Camera calibration plays a critical role in 3D computer vision tasks. The most commonly used calibration method utilizes a planar checkerboard and can be done nearly fully automatically. However, it requires the user to move either the camera or the checkerboard during the capture step. This manual operation is time consuming and makes the calibration results unstable. In order to solve the above problems caused by manual operation, this paper presents a full-automatic camera calibration method using a virtual pattern instead of a physical one. The virtual pattern is actively transformed and displayed on a screen so that the control points of the pattern can be uniformly observed in the camera view. The proposed method estimates the camera parameters from point correspondences between 2D image points and the virtual pattern. The camera and the screen are fixed during the whole process; therefore, the proposed method does not require any manual operations. Performance of the proposed method is evaluated through experiments on both synthetic and real data. Experimental results show that the proposed method can achieve stable results and its accuracy is comparable to the standard method by Zhang.

Traditionally, magnetic loop detectors are often used to count vehicles passing over them in intelligent transportation system. Real-time image sequences are captured by video surveillance system. Virtual loop, which emulates the functionality of inductive loop detectors, is placed on images. It is more convenient, but it occurs in false detection and discrimination when vehicles are lane departure due to overtaking or crossing. This paper presents an effective approach for vehicle counting based on double virtual lines (DVL). Double virtual lines are assigned on images, which are across bidirectional multi-lane. The region between DVL is the detection zone, rather than virtual loop zone in each lane, so as to reduce the proportion of false detection and misjudgment from lane departure for vehicles. Then, in the detection zone, the dual-template convolution is designed to detect and locate moving vehicles to eliminate the mapping of one to many, many to one. The effective rules are given in terms of the constraint of the horizontal and vertical distances to improve the accuracy of vehicle counting. Experimental comparisons with the other method demonstrate the performance of the proposed method.

Location and mapping are fundamental problems for a mobile robot to implement a series of high-level applications. Traditional solutions to Simultaneous Location and Mapping (SLAM) are probabilistic reasoning. This paper proposes an analytic solution to 3D visual SLAM with a Time-of-Flight (TOF) camera. According to the visual registration, 3D visual SLAM problem is decomposed into such steps as environment sensing, data matching, motion estimation, as well as location update and registration of new landmarks. First, TOF range camera enables the robot to capture images of distance and intensity of a scene. Scale-Invariant feature transform (SIFT) algorithms for visual feature extraction is applied to the captured intensity images. These visual features combined with the corresponding distance information give a full measurement of 3D landmarks. Then, the process of data association and match is developed through SIFT and the Iterative Closest Point (ICP) to minimize the relative and global error in SLAM process, while obtaining motion estimation. Finally, based on the visual theory of structure from motion (SFM), an analytic solution to location and mapping is presented to 3D Visual SLAM, instead of conventional probabilistic reasoning. We provide 3D visual SLAM experimental results from simulation and the indoor environment. It turns out that the proposed scheme is feasible.

Inspired by the human 3D visual perception system, we present an obstacle detection and classification method based on the use of Time-of-Flight (ToF) cameras for robotic navigation in unstructured environments. The ToF camera provides 3D sensing by capturing an image along with per-pixel 3D space information. Based on this valuable feature and human knowledge of navigation, the proposed method first removes irrelevant regions which do not affect robot’s movement from the scene. In the second step, regions of interest are detected and clustered as possible obstacles using both 3D information and intensity image obtained by the ToF camera. Consequently, a multiple relevance vector machine (RVM) classifier is designed to classify obstacles into four possible classes based on the terrain traversability and geometrical features of the obstacles. Finally, experimental results in various unstructured environments are presented to verify the robustness and performance of the proposed approach. We have found that, compared with the existing obstacle recognition methods, the new approach is more accurate and efficient.

