Search

Tese de Doutoramento em Engenharia Electrónica e de Computadores (Especialidade em Tecnologia dos Sistemas de Informação), As doenças neurológicas provocam frequentemente graves sequelas, nomeadamente problemas de mobilidade. Estes pacientes necessitam de programas regulares de fisioterapia para manterem a sua independência. No entanto, os exercícios da fisioterapia tradicional são monótonos e fastidiosos, deixando-os, muitas vezes, desmotivados. Vários estudos têm revelado que os jogos sérios criam ambientes digitais agradáveis e usam formas de feedback que encorajam os pacientes a executar os exercícios fisioterapêuticos com satisfação e tranquilidade, incentivando-os a permanecer cada vez mais motivados. Neste sentido, foi objetivo geral desta investigação desenvolver um jogo sério, para motivar e monitorizar a prática de fisioterapia de pacientes com doenças neurológicas, atendendo ao conceito de interação homem-máquina. Este objetivo foi dividido em objetivos parciais, que se fizeram corresponder a questões de investigação. Através da estratégia da investigação-ação, desenvolveu-se um sistema não invasivo que utiliza técnicas de processamento de imagem para monitorizar os pacientes, e adapta, a situações de jogo eletrónico, exercícios propostos pela fisioterapia tradicional. Com a estratégia de estudo de caso, experimentou-se o jogo desenvolvido em ambiente clínico, utilizando duas amostras: uma relativa a profissionais de saúde, e outra a pacientes com doenças neurológicas. A observação participante não estruturada e o inquérito por questionário foram as técnicas escolhidas para a recolha de dados para responder às questões de investigação. Além destes, utilizaram-se, ainda, os dados recolhidos pelo software. Os dados do questionário revelaram que o jogo é capaz de motivar os pacientes para a prática da fisioterapia, e os do software mostraram a sua capacidade de adaptação ao estado físico do paciente, de monitorização dos exercícios e de armazenamento de toda a informação. Os resultados obtidos permitiram responder afirmativamente a todas as questões de investigação, ou seja, foram atingidos os objetivos parciais definidos para esta investigação, concluindo-se que o jogo desenvolvido é um complemento válido à fisioterapia tradicional., Neurological diseases may cause serious consequences in their victims, including mobility problems. These patients require physical therapy programs to maintain their independence for the longest possible time. But traditional therapy exercises are often monotonous and boring, leaving them unmotivated. Several studies have revealed that serious games create pleasant digital environments and use feedback forms that encourage patients to perform the physical therapy exercises with pleasure and tranquility, encouraging them to remain increasingly motivated. In this sense, it was the general objective of this research to develop a serious game to motivate and monitor the practice of physical therapy of patients with neurological disease, taking into account the concept of man-machine interaction. This goal was divided into sub-goals, which correspond to research questions. Through the research-action strategy, it was developed a non-invasive system that uses image processing techniques monitoring of patients and that adapts some exercises of traditional physical therapy to electronic game situations. With the case study strategy, it was tested the developed game in a clinical environment, using two samples: one corresponding to health professionals and another to patients with neurological diseases. The unstructured participant observation and the questionnaire survey were the techniques used to collect data that allow to answer the research questions. In addition, the data collected by the software was also analyzed. The questionnaire data revealed that the game is able to motivate patients to practice physical therapy, and the software showed its ability to adapt to the physical condition of the patient, to monitor the exercises and to store all information. The results allowed to answer positively to all research questions or, in other words, the partial objectives defined for this study were achieved, and it may be concluded that the developed game is a good complement to traditional physiotherapy.

Wine growers prefer cultivars with looser bunch architecture because of the decreased risk for bunch rot. As a consequence, grapevine breeders have to select seedlings and new cultivars with regard to appropriate bunch traits. Bunch architecture is a mosaic of different single traits which makes phenotyping labor-intensive and time-consuming. In the present study, a fast and high-precision phenotyping pipeline was developed. The optical sensor Artec Spider 3D scanner (Artec 3D, L-1466, Luxembourg) was used to generate dense 3D point clouds of grapevine bunches under lab conditions and an automated analysis software called 3D-Bunch-Tool was developed to extract different single 3D bunch traits, i.e., the number of berries, berry diameter, single berry volume, total volume of berries, convex hull volume of grapes, bunch width and bunch length. The method was validated on whole bunches of different grapevine cultivars and phenotypic variable breeding material. Reliable phenotypic data were obtained which show high significant correlations (up to r² = 0.95 for berry number) compared to ground truth data. Moreover, it was shown that the Artec Spider can be used directly in the field where achieved data show comparable precision with regard to the lab application. This non-invasive and non-contact field application facilitates the first high-precision phenotyping pipeline based on 3D bunch traits in large plant sets.

