基于霍夫变换与模板匹配的栅格地图拼接方法

单个机器人在面对大尺度环境时,建图结果往往存在较大误差。由多个机器人创建出局部地图后进行拼接可以很好地解决这一问题。本文提出了一种基于霍夫变换与模板匹配的地图拼接方法。首先,利用霍夫变换检测局部地图中的直线段,根据最长直线段斜率将局部地图正交方向重对准。然后利用Harris角点检测算法检测局部地图中的角点,并以角点为中心,提取模板图像。最后利用模板匹配算法将模板图像与另一局部地图进行匹配,获取两局部地图之间的刚体变换矩阵,实现地图拼接。实验表明,该方法对尺度较大的现实环境具有很好的实用性。     

基于Radon变换的图像角点角度提取算法

针对现有角点检测算法只提取图像角点位置而无法提取角点角度的缺陷,本文在SUSAN角点检测算法的基础上,提出基于Radon变换的图像角点角度提取算法.该算法首先利用SUSAN算子检测候选角点,然后得到候选角点子邻域图像,对其进行角点角度可测量性判定,对满足可测量性的角点子邻域图像中每个像素点进行梯度映射后,利用Radon变换提取角点边缘直线,最后利用坐标几何关系提取角度.并针对Radon变换无法检测直线起始点导致角点角度提取时存在二值歧义性的缺陷,本文借鉴USAN思想,把同值核引入到角度提取中,从而实现了对角度的正确提取.该算法实现简单,实验结果证明了其有效性.     

基于兴趣点特征匹配的印刷图像缺陷检测

提出了1种基于兴趣点特征匹配的印刷图像缺陷检测方法.首先利用SIFT特征检测器,获取图像中具有旋转和平移不变性的特征点,然后采用最近邻方法进行特征点匹配,最后使用最小二乘优化算法,计算2幅图像特征点对的变换矩阵来实现缺陷图像的检测.实验结果表明:该方法能够快速准确地提取出2幅图像间的对应特征点,大大降低了误匹配的概率,在图像发生较大平移和旋转的情况下,仍能准确地实现图像的对准及缺陷图像的检测.     

基于平面重建与自然特征匹配的虚实注册方法

针对增强现实系统中的三维注册问题,提出了一种新的基于多平面结构三维重建与自然特征匹配相结合的虚实注册方法,该方法分为离线三维重建和在线实时注册两个阶段.离线阶段系统利用两幅参考图像和双目视觉原理恢复出平面坐标系,并根据用户指定的世界坐标系计算两者之间的变换关系.实时注册阶段,系统首先将当前图像与参考图像做特征匹配以获取平面所对应的单应性关系,进而恢复出每个平面所对应的投影矩阵,由上述投影矩阵以及离线计算所获得的世界坐标系与多个平面坐标系之间的变换关系合成三维注册所需的变换矩阵.实验结果表明,该方法是有效可行的,而且在克服误差积累方面较传统方法有明显改善.     

基于图像处理的非规则形状检测方法

针对工业产品中非规则形状检测方法少、检测效率低的现状,本文提出了一种基于图像处理的快速高效检测方法。该方法首先对实际拍摄图像与参考模型进行特征点匹配,通过匹配寻找两幅图像中轮廓位置的映射关系,进而求解实际拍摄图像到参考模型的透视变换参数;然后对实际拍摄图像中的非规则形状进行高精度边缘提取,利用求取的透视变换参数对提取的边缘点进行变换;最后将计算得到的数据与参考图像中的数据进行比较,以此进行非规则形状检测。该方法为非接触式,设备少,环境影响小,数据对象仅是摄影图像,检测速度快,适用于各种非规则形状。标准实验分析表明,该方法检测精度优于0.2 mm,具有较高的检测精度及较好的检测效率。     

一种用于图像拼接的角点匹配算法

针对基于自适应光学技术的人眼眼底待拼接图像灰度、清晰度不一致以及细节的差异使检测出的角点无法匹配,进而导致图像不能拼接的问题,提出一种在OpenCV配置环境下角点检测的基础上,进行角点匹配的新方法.该方法利用角点间的距离和斜率在眼底很小的区域内基本不变的特性,将两幅图像的角点对相互匹配,进而找到两幅图像中三对一一对应的角点.实验结果表明,用该方法可以解决人眼图像拼接的角点检测和角点匹配问题,而且运算速度快,检测效率高.     

结合Hough变换与Kalman滤波的车道线检测方法

车道线检测是智能驾驶系统的重要组成部分,它提供了车辆与车道位置关系的信息.针对智能车辆驾驶系统在视觉导航过程中车道线检测的精确性和鲁棒性的问题,提出一种有效的车道线检测方法.首先对原始RGB图像分别进行感兴趣区域设定、逆透视变换、灰度化和阈值处理;然后进行霍夫变换处理,利用斜率和中心点位置筛选检测结果;最后利用卡尔曼滤波对检测到的线段进行跟踪,预测当前车道线位置.实验结果表明,该算法能够有效解决图像中车道线不清晰以及一些干扰遮挡的问题,车道线检测准确率可达94%,具有较好的准确性、鲁棒性和较低的计算复杂度,有利于实时性检测系统的构建.     

