This PDF file contains the front matter associated with SPIE Proceedings Volume 10253, including the Title Page, Copyright information, Table of Contents, Introduction (if any), and Conference Committee listing.

Applications of computer vision techniques assume an image acquisition from one or another sensing system. This system should be calibrated before the usage to obtain proper results. In this paper a calibration technique for the stereo camera system with the laser illumination is proposed. Modern approaches to the calibration of the different sensing systems are indicated. The characteristics of the specific system required a calibration are described. The main calibration tasks and subtasks for the given system and also the main stages of the proposed technique are highlighted. The need to rotate the laser illumination relative to the axis between the cameras through 8 degrees is proven. An approach to the calibration of the illumination laser beam directions is developed. The accounting of the parameters which can be obtained as an issue of the complex calibration of the stereo camera system with the laser illumination makes it possible to improve the results of the analyzed system utilization for mobile robots.

This article presents an approach to the optical flow (OF) usage as a general navigation means providing the information about the linear and angular vehicle's velocities. The term of “OF” came from opto-electronic devices where it corresponds to a video sequence of images related to the camera motion either over static surfaces or set of objects. Even if the positions of these objects are unknown in advance, one can estimate the camera motion provided just by video sequence itself and some metric information, such as distance between the objects or the range to the surface. This approach is applicable to any passive observation system which is able to produce a sequence of images, such as radio locator or sonar. Here the UAV application of the OF is considered since it is historically

Fighting detection is an important issue in security aimed to prevent criminal or undesirable events in public places. Many researches on computer vision techniques have studied to detect the specific event in crowded scenes. In this paper we focus on fighting detection using social-based Interaction Energy Force (IEF). The method uses low level features without object extraction and tracking. The interaction force is modeled using the magnitude and direction of optical flows. A fighting factor is developed under this model to detect fighting events using thresholding method. An energy map of interaction force is also presented to identify the corresponding events. The evaluation is performed using NUSHGA and BEHAVE datasets. The results show the efficiency with high accuracy regardless of various conditions.

In this paper we study multiple reflection effect in a fold of material with regard to color constancy problem. Namely we consider light source chromaticity estimation using perceived material color. We measured relative spectra of reflected light source emission for different positions under folds. Experiment was performed on 105 fabric samples. Using this data we discuss applicability of different spectral models for description of observed chromaticity deviation in different fold’s areas. Obtained experimental data was released in open access.

Computing image patch descriptors for correspondence problems relies heavily on hand-crafted feature transformations, e.g. SIFT, SURF. In this paper, we explore a Siamese pairing of fully connected neural networks for the purpose of learning discriminative local feature descriptors. Resulting ANN computes 128-D descriptors, and demonstrates consistent speedup as compared to such state-of-the-art methods as SIFT and FREAK on PCs as well as in embedded systems. We use L₂ distance to reflect descriptor similarity during both training and testing. In this way, feature descriptors we propose can be easily compared to their hand-crafted counterparts. We also created a dataset augmented with synthetic data for learning local features, and it is available online. The augmentations provide training data for our descriptors to generalise well against scaling and rotation, shift, Gaussian noise, and illumination changes.

We study a technique for improving visualization quality of noisy multispectral images. Contrast form visualization approach is considered, which guarantees a non-zero contrast in the output image when there is a difference between the spectra of the object and the background in the input image. The improvement is based on channel weighting according to estimation of the noise level. We show this approach to reduce noise in color visualization of real multispectral images. The low-noise visualizations are demonstrated to be more comprehensive to a human on examples from a publicly available dataset of Earth surface images. Noise variance estimation needed for weighting uses the method proposed earlier by the authors. The validation dataset consists of publicly available images of Earth surface.

This paper presents a method of radial distortion automatic compensation on video from an unknown camera. The proposed algorithm estimates the distortion parameters by analyzing a sequence of video frames. It does not require any calibration objects, but is based on the assumption that the original scene contained straight lines. The method tries to perform such radial distortion correction that makes lines look generally straighter. To estimate the overall curvature of the lines we propose to use the fast Hough transform; without actually detecting them in the image. The proposed algorithm has been tested on real data.

Systems of panoramic photography and video are developed, as a rule, on the basis of an array of light-sensitive sensors, with different ways of positioning in space. The authors developed a high-quality portable system panoramic photo and video using a 12 light-sensitive sensors and the formation of the video standard Ultra HD 4K. According to the simulation results, it was found that the optimal arrangement of 12 the light-sensitive sensors in conjunction with lenses is their location in the center of the dodecahedron faces. In this case, part of the image formed on each photosensitive lens sensor is unique (not repeating other sensors) as part of the panorama of a regular pentagon. This design allows you to create a panorama of 360-degrees. The developed system is a panoramic photo and video, using PLD (programmable logic devices) and includes modules removal of distortions, masking, calibration, image formation on the sphere describing the dodecahedron and obtaining equidistant projection, modules and compression encoding an image. The article presents the basic elements of the developed system of panoramic photography and video.