Electrochemical bio-sensing using paper-based detection systems is the main focus of this review. The different existing designs of 2-dimensional and 3-dimensional sensors, and fabrication techniques are discussed. This review highlights the effect of adopting different sensor designs, distinct fabrication techniques, as well as different modification methods, in order to produce reliable and reproducible reading. The use of various nanomaterials have been demonstrated in order to modify the surface of electrodes during fabrication to further enhance the signal for subsequent analysis. The reviewed sensors were classified into categories based on their applications, such as diagnostics, environmental and food testing. One of the major advantages of using paper-based electrochemical sensors is the potential for miniaturization, which only requires relatively small amount of samples, and the low cost for the purpose of mass production. Additionally, most of the devices reviewed were made to be portable, making them well-suited for on-site detection. Finally, paper-based detection is an ideal platform to fabricate cost-effective, user-friendly and sensitive electrochemical biosensors, with large capacity for customization depending on functional needs.

Large-screen human-computer interaction technology is reflected in all aspects of daily life. The dynamic gesture tracking algorithm commonly used in recent large-screen interactive technologies demonstrates compelling results but suffers from accuracy and real-time problems. This paper systematically addresses these issues by a switching federated filter method that combines particle filtering and Mean Shifting algorithms based on a 3D sensor. Compared with several algorithms, the results show that the one-hand and two-hand large-screen gesture tracking based on the switched federated filtering algorithm work well, and there is no tracking failure and loss of target. Therefore, the switching federated tracking positioning algorithm can be well applied to the design of human-computer interaction system and provide new ideas for future human-computer interaction.

A new generation of 3D silicon pixel detectors with a small pixel size of 50 \(\times \) 50 and 25 \(\times \) 100 \(\upmu \)m\(^{2}\) is being developed for the HL-LHC tracker upgrades. The radiation hardness of such detectors was studied in beam tests after irradiation to HL-LHC fluences up to \(1.4\times 10^{16}\) n\(_{\mathrm {eq}}\)/cm\(^2\). At this fluence, an operation voltage of only 100 V is needed to achieve 97% hit efficiency, with a power dissipation of 13 mW/cm\(^2\) at \(-25\,^{\circ }\)C, considerably lower than for previous 3D sensor generations and planar sensors.

Repetitive movements in hands may cause strain injuries that negatively influence the employee performance. The lack of awareness for prevention this type of working injuries often leads the employers to believe that costs associated to prevention are unnecessary expenditures. So, the objective of this paper is to develop a serious game for preventing repetitive strain injuries in hands. The game activities are developed in Unity 3D software and using the 3D sensor, Intel RealSense 3D Camera F200. There are two sets of exercises, the warming up and the stretching off exercises, to be executed during the playing activity. Three game scenarios were developed. The tests performed so far allow concluding that the proposed system correctly detects the hands movements., The authors would like to express their acknowledgments to COMPETE:POCI-01-0145-FEDER-007043 and FCT – Fundação para a Ciência e Tecnologia within the Project Scope: UID/CEC/00319/2013., info:eu-repo/semantics/publishedVersion

We propose an automatic calibration method of a 3D sensor using a real-time simulator. With this method, the relative posture of a real sensor is estimated by obtaining the rigid body coordinate transformation that associates the virtual and real sensors in real-time. Moreover, the results of applying the method to a real system and verifying its effectiveness are shown.

Consumer-grade RGB-D cameras are widely accessible, but they suffer from a lack of accuracy when compared to professional-grade 3D scanning solutions. In this paper, we propose a new method for calibrating an Intel RealSense SR300 camera, adaptable to other structured light sensors. The method uses classical checkerboard calibration and a coordinate-measuring machine (CMM) based setup with a calibrating plane. It delivers better results than the manufacturers settings.

Three-dimensional (3D) sensing and printing technologies have reshaped our world in recent years. In this article, a comprehensive overview of techniques related to the pipeline from 3D sensing to printing is provided. We compare the latest 3D sensors and 3D printers and introduce several sensing, postprocessing, and printing techniques available from both commercial deployments and published research. In addition, we demonstrate several devices, software, and experimental results of our related projects to further elaborate details of this process. A case study is conducted to further illustrate the possible tradeoffs during the process of this pipeline. Current progress, future research trends, and potential risks of 3D technologies are also discussed.