基于灰度相关特征点的图像拼接算法

提出了一种基于特征点的图像拼接算法,首先利用Harris角点检测提取两幅图像中的特征点,用特征点邻域灰度相关性将特征点进行匹配,并在此基础上根据变换参数向量的欧式距离对误匹配进行剔除,进而实现图像拼接.实验表明,这种算法能有效的排除误匹配的干扰,降低误匹配率,拼接效果良好.     

基于投票策略的图像配准算法

针对二维刚体图像配准,本文提出了一种由粗到细的图像配准方法.该方法首先由多尺度Harris角点检测算法提取出参考图像和目标图像的特征角点,采用基于投票策略的特征点匹配算法进行初步配准,有效地减少假点和不稳定点,然后对变换矩阵参数进行初估计,采用基于窗的精细算法对变换矩阵参数进行精估计,并得到最终的变换矩阵.实验结果表明,该算法具有高精度,对噪声鲁棒和低计算量的特点.     

Hough变换和最小二乘拟合的车道线协调检测

针对车道线检测的实时性和鲁棒性要求,借助车道线参数置信区间和Kalman滤波器,采用基于Hough变换的车道线参数全局提取和基于最小二乘拟合的车道线参数局部小窗口提取相协调的方法对车道线进行检测.通过设定车道线感兴趣区域,利用改进快速中值滤波、车道线特征滤波、梯度方向角直方图筛选和车道线连通性分析等图像预处理算法削减图...     

利用鱼眼照片恢复景物的深度信息

通过照片获取深度信息的技术已经研究多年，也产生了各种算法，但这些算法或者需要一张以上的照片来计算，或者运算量比较大。此外，它们的共同点是：各种方法采用的照片是用普通镜头拍摄的。笔者给出一种新的方法：利用一张鱼眼镜头照片和少量的辅助手段来获取深度信息。建立了一种简便而又准确的辅助手段，从而可以构造景物的三维框架，最终可以实现在景物内的任意位置向任意方向进行虚拟浏览。     

AHSPP Annual Conference

Abstract

A calibration and three-dimensional (3D) reconstruction method is presented based on images reflected from planar mirrors and acquired with one camera. The geometric model of the camera–mirror set and a method of calibrating it are described. The calibration technique computes the model parameters using linear equations, and it is proved that the calibration is possible with the knowledge of only six 3D points. The reconstruction method is based on a volumetric representation. The 3D reconstruction is based on a space carving algorithm and the calibration method described in the paper. The results of the calibration and reconstruction method show the efficiency of both techniques. This set-up enables a simple and inexpensive multi-ocular system to be built to recover the 3D structure of volumes.     

Three-dimensional multiexcitation magnetoacoustic tomography with magnetic induction

Magnetoacoustic tomography with magnetic induction (MAT-MI) is a hybrid imaging modality proposed to image electrical conductivity contrast of biological tissue with high spatial resolution. This modality combines magnetic excitations with ultrasound detection through the Lorentz force based coupling mechanism. However, previous studies have shown that MAT-MI method with single type of magnetic excitation can only reconstruct the conductivity boundaries of a sample. In order to achieve more complete conductivity contrast reconstruction, we proposed a multiexcitation MAT-MI approach. In this approach, multiple magnetic excitations using different coil configurations are applied to the object sequentially and ultrasonic signals corresponding to each excitation are collected for conductivity image reconstruction. In this study, we validate the new multiexcitation MAT-MI method for three-dimensional (3D) conductivity imaging through both computer simulations and phantom experiments. 3D volume data are obtained by utilizing acoustic focusing and cylindrical scanning under each magnetic excitation. It is shown in our simulation and experiment results that with a common ultrasound probe that has limited bandwidth we are able to correctly reconstruct the 3D relative conductivity contrast of the imaging object. As compared to those conductivity boundary images generated by previous single-excitation MAT-MI, the new multiexcitation MAT-MI method provides more complete conductivity contrast reconstruction, and therefore, more valuable information in possible clinical and research applications.     

角检测在彩色伪随机编码图像特征提取中的应用

为研究基于单幅二维图像不标定欧氏重构三维场景的理论,有效提取伪随机编码图像特征角点,介绍了一种角检测方法,该角检测方法是基于轮廓线曲率的测试,实验数据和重构结果验证了该检测方法的适用性和有效性。     

多视角几何三维重建法识别工程结构缺损与变形

裂缝等工程结构表面缺损与弯曲等构件变形是既有工程结构安全性检测鉴定与监测的重要指标,采用数字图像法可有效改善当前人工检查技术的缺点,但在工程实践中显现出定量检测中图像几何变形修正困难、局部损伤难以在整体结构中定位、无法测量钢结构空间变形等问题。结合多视角几何三维重建法的研究与应用可有效解决上述问题。该文对多视角几何三维重建的方法原理与算法实现进行讨论,列举方法实践的经典算法与高效算法;围绕数字图像法检测在工程实践中存在的问题,提出表面投影法解决成像几何变形与损伤定位、逆向工程建模与特征提取识别几何变形损伤;通过结构表面裂缝识别、大型构筑物表面损伤定位、钢结构构件变形识别等3个应用研究,阐述多视角几何三维重建法在结构表面损伤与变形识别中的具体操作方法与比较优势。多视角几何三维重建法具有设备要求低、操作方便快捷、模型色彩丰富、点云精度较高等特点,与数字图像法结合后在工程结构检测监测领域的研究与应用潜力巨大。     