Image fractal coding possesses very high compression ratio, the main problem is low speed of coding. The algorithm based on Complex Exponent Moments(CEM) and minimum variance is proposed to speed up the fractal coding compression. The definition of CEM and its FFT algorithm are presented, and the multi-distorted invariance of CEM are discussed. The multi-distorted invariance of CEM is fit to the fractal property of an image. The optimal matching pair of range blocks and domain blocks in an image is determined by minimizing the variance of their CEM. Theory analysis and experimental results have proved that the algorithm can dramatically reduce the iteration time and speed up image encoding and decoding process.

Obtaining high quality images from Computed Tomography (CT) is important for correct image interpretation. In this paper, we propose novel procedures that can be used for a quantitative description of the degree of artifact expressiveness in CT images, and show that the use of this type of metric allows to assess the dynamics of image degradation. We perform different image reconstruction tests in order to analyse our approach, and the obtained results confirm the usefulness of the proposed method. We conclude that the use of the proposed estimates allows moving from image quality assessment based on visual scoring to a quantitative approach and consequently to support a CT setup providing high quality reconstructed images obtained by appropriate changes of the reconstruction parameters or algorithms.

Reconstruction of a drivability map for a moving vehicle is a well-known research topic in applied robotics. Here creating such a map for an autonomous truck on a generally planar surface containing separate obstacles is considered. The source of measurements for the truck is a calibrated pair of cameras. The stereo system detects and reconstructs several types of objects, such as road borders, other vehicles, pedestrians and general tall objects or highly saturated objects (e.g. road cone). For creating a robust mapping module we use a modification of Bayes filtering, which introduces some novel techniques for occupancy map update step. Specifically, our modified version becomes applicable to the presence of false positive measurement errors, stereo shading and obstacle occlusion. We implemented the technique and achieved real-time 15 FPS computations on an industrial shake proof PC. Our real world experiments show the positive effect of the filtering step.

In this paper we propose a novel method for localization based on matching two stereo images. It is based on minimizing the sum of square distances between each 3D point and four corresponding 3D rays. The method shows good results for practical localization purposes. Moreover it is robust to the presence of feature point correspondences with zero disparity, which is usually a problem for classical methods. The algorithm is tested in comparison to the classical method. It has linear complexity with respect to feature point correspondence number.

This paper presents the design of a control structure that ensures no overshoot for the movement of an industrial robot, used for the evacuation of round steel blocks from inside a rotary hearth furnace. First, a mathematical model for the positioning system is derived from a set of experimental data, and further, the paper focuses on obtaining a PID type controller, using the relay method as tuning method in order to obtain a stable closed loop system. The controller parameters are further tuned in order to achieve the imposed set of performances for the positioning of the industrial robot through computer simulation, using trial and error method. Further, a fractional – order PID controller is obtained in order to improve the control signal variation, so as to fit within the range of unified current’s variation, 4 to 20 mA.

The image data that object recognition systems are designed for changes over time. As soon as a new imaging technology is developed or becomes affordable new algorithms are inspired or known algorithms are adapted. Thus, different object recognition algorithms were developed and used on our mobile robot Lisa. In this work we compare the different approaches and investigate how they can be combined to best use 2D and 3D data. The individual approaches as well as their combinations will be introduced. Evaluation is performed on a large public dataset and a dataset acquired during the RoboCup competition.

Smart algorithms are used in Machine vision and Robotics to organize or extract high-level information from the available data. Nowadays, Machine learning is an essential and ubiquitous tool to automate extraction patterns or regularities from data (images in Machine vision; camera, laser, and sonar sensors data in Robotics) in order to solve various subject-oriented tasks such as understanding and classification of images content, navigation of mobile autonomous robot in uncertain environments, robot manipulation in medical robotics and computer-assisted surgery, and other. Usually such data have high dimensionality, however, due to various dependencies between their components and constraints caused by physical reasons, all „feasible and usable data‟ occupy only a very small part in high dimensional „observation space‟ with smaller intrinsic dimensionality. Generally accepted model of such data is manifold model in accordance with which the data lie on or near an unknown manifold (surface) of lower dimensionality embedded in an ambient high dimensional observation space; real-world high-dimensional data obtained from „natural‟ sources meet, as a rule, this model. The use of Manifold learning technique in Machine vision and Robotics, which discovers a low-dimensional structure of high dimensional data and results in effective algorithms for solving of a large number of various subject-oriented tasks, is the content of the conference plenary speech some topics of which are in the paper.