There is fast growing demand for antennas and sensors directly integrated into products including smartphones, wearables, structural components, and other products related to the Internet of Things (IoT). While IoT applications are varied, the common requirement for such “Smart Connected Devices” is an inherent need for sensors (“smart”) and antennas (“connected”). Emerging 3-Dimensional “Direct Write” approaches for creating antenna and sensors enable rapid design and prototyping, reduce manufacturing steps and associated costs, and broaden the choice of target substrate shapes and materials. Optomec has developed a high volume printing solution for the production of 3D antenna and 3D sensors that are tightly integrated with an underlying product ranging from Smartphones to Industrial Components. Aerosol Jet® is a non-contact, room temperature deposition method that is used to print a wide variety of materials, including conductive, semiconducting, ceramic, CNT, graphene and other functional inks suitable for the fabrication of antenna and sensor that are directly integrated onto existing 3D products and structures. Case studies involving direct printing of various antenna (ie: Broadband, Bluetooth, NFC, etc) and sensors (ie: Strain, Creep, etc.) will be presented. Material considerations, including performance and reliability metrics will be discussed; and examples of via filling, wrap-around printing, and five axis motion will be shown.

Industrial robots are mostly designed for tasks where working conditions are well known and do not change over time. However, there are many examples where robots require vision sensor-based control in order to perform complex operations and react to changes in the environment. Other areas where there is need for combining vision sensors with manipulators are quality inspection tasks of large or complex structured objects. The paper presents an overview of the methods used in robotic vision systems, taking into account their specific requirements and advantages. Experimental studies of a selected 2D/3D sensors mounted on industrial manipulators are presented including analysis of collected 3D points clouds. Additionally, a system of a reconfigurable manipulator that interchangeably uses various sensors and tools in complex inspection tasks is presented.

In this paper, we propose an algorithm for the automatic registration of two rigid parts using multiple 3D sensor systems on a robot. Four sets of structured laser stripe system consisted of a camera and a visible laser stripe is used for the acquisition of 3D information. Detailed procedures including extrinsic calibration among four 3D sensor systems and hand/eye calibration of 3D sensing system on robot arm are presented. We find a best pose using search-based pose estimation algorithm where cost function is proposed by reflecting geometric constraints between sensor systems and target objects. A pose with minimum gap and height difference is found by greedy search. Experimental result using demo system shows the robustness and feasibility of the proposed algorithm.

Underground displacement monitoring is a key means to monitor and evaluate geological disasters and geotechnical projects. There exist few practical instruments able to monitor subsurface horizontal and vertical displacements simultaneously due to monitoring invisibility and complexity. A novel underground displacement 3D measuring sensor had been proposed in our previous studies, and great efforts have been taken in the basic theoretical research of underground displacement sensing and measuring characteristics by virtue of modeling, simulation and experiments. This paper presents an innovative underground displacement joint inversion method by mixing a specific forward modeling approach with an approximate optimization inversion procedure. It can realize a joint inversion of underground horizontal displacement and vertical displacement for the proposed 3D sensor. Comparative studies have been conducted between the measured and inversed parameters of underground horizontal and vertical displacements under a variety of experimental and inverse conditions. The results showed that when experimentally measured horizontal displacements and vertical displacements are both varied within 0 ~ 30 mm, horizontal displacement and vertical displacement inversion discrepancies are generally less than 3 mm and 1 mm, respectively, under three kinds of simulated underground displacement monitoring circumstances. This implies that our proposed underground displacement joint inversion method is robust and efficient to predict the measuring values of underground horizontal and vertical displacements for the proposed sensor.

Obstacle detection in a greenhouse using the Kinect 3D sensor.Depth is processed into slope information.Obstacle decision is based on color, texture and slope.The system is real-time and provides good results. In many agricultural robotic applications, the robotic vehicle must detect obstacles in its way in order to navigate correctly. This is also true for robotic spray vehicles that autonomously explore greenhouses. In this study, we present an approach to obstacle detection in a greenhouse environment using the Kinect 3D sensor, which provides synchronized color and depth information. First, the depth data are processed by applying slope computation. This creates an obstacle map which labels the pixels that exceed a predefined slope as suspected obstacles, and others as surface. Then, the system uses both color and texture features to classify the suspected obstacle pixels. The obstacle detection decision is made using information on the pixel slope, its intensity and surrounding neighbor pixels. The obstacle detection of the proposed sensor and algorithm is demonstrated on data recorded by the Kinect in a greenhouse. We show that the system produces satisfactory results (all obstacles were detected with only few false positive detections) and is fast enough to run on a limited computer.

In this paper, we present a new algorithm that utilizes low-quality red, green, blue and depth (RGB-D) data from the Kinect sensor for face recognition under challenging conditions. This algorithm extracts multiple features and fuses them at the feature level. A Finer Feature Fusion technique is developed that removes redundant information and retains only the meaningful features for possible maximum class separability. We also introduce a new 3D face database acquired with the Kinect sensor which has released to the research community. This database contains over 5,000 facial images (RGB-D) of 52 individuals under varying pose, expression, illumination and occlusions. Under the first three variations and using only the noisy depth data, the proposed algorithm can achieve 72.5 % recognition rate which is significantly higher than the 41.9 % achieved by the baseline LDA method. Combined with the texture information, 91.3 % recognition rate has achieved under illumination, pose and expression variations. These results suggest the feasibility of low-cost 3D sensors for real-time face recognition.

