integral-imaging reconstruction (CIIR) technique; the depth-dependent plane images are reconstructed on the output plane. The reconstructed 3D image quality of the proposed scheme can be greatly improved, because the proposed encryption scheme carries out in a computer which can avoid the light diffraction caused by optical device CIIR, and solves blur problem caused by CIIR by using the pixel-averaging algorithm. Furthermore, the CAT-based encryption algorithm is an error-free encryption method; CAT as an orthogonal transformation offers considerable simplicity in the calculation of the transform coefficient, that is, it can improve the quality of the reconstructed image by reducing energy loss compared with the traditional complicated transform process. To show the effectiveness of the proposed scheme, we perform computational experiments. Experimental results show that the proposed scheme outperforms conventional encryption methods.     

The infrared and visible image fusion algorithm based on target separation and sparse representation

Although the fused image of the infrared and visible image takes advantage of their complementary, the artifact of infrared targets and vague edges seriously interfere the fusion effect. To solve these problems, a fusion method based on infrared target extraction and sparse representation is proposed. Firstly, the infrared target is detected and separated from the background rely on the regional statistical properties. Secondly, DENCLUE (the kernel density estimation clustering method) is used to classify the source images into the target region and the background region, and the infrared target region is accurately located in the infrared image. Then the background regions of the source images are trained by Kernel Singular Value Decomposition (KSVD) dictionary to get their sparse representation, the details information is retained and the background noise is suppressed. Finally, fusion rules are built to select the fusion coefficients of two regions and coefficients are reconstructed to get the fused image. The fused image based on the proposed method not only contains a clear outline of the infrared target, but also has rich detail information.     

A study on the evaluation of usefulness depending on the image reconstruction method in bone SPECT scan

The recent advances in image-processing techniques have led to the development of many methods to reduce the scan time without degrading the image quality. In particular, tomography has improved image reconstruction methods with the concomitant improvement of high-quality images. In this study, PRECEDENCE 16 was used to reconstruct images using the filtered back projection method, which is generally used, and the Astonish method and three-dimensional ordered-subsets expectation maximization method, which are based on repetition techniques. In qualitative and quantitative analysis of the reconstructed images, a comparison was made between images with different acquisition times and between images with the same acquisition time, which aimed at determining the optimal method for reconstructing high-quality images. A blind test for qualitative analysis confirmed almost no difference in image quality depending on the image acquisition time. Furthermore, in quantitative analysis, there was no significant difference in image quality depending on the image acquisition time. On the other hand, the results of the analysis in the image reconstruction method with the same acquisition time demonstrated a significant difference. The images reconstructed by the Astonish method, which uses a repetition technique, are believed to be excellent because they have high resolution and provide clinical diagnostic information. This study confirmed that the reconstruction method with a repetition technique could be used to improve image quality and reduce the scan time, despite not being in general use until recently due to the lengthy time needed for image reconstruction and lack of storage space.