Using multiple Lyapunov function method, the H∞ control theory of switched Hamiltonian systems is discussed. The sufficient conditions in the form of linear matrix inequality (LMI) are presented to guarantee stability of switched Hamilton system. At the same time, an switching law based on switching dwell time is constructed. In order to prove the effectiveness of this conclusion, a numerical simulation example is given, in which the corresponding parameters are calculated by LMI.

In this work we address the problem of detecting and recognizing transparent objects using depth images from an RGB-D camera. Using this type of sensor usually prohibits the localization of transparent objects since the structured light pattern of these cameras is not reflected by transparent surfaces. Instead, transparent surfaces often appear as undefined values in the resulting images. However, these erroneous sensor readings form characteristic patterns that we exploit in the presented approach. The sensor data is fed into a deep convolutional neural network that is trained to classify and localize drinking glasses. We evaluate our approach with four different types of transparent objects. To our best knowledge, no datasets offering depth images of transparent objects exist so far. With this work we aim at closing this gap by providing our data to the public.

Rail wear measurement is a necessary task in railway infrastructure inspection. To acquire the wear amounts accurately with more continuous scanning data, a rail wear measurement method based on structured light scanning is proposed in this paper. The CAD model of the rail is converted into a point set, and the data registration is implemented by aligning the scanning data to the point cloud generated by the CAD model. On a cross section plane of the rail, the vertical and lateral wear amounts are calculated by the nearby points projected onto the plane. To verify the accuracy of wear measurement based on structured light scanning, the wear amounts calculated by laser scanning data are compared. For the comparison, an experiment is designed to ensure that the same plane is sliced in two different kinds of measurement. On the cross section plane, the wear amounts are calculated by the distances from these points to the 2D profile of the rail CAD model, and then the results are compared with those from laser scanning data for the accuracy evaluation. It indicates that the accuracy of the structured light scanning is sufficient for railway track wear measurement.

To deal with the problems of the local stress concentration caused by the internal hydraulic pressure during the operation of cubic press, we have used the theory of thick-wall block to analyze and calculate the loads of the hinge beam block. The results demonstrate that the working block suffering from the excessive hydraulic pressure contacts with the bottom of hinge beam each other, which gives rise to the stress concentration in the arc of the bottom of the hinge beam block. According to the data of the theoretical calculation, the strain tests of the arc are completed by the strain testing methods in the bottom of the hinge beam block. The Comparison between the data of the two groups has verified their consistency. It also confirms the theoretical calculation is reasonable and accurate at the same time.

The five-point algorithm is an efficient way of evaluating camera motion parameters from five point pairs from two distinct views. However there is a need of tenth degree polynomial solving emerges during the computational process. In the paper we investigate the statistical properties of polynomial solvers used as a part of the five-point algorithm. We adduce the mathematical background of the problem and study briefly the main four polynomial solving methods. Finally, we investigate the essential characteristics of the algorithms such as parameters of distribution of an error value, rate of fails and average computation time. To evaluate the solvers we conduct an experiment using synthetic data.

The problem of decentralized control of group of robots, described by kinematic and dynamic equations of motion in the plane, is considered. Group performs predetermined rectangular area passing at a fixed speed, keeping the line and a uniform distribution. The environment may contain a priori unknown moving or stationary obstacles. Decentralized control algorithms, based on the formation of repellers in the state space of robots, are proposed. These repellers form repulsive forces generated by dynamic subsystems that extend the state space of robots. These repulsive forces are dynamic functions of distances and velocities of robots in the area of operation of the group. The process of formation of repellers allows to take into account the dynamic properties of robots, such as the maximum speed and acceleration. The robots local control law formulas are derived based on positionally-trajectory control method, which allows to operate with non-linear models. Lyapunov function in the form of a quadratic function of the state variables is constructed to obtain a nonlinear closed-loop control system. Due to the fact that a closed system is decomposed into two independent subsystems Lyapunov function is also constructed as two independent functions. Numerical simulation of the motion of a group of five robots is presented. In this simulation obstacles are presented by the boundaries of working area and a movable object of a given radius, moving rectilinear and uniform. Obstacle speed is comparable to the speeds of the robots in a group. The advantage of the proposed method is ensuring the stability of the trajectories and consideration of the limitations on the speed and acceleration at the trajectory planning stage. Proposed approach can be used for more general robots’ models, including robots in the three-dimensional environment.