This paper is focused on the problem of repetitive strain injuries in hands. These injuries are mostly related to professional activities where people are subjected to a high rate of work and the performed tasks often lead to repetitive actions. The objective of this paper is to develop a serious game to prevent strain injuries in hands. The game scenarios promote warm-up and stretching off hand exercises that should be executed before and after the working period. The game is developed in Unity software, associated with a 3D sensor, Intel RealSense 3D Camera F200 for hand and movements detection. With the activity implementation in companies and establishments the employees will be able to exercise their hands, thus reducing the risk of being affected by strain injuries.

This paper presents an inverse kinematic based control system of a novel self-developed humanoid robot using a Microsoft Kinect as 3D sensor. Based on the processed sensory data from the Kinect, human gestures and motions are captured, analyzed and mapped into joint variables by using inverse kinematic equations. A significant improvement in the inverse kinematic algorithm has been made to achieve better perception of human gestures. The performance of the proposed system is evaluated in a newly developed test-bed under real-world conditions, which effectively shows perception-related error control in order to further improve the system.

Hoy en día el desarrollo de nuevos productos enfocados a individuos que presentan alguna limitación es un punto clave en el proceso de inclusión y aprendizaje, ya que gran parte de la información se encuentra en medios virtuales. En la actualidad existen diversos dispositivos alternativos de entrada para computador entre los cuales se encuentran los denominados Free-hand, los cuales permiten interactuar de forma directa sin utilizar los medios tradicionales como lo son el teclado y el ratón. Dado este contexto, a pesar de que existe una gran cantidad de dispositivos capaces de detectar el movimiento y reconocer gestos, no se cuenta con muchos desarrollos específicos sobre el área de la inclusión social a personas con limitaciones motoras, las cuales permitan el acceso de forma sencilla a los medios de aprendizaje y comunicación. En esta investigación se propone diseñar un sistema de detección de movimientos usando técnicas de visión artificial 3D, con el fin de mejorar la interacción de personas con movilidad reducida. Abstract: Nowadays the development of new products focused on individuals that present some limitation is a key point in the process of inclusion and learning, since much of the information is found in virtual media. At present there are several alternative computer input devices among which are the so-called Free-hand, which allow direct interaction without using traditional media such as the keyboard and mouse. Given this context, although there is a large number of devices capable of detecting movement and recognizing gestures, there are not many specific developments in the area of social inclusion for people with motor limitations, which allow easy access to the means of learning and communication. In this research it is proposed to design a movement detection system using 3D artificial vision techniques, in order to improve the interaction of people with reduced mobility. Maestría

Hoy en día el desarrollo de nuevos productos enfocados a individuos que presentan alguna limitación es un punto clave en el proceso de inclusión y aprendizaje, ya que gran parte de la información se encuentra en medios virtuales. En la actualidad existen diversos dispositivos alternativos de entrada para computador entre los cuales se encuentran los denominados Free-hand, los cuales permiten interactuar de forma directa sin utilizar los medios tradicionales como lo son el teclado y el ratón. Dado este contexto, a pesar de que existe una gran cantidad de dispositivos capaces de detectar el movimiento y reconocer gestos, no se cuenta con muchos desarrollos específicos sobre el área de la inclusión social a personas con limitaciones motoras, las cuales permitan el acceso de forma sencilla a los medios de aprendizaje y comunicación. En esta investigación se propone diseñar un sistema de detección de movimientos usando técnicas de visión artificial 3D, con el fin de mejorar la interacción de personas con movilidad reducida. Abstract: Nowadays the development of new products focused on individuals that present some limitation is a key point in the process of inclusion and learning, since much of the information is found in virtual media. At present there are several alternative computer input devices among which are the so-called Free-hand, which allow direct interaction without using traditional media such as the keyboard and mouse. Given this context, although there is a large number of devices capable of detecting movement and recognizing gestures, there are not many specific developments in the area of social inclusion for people with motor limitations, which allow easy access to the means of learning and communication. In this research it is proposed to design a movement detection system using 3D artificial vision techniques, in order to improve the interaction of people with reduced mobility. Maestría

What is an essential difference between the expert movement and the novice one at where the movement plays important role such as dancing, playing darts, or preparing tea at Japanese tea ceremony? We would like to figure out it from the movement of the human joints at an activity. Which will give us a hint how to improve our behavior. In our study, Structure Sensor, a Kinect-like 3D sensor, is used to track and record 3-dimensional positions of the human joints during an action. The device is portable and does not require any sensors nor markers attached to observed persons. We then investigate the effectiveness of a statistical approach. Results of preliminary experiments established for dancing will be reported.