Single‐Particle Cryo‐EM and 3D Reconstruction of Hybrid Nanoparticles with Electron‐Dense Components

Single‐particle cryo‐electron microscopy (cryo‐EM), accompanied with 3D reconstruction, is a broadly applicable tool for the structural characterization of macromolecules and nanoparticles. Recently, the cryo‐EM field has pushed the limits of this technique to higher resolutions and samples of smaller molecular mass, however, some samples still present hurdles to this technique. Hybrid particles with electron‐dense components, which have been studied using single‐particle cryo‐EM yet with limited success in 3D reconstruction due to the interference caused by electron‐dense elements, constitute one group of such challenging samples. To process such hybrid particles, a masking method is developed in this work to adaptively remove pixels arising from electron‐dense portions in individual projection images while maintaining maximal biomass signals for subsequent 2D alignment, 3D reconstruction, and iterative refinements. As demonstrated by the success in 3D reconstruction of an octahedron DNA/gold hybrid particle, which has been previously published without a 3D reconstruction, the devised strategy that combines adaptive masking and standard single‐particle 3D reconstruction approach has overcome the hurdle of electron‐dense elements interference, and is generally applicable to cryo‐EM structural characterization of most, if not all, hybrid nanomaterials with electron‐dense components.     

Convergence of Stereo Vision-Based Multimodal YOLOs for Faster Detectionof Potholes

Road potholes can cause serious social issues, such as unexpected damages to vehicles and traffic accidents. For efficient road management, technologies that quickly find potholes are required, and thus researches on such technologies have been conducted actively. The three-dimensional (3D) reconstruction method has relatively high accuracy and can be used in practice but it has limited application owing to its long data processing time and high sensor maintenance cost. The two-dimensional (2D) vision method has the advantage of inexpensive and easy application of sensor. Recently, although the 2D vision method using the convolutional neural network (CNN) has shown improved pothole detection performance and adaptability, large amount of data is required to sufficiently train the CNN. Therefore, we propose a method to improve the learning performance of CNN-based object detection model by artificially generating synthetic data similar to a pothole and enhancing the learning data. Additionally, to make the defective areas appear more contrasting, the transformed disparity map (TDM) was calculated using stereo-vision cameras, and the detection performance of the model was further improved through the late fusion with RGB (Red, Green, Blue) images. Consequently, through the convergence of multimodal You Only Look Once (YOLO) frameworks trained by RGB images and TDMs respectively, the detection performance was enhanced by 10.7% compared with that when using only RGB. Further, the superiority of the proposed method was confirmed by showing that the data processing speed was two times faster than the existing 3D reconstruction method.     

彩色伪随机阵列在主动三维视觉检测中的应用

要用一幅图像在三维欧氏空间中重构三维场景,尤其是动态的三维场景,常用的点结构光、线结构光扫描法不能满足要求。一个有效的方法是采用伪随机编码结构光照明主动视觉技术:用编码结构光照明被测场景,因为伪随机编码阵列中任一子窗口阵列都唯一,所以通过对场景图像中编码窗口的查寻,可以确定图像中编码点在编码阵列中的位置,从而使场景被测表面每一个特征点都具有唯一的代码,能唯一被辨识,并最终得以重构。     

Factor graph methods for three-dimensional shape reconstruction as applied to LIDAR imaging

Two methods based on factor graphs for reconstructing the three-dimensional (3D) shape of an object from a series of two-dimensional images are presented. First, a factor graph model is developed for image segmentation to obtain silhouettes from raw images; the shape-from-silhouette technique is then applied to yield the 3D reconstruction of the object. The second method presented is a direct 3D reconstruction of the object using a factor graph model for the voxels of the reconstruction. While both methods should be applicable to a variety of input data types, they will be developed and demonstrated for a particular application involving the LIDAR imaging of a submerged target. Results from simulations and from real LIDAR data are shown that detail the performance of the methods.     

Tomographic reconstruction of three-dimensional refractive index fields by use of a regularized phase-tracking technique and a polynomial approximation method

We present a complete data-processing procedure for quantitative reconstruction of three-dimensional (3D) refractive index fields from limited multidirectional interferometric data. The proposed procedure includes two parts: (1) extraction of the projection data from limited multidirectional interferograms by a regularized phase-tracking technique and wavefront fitting and (2) reconstruction of the 3D refractive index fields by a modified polynomial approximation method. It has been shown that the procedure gives a satisfactory solution to the reconstruction issue in interferometric tomography, from the initial projection data extraction to the final image reconstruction. Computer simulation and experimental results are both presented.