Accessible frontier is an important factor for mobile robot autonomous exploration. This paper presents a fast and robust frontier line segment extracting method based on fuzzy c-means clustering algorithm for robot exploration. Firstly, the proposed method divides robot's local occupancy map into sub-regions with same size. In the next step, this paper analyzes the characteristic of robot exploration frontier with occupancy grid map, and the optimal number of FCM cluster center in each sub-region is defined. Consequently, line segments corresponding to exploration frontiers based on fuzzy c-mean algorithm are calculated in sub-region level to alleviate the extensive computation. Following those steps, line segments merging, line endpoints extending and line excluding are conducted to get more accurate frontier segment parameters in global level. In the end, the effectiveness of proposed method is verified by experiments results in lab environment.

This paper presents a robot navigation method based on fuzzy inference and behavior control. Stroll, Avoiding, Goal-reaching, Escape and Correct behavior are defined for robot navigation. The detailed scheme for each behavior is described in detail. Furthermore, fuzzy rules are used to switch those behaviors for best robot performances in real time. Experiments about five navigation tasks in two different environments were conducted on pioneer 2-DXE mobile robot. Experiment results shows that the proposed method is robust and efficiency in different environments.

针对TOF相机三维信息可视化中的突变、拉伸、毛刺等的问题,提出了一种基于仿人视觉映射的TOF相机三维数据信息可视化效果改善处理方法.方法首先研究分析了TOF三维相机和人类视觉系统感知原理,探寻突变、拉伸和毛刺现象的原因.在此基础上,提出了将TOF三维信息的锥形区域显示映射为人类视觉系统平行光模式的视觉映射模型,通过该映射处理达到逼近人类视觉的效果.该算法对多个不同场景进行了实验验证分析,结果表明该方法有效解决了三维显示中的变形和异常问题,有效改善了三维视觉可视化效果,并保持了三维目标的良好空间连续性.

We investigate the motion estimation from image sequence features,and propose for it a linear algorithm based on the parallel-line-segment(PLS) correspondences.The line segment is represented by two elements:end points and the line in between.The space line segment structure is reconstructed step by step from image lines,using the principle of parallelism.Then,the two elements of a space line segment are established based on the motion parameter equations which are solved by using the screw theory and quaternion.Finally,the motion parameters are optimized by the particle swarm optimization(PSO) algorithm.Our algorithm needs at least two lines and two perspective views to obtain the parameters;thus,the multisolution phenomenon is avoided and the magnitude of the translation is estimated with high computation efficiency.Simulations and real experiments illustrate the effectiveness of the proposed algorithm.

Environment perception is the foundation of mobile robot autonomous navigation. In order to correctly perceive slope terrain, a description model is discussed. Based on the change trend of distance between the mobile robot and the slope, their relative position can be determined. Based on the strong local nonlinear approximation ability of RBF neural network, a novel method for estimating terrain slope is proposed. Experiment results confirm that the proposed approach can be used for mobile robot incline terrain perception and has the merits of simplicity, trustiness and robustness.

针对电厂120MW以上容量机组的大型冷凝设备污垢在线清洗,研发了一种清洗机器人。根据冷凝水室的工作环境,对机器人的行走机构、供水管道和机械臂的传动机构作了针对性设计,并研制了样机。采用基于视觉伺服的方法,实现了冷凝管口的自主定位控制。经实验测试,机器人的性能指标基本满足大型冷凝设备清洗自动化的需求,可推广应用于电力、化工、制药等行业,对促进大规模安全生产、节能降耗有重大的意义。

In order to correctly sense incline terrain, its geometrical calculated model is analyzed. Based on the change trend of distance between the mobile robot and slope, their relative position can be determined. Then a novel method which takes the use of the powerful nonlinearity approach capability of RBF network is introduced to estimate the slope of the terrain with respect to the robot's current angular tilt. The experimental results show that the proposed approach can be used for mobile robot incline terrain perception and has the merits of easy, trustiness and robustness.

This paper presents a novel method for real-time obstacle detection and recognition in natural terrain for a field mobile robot using a image information, geometric information and support vector machine(SVM). Firstly, the scene is divided into two distinct regions: interest regions and uninterested regions. Then detected obstacle points are clustered into objects on the basis of their geometric information, i.e., depth and horizontal information, connectivity. The key obstacle characteristics are identified as width, height, their ratio, the ratio of area and depth. In the paper, the SVM method is used to classifying the objects into four classes. In order to determine the slope value, a SVM slope estimation approximation model was also proposed. Experimental results are presented to demonstrate the capability of the proposed approach for recognition of different obstacle in natural terrain.