Repetitive movements in hands may cause strain injuries that negatively influence the employee performance. The lack of awareness for prevention this type of working injuries often leads the employers to believe that costs associated to prevention are unnecessary expenditures. So, the objective of this paper is to develop a serious game for preventing repetitive strain injuries in hands. The game activities are developed in Unity 3D software and using the 3D sensor, Intel RealSense 3D Camera F200. There are two sets of exercises, the warming up and the stretching off exercises, to be executed during the playing activity. Three game scenarios were developed. The tests performed so far allow concluding that the proposed system correctly detects the hands movements.

Calibration and registration are two challenging tasks required by a real-time inspection system for long steel products based on multiple three-dimensional sensors. Spatial calibration is required to transform the information acquired by each sensor into a common reference system. Registration is used to align the calculated shape with the model of the product being inspected, which enables the calculation of quality control parameters. This paper proposes efficient solutions for these two problems. Calibration is performed using a specially designed calibration plate that can be used to detect the position and orientation of each sensor. Registration is applied using an analytic alignment between the data and the model. The proposed methods are specially designed to perform accurately in industrial environments where inspection systems are exposed to adverse conditions, such as dust, oil, or water, which provoke noise in the signals. Moreover, vibrations affect the position of the inspected product randomly, which makes accurate measurements much more difficult to achieve. The proposed procedures are validated with extensive tests using real and simulated data showing excellent performance.

Current society has observed an increasing number of victims of neurological disease, with reduced mobility, leading to a necessity to perform physical therapy to optimize their quality of life. This action results in physiotherapeutic programs filled with repetitive exercises, often fastidious, that lead to the demotivation of patients and consequent poor adherence and withdrawal. As a result of the technological evolution, new tools such as serious games are emerging, so their use in the field of physical therapy can modify the way patients face their treatments, promoting their motivation. Thus, we have developed a serious game based on image processing techniques to motivate and monitor patients with neurological diseases in their physical therapy practice.

The aim of this work is to investigate the problem of static hand gesture recognition using Kinect v2 depth sensor. We have dealt not only with the theory of recognition itself, but also with the use of depth sensors and the latest approaches in this field. We have successfully applied different convolutional neural network architectures to this classification problem and evaluated the impact of kernel size on the recognition score. By creating our own static gesture database, we have gained an objective way of comparing each test method.

3D foot scanner is a non-contact type device that measures the shape of a foot using laser triangulation sensors. In general, a laser triangulation sensor is equipped with a single camera and a line-laser. Using only one laser triangulation sensor is insufficient to acquire accurate surface information of the full foot. In addition, the single camera in the laser triangulation sensor could cause occlusions when the projected laser line is blocked by protruding surface areas of the foot. To overcome these problems, modern 3D foot scanners use multiple laser triangulation sensors that consist of dual cameras. Most of the general 3D foot scanners use a total of eight cameras and four line-lasers making the device costly. In this paper, we propose an architecture of a 360' rotation-based 3D foot scanner that consists of a single rotating-type 3D sensor and a linear-type 3D sensor. The rotating-type 3D sensor consists of two line-lasers and a single camera and acquires 360' surface information of the upper part of the foot. The linear-type 3D sensor consists of a single line-laser and a camera and acquires sole surface information. Since the proposed 3D foot scanner has only two cameras and three line-lasers, it can be used as a cheap alternative to many existing models.

The history of sensors array (SA) X-ray radiography receptors development shows a big possibility in characteristics optimization and stable position of SA receptors on world market. New generation of SA receptors has 3D architecture. That allows to optimize parameters of CMOS SA receptors for fluoroscopic, linear and cone beam tomography units. Extra big field of view and frame rate (more than 300 fps) of CMOS SA receptors gives unique possibility for system development.

An important step for the classification of plants in organic farming is the separation of the objects. In our approach a 3D camera will be used for this task. One problem is that most of the available 3D sensors are not suitable due to their bad resolution, low-speed and high costs. The Kinect II is an affordable alternative but designed for another workspace. In this paper it is shown why a 3D sensor is required for the separation of objects in organic farming and how the modification of a Kinect II 3D camera increases the range resolution to solve the given problem.

In order to increase the measurement speed of pattern projection-based three-dimensional (3-D) sensors, in 2014, we introduced the so-called array projector which allows pattern projection at several 1,000 fps. As the patterns are switched by switching on and off the light sources of multiple slide projectors, each pattern originates from a different projection center. This may lead to a 3-D point deviation when measuring glossy objects. In this contribution, we theoretically and experimentally investigate the dependence of this deviation on the measurement object's reflective properties. Furthermore, we propose a procedure for compensating for this deviation.

Marine seismic data is typically acquired by using towed-streamers due to their high efficiency, easy implementation, and narrow-azimuth (NAZ) acquisition geometry. This method usually results in good-quality seismic data for noncomplex geological environments and is a well-established methodology. However, this method is limited in obstructed areas or in areas with very shallow water where streamer vessels cannot be used. We present two field tests of autonomous marine vehicles (AMVs) equipped with 3D seismic sensor arrays to record marine seismic data concurrently with OBC seismic acquisition missions ongoing in the same area. The field tests evaluated the operational aspects of the AMVs and demonstrated that the vehicles equipped with 3D arrays can be used to acquire seismic data that is usable and comparable with seismic data obtained from the associated ocean-bottom cable (OBC) surveys. The vehicles proved their ability to hold station and record and transmit QC data back to the survey vessel. The ability to use the vehicles to acquire data in areas previously hard to reach opens the door to potentially deploy the vehicles during streamer surveys to acquire data with offset in excess of 20 km. Shots can be recorded by the vehicle placed at predetermined locations to record ultra-long offset data, which can assist in building velocity models with full waveform inversion. The AMV is capable of carrying out a number of functions over a longer time frame and at lower cost than when using traditional methods. Its ability to maintain position in a holding area or travel to a new location, when needed, through GPS coordinates optimizes seismic surveys without compromising results. As offshore exploration and production move into deeper and more remote areas, the seismic industry can increasingly rely on autonomous vehicles for support.

Mapping the interior of a building is a nice step to obtain the description of that building. In that context, this article presents the use of a low cost 3D sensor, the Microsoft Kinect for Windows v1, associated to RGBDSLAM in order to perform that mapping. Even if the method is well-known and has showed its performance, its use is not as simple as it looks like. Limitations of both the sensor and the method have to be dealt with and great care has to be taken during the measurements. Some post-processing is also needed to obtain nice point cloud as results of the experiments. Emphasis is put on the point cloud' size and the storage needed. Finally, the data obtained after an internship measurement campaign are presented and show to fulfill nicely the goals and requirements of the study.

X-ray digital receptor structure in a form of array partial picture forming sensors (sensors array — SA), allows to optimize parameters of X-ray receptors for fluoroscopic, linear and cone beam tomography units. Analysis efficiency of sensor arrays, based on detective quantum efficiency is the subject of this article.

To guarantee autonomous and safely navigation for a mobile robot, the processing achieved for its localization must be fast and accurate enough to enable the robot to perform high-level tasks for navigation and obstacle avoidance. The authors of Simultaneous Localization And Mapping (SLAM) based works, are trying since year, to ensure the speed/accuracy trade-off. Most existing works in the field of monocular (SLAM) has largely centered around sparse feature-based representations of the environment. By tracking salient image points across many frames of video, both the positions of the features and the motion of the camera can be inferred live. Within the visual SLAM community, there has been a focus on both increasing the number of features that can be tracked across an image and efficiently managing and adjusting this map of features in order to improve camera trajectory and feature location accuracy. However, visual SLAM suffers from some limitations. Indeed, with a single camera and without any assumptions or prior knowledge about the camera environment, rotation can be retrieved, but the translation is up to scale. Furthermore, visual monocular SLAM is an incremental process prone to small drifts in both pose measurement and scale, which when integrated over time, become increasingly significant over large distances.To cope with these limitations, we have centered our work around the following issues : integrate additional information into an existing monocular visual SLAM system, in order to constrain the camera localization and the mapping points. Provided that the high speed of the initial SLAM process is kept and the lack of these added constraints should not give rise to the failure of the process. For these last reasons, we have chosen to integrate the depth information provided by a 3D sensor (e.g. Microsoft Kinect) and geometric information about scene structure.The primary contribution of this work consists of modifying the SLAM algorithm proposed by Mouragnon et al. (2006b) to take into account the depth measurement provided by a 3D sensor. This consists of several rather straightforward changes, but also on a way to combine the depth and visual data in the bundle adjustment process. The second contribution is to propose a solution that uses, in addition to the depth and visual data, the constraints lying on points belonging to the plans of the scene.The proposed solutions have been validated on a synthetic sequences as well as on a real sequences, which depict various environments. These solutions have been compared to the state of art methods. The performances obtained with the previous solutions demonstrate that the additional constraints developed, improves significantly the accuracy and the robustness of the SLAM localization. Furthermore, these solutions are easy to roll out and not much time consuming.; Pour assurer la navigation autonome d’un robot mobile, les traitements effectués pour sa localisation doivent être faits en ligne et doivent garantir une précision suffisante pour permettre au robot d’effectuer des tâches de haut niveau pour la navigation et l’évitement d’obstacles. Les auteurs de travaux basés sur le SLAM visuel (Simultaneous Localization And Mapping) tentent depuis quelques années de garantir le meilleur compromis rapidité/précision. La majorité des solutions SLAM visuel existantes sont basées sur une représentation éparse de l’environnement. En suivant des primitives visuelles sur plusieurs images, il est possible d’estimer la position 3D de ces primitives ainsi que les poses de la caméra. La communauté du SLAM visuel a concentré ses efforts sur l’augmentation du nombre de primitives visuelles suivies et sur l’ajustement de la carte 3D, afin d’améliorer l’estimation de la trajectoire de la caméra et les positions 3D des primitives. Cependant, la localisation par SLAM visuel présente souvent des dérives dues au cumul d’erreurs, et dans le cas du SLAM visuel monoculaire, la position de la caméra n’est connue qu’à un facteur d’échelle près. Ce dernier peut être fixé initialement mais dérive au cours du temps.Pour faire face à ces limitations, nous avons centré nos travaux de thèse sur la problématique suivante : intégrer des informations supplémentaires dans un algorithme de SLAM visuel monoculaire afin de mieux contraindre la trajectoire de la caméra et la reconstruction 3D. Ces contraintes ne doivent pas détériorer les performances calculatoires de l’algorithme initial et leur absence ne doit pas mettre l’algorithme en échec. C’est pour cela que nous avons choisi d’intégrer l’information de profondeur fournie par un capteur 3D (e.g. Microsoft Kinect) et des informations géométriques sur la structure de la scène.La première contribution de cette thèse est de modifier l’algorithme SLAM visuel monoculaire proposé par Mouragnon et al. (2006b) pour prendre en compte la mesure de profondeur fournie par un capteur 3D, en proposant particulièrement un ajustement de faisceaux qui combine, d’une manière simple, des informations visuelles et des informations de profondeur. La deuxième contribution est de proposer une nouvelle fonction de coût du même ajustement de faisceaux qui intègre, en plus des contraintes sur les profondeurs des points, des contraintes géométriques d’appartenance aux plans de la scène.Les solutions proposées ont été validées sur des séquences de synthèse et sur des séquences réelles, représentant des environnements variés. Ces solutions ont été comparées aux récentes méthodes de l’état de l’art. Les résultats obtenus montrent que les différentes contraintes développées permettent d’améliorer significativement la précision de la localisation du SLAM. De plus les solutions proposées sont faciles à déployer et peu couteuses en temps de calcul.