环境感知是实现移动机器人自主导航的基础与根本保证.为了正确感知斜坡地形,分析了其描述模型.根据移动机器人从不同位置观测斜坡时深度信息的变化趋势,确定移动机器人观测斜坡的方向.在此基础上,提出了应用径向基(RBF)神经网络强非线性逼近能力估算地形坡度值的新方法.实验结果表明,所提出的算法能满足移动机器人感知斜坡地形的要求,同时算法具有简单,准确,鲁棒性强的优点.更多还原

The establishment and solution of a PnP problem are investigated here,and a vector difference based solution algorithm for linear P5P problem with an un-calibrated camera is proposed.Five control points are divided into collections containing four points of non-planar.According to the process of image formation,the constraint equations are set up between two collections of vector differences of control points.Then, these equations are simplified and determined in terms of linear theory and the orthogonal relation of rotation R.As a result, the analytic solutions of camera pose and intrinsic parameter matrix are obtained via vector cross or inner product operation. Simulated and real experiments illustrate that the proposed algorithm is effective.

We propose a hierarchical mean shift (HMS) algorithm for object tracking. Firstly, cluster modal points are obtained by mean-shift iteratively processing all the data points in the region so that they can represent foreground object in a succinct manner. The target model and the target candidate model are described by the cluster modal points, and match processes of clustered blocks are performed. Then, on the basis of cluster blocks match, similarity measure function is set up to match between target model and target candidate at pixel level. And the pixel shift vector of target is calculated with the introduction of the neighborhood consistency concept. So, the centroid of tracking object is got layer by layer in the consecutive frames, and the HMS match iteration for object tracking is presented. Experimental comparisons with other two MS algorithms demonstrate the validity and performance of the proposed algorithm. © 2009 Acta Automatica Sinica. All rights reserved.

For keeping the details of an image when removing noise,a novel adaptive filter is proposed.According to whether the gray value of a pixel is its neighborhood extremum,all pixels are divided into two types,doubtful noise and signals.If the center pixel is doubtful noise,the signals,together with the center pixel,are recognized as a new type.By comparing the difference between the median of the signals and the new type with the threshold,it can be decided that if the original gray value should be instead by the median of the signals.The size of the window can change adaptively along with the density of the noise.In simulation experiment,it is proved that the filter has an excellent performance and is more effective than ordinary median filter especially on high noisy rate situations.

针对经典Mean shift(MS)目标跟踪算法的颜色特征鲁棒差、匹配迭代复杂的缺点,提出一种分层Mean shift (Hierarchical mean shift,HMS)目标跟踪算法.首先通过MS迭代将目标区域特征空间的数据点聚类于模式点,使得以简洁的方式描述前景跟踪目标,建立目标模型与目标候选模型的聚类模式点描述,进行聚类块匹配.然后,导出聚类块模式点匹配下的相似度量函数,进行像素点匹配,结合邻域一致性,计算像素平移量,分层估计序列帧中跟踪目标质心模式点的位置,并给出HMS匹配迭代跟踪算法.实验结果表明,与其他两种MS跟踪算法相比,HMS既能提高序列帧跟踪目标表达与匹配的鲁棒性,又无需匹配所有数据点,算法简洁且有效可行.

A self-calibration approach to hand-eye relation of robot based on a single point in the scene is proposed.The manipulator end-effector is accurately controlled to perform five(or more) pure translational motions and two(or more) pure rotational motions,and the camera is required to image a single point in the scene.Motions of the camera are estimated based on the disparity and depth value of the point,relative position constraint equations between coordinates of the manipulator end-effector and the camera are set up,and the intrinsic parameters of camera and the hand-eye relation are obtained linearly.In the calibration process,only one point in the scene needs to be extracted,and neither matching nor orthogonal motion is required,therefore,motions of the manipulator can be controlled conveniently and the algorithm can be implemented simply.Experiment results of both simulation data and real image data show that the proposed method is effective and feasible.