To guarantee autonomous and safely navigation for a mobile robot, the processing achieved for its localization must be fast and accurate enough to enable the robot to perform high-level tasks for navigation and obstacle avoidance. The authors of Simultaneous Localization And Mapping (SLAM) based works, are trying since year, to ensure the speed/accuracy trade-off. Most existing works in the field of monocular (SLAM) has largely centered around sparse feature-based representations of the environment. By tracking salient image points across many frames of video, both the positions of the features and the motion of the camera can be inferred live. Within the visual SLAM community, there has been a focus on both increasing the number of features that can be tracked across an image and efficiently managing and adjusting this map of features in order to improve camera trajectory and feature location accuracy. However, visual SLAM suffers from some limitations. Indeed, with a single camera and without any assumptions or prior knowledge about the camera environment, rotation can be retrieved, but the translation is up to scale. Furthermore, visual monocular SLAM is an incremental process prone to small drifts in both pose measurement and scale, which when integrated over time, become increasingly significant over large distances.To cope with these limitations, we have centered our work around the following issues : integrate additional information into an existing monocular visual SLAM system, in order to constrain the camera localization and the mapping points. Provided that the high speed of the initial SLAM process is kept and the lack of these added constraints should not give rise to the failure of the process. For these last reasons, we have chosen to integrate the depth information provided by a 3D sensor (e.g. Microsoft Kinect) and geometric information about scene structure.The primary contribution of this work consists of modifying the SLAM algorithm proposed by Mouragnon et al. (2006b) to take into account the depth measurement provided by a 3D sensor. This consists of several rather straightforward changes, but also on a way to combine the depth and visual data in the bundle adjustment process. The second contribution is to propose a solution that uses, in addition to the depth and visual data, the constraints lying on points belonging to the plans of the scene.The proposed solutions have been validated on a synthetic sequences as well as on a real sequences, which depict various environments. These solutions have been compared to the state of art methods. The performances obtained with the previous solutions demonstrate that the additional constraints developed, improves significantly the accuracy and the robustness of the SLAM localization. Furthermore, these solutions are easy to roll out and not much time consuming.; Pour assurer la navigation autonome d’un robot mobile, les traitements effectués pour sa localisation doivent être faits en ligne et doivent garantir une précision suffisante pour permettre au robot d’effectuer des tâches de haut niveau pour la navigation et l’évitement d’obstacles. Les auteurs de travaux basés sur le SLAM visuel (Simultaneous Localization And Mapping) tentent depuis quelques années de garantir le meilleur compromis rapidité/précision. La majorité des solutions SLAM visuel existantes sont basées sur une représentation éparse de l’environnement. En suivant des primitives visuelles sur plusieurs images, il est possible d’estimer la position 3D de ces primitives ainsi que les poses de la caméra. La communauté du SLAM visuel a concentré ses efforts sur l’augmentation du nombre de primitives visuelles suivies et sur l’ajustement de la carte 3D, afin d’améliorer l’estimation de la trajectoire de la caméra et les positions 3D des primitives. Cependant, la localisation par SLAM visuel présente souvent des dérives dues au cumul d’erreurs, et dans le cas du SLAM visuel monoculaire, la position de la caméra n’est connue qu’à un facteur d’échelle près. Ce dernier peut être fixé initialement mais dérive au cours du temps.Pour faire face à ces limitations, nous avons centré nos travaux de thèse sur la problématique suivante : intégrer des informations supplémentaires dans un algorithme de SLAM visuel monoculaire afin de mieux contraindre la trajectoire de la caméra et la reconstruction 3D. Ces contraintes ne doivent pas détériorer les performances calculatoires de l’algorithme initial et leur absence ne doit pas mettre l’algorithme en échec. C’est pour cela que nous avons choisi d’intégrer l’information de profondeur fournie par un capteur 3D (e.g. Microsoft Kinect) et des informations géométriques sur la structure de la scène.La première contribution de cette thèse est de modifier l’algorithme SLAM visuel monoculaire proposé par Mouragnon et al. (2006b) pour prendre en compte la mesure de profondeur fournie par un capteur 3D, en proposant particulièrement un ajustement de faisceaux qui combine, d’une manière simple, des informations visuelles et des informations de profondeur. La deuxième contribution est de proposer une nouvelle fonction de coût du même ajustement de faisceaux qui intègre, en plus des contraintes sur les profondeurs des points, des contraintes géométriques d’appartenance aux plans de la scène.Les solutions proposées ont été validées sur des séquences de synthèse et sur des séquences réelles, représentant des environnements variés. Ces solutions ont été comparées aux récentes méthodes de l’état de l’art. Les résultats obtenus montrent que les différentes contraintes développées permettent d’améliorer significativement la précision de la localisation du SLAM. De plus les solutions proposées sont faciles à déployer et peu couteuses en temps de calcul.

Feature matching forms the basis for numerous computer vision applications. With the rapid development of 3D sensors, the availability of RGB-D images has been increased stably. Compared to traditional 2D images, the additional depth images in RGB-D images can provide more geometric information. In this paper, we propose a new efficient binary descriptor (namely BAG) for RGB-D image representation by combining appearance and geometric cues. Experimental results show that the proposed BAG descriptor produces better feature matching performance with faster matching speed and less memory than the existing methods.

Recently, the interest in development of service robots highly increases. The annual turnover in this area is estimated to a number close to 17 milliard Euro in the past years. The annual turnover could rise to 0.1 billion Euro by 2020 by IFR considering 30 % growth every year. Investments are expected to flow into all areas related to service robotics, mainly into the development of assistant robots for seniors and the handicapped people, health monitoring and surgical robots, robots in agriculture, pilotless drones and helicopters and ground vehicles without a driver. Very promising years seems to be coming for all the new and already existing companies focused on this area with their software and sensor engineers together with producers of important accessories. © The Authors, published by EDP Sciences, 2017.

Modern research in mobile robotics proposes to combine localization and perception in order to recognize previously visited locations and thus to improve localization as well as the object recognition processes recursively. A crucial issue is to perform updates of the scene geometry when novel observations become available. The reason is that a practical application often requires a system to model large 3D environments at high resolution which exceeds the storage of the local memory. The underlying work presents an optimized volume data structure for infinite 3D environments which facilitates (i) successive world model updates without the need to recompute the full dataset, (ii) very fast in-memory data access scheme enabling the integration of high resolution 3D sensors in real-time, (iii) efficient level-of-detail for visualization and coarse geometry updates. The technique is finally demonstrated on real world application scenarios which underpin the feasibility of the research outcomes.

Each individual describes unique patterns during their gait cycles. This information can be extracted from the live video stream and used for subject identification. In appearance based recognition methods, this is done by tracking silhouettes of persons across gait cycles. In recent years, there has been a profusion of sensors that in addition to RGB video images also provide depth data in real-time. When such sensors are used for gait recognition, existing RGB appearance based methods can be extended to get a substantial gain in recognition accuracy. In this paper, this is accomplished using information fusion techniques that combine features from extracted silhouettes, used in traditional appearance based methods, and the height feature that can now be estimated using depth data. The latter is estimated during the silhouette extraction step with minimal additional computational cost. Two approaches are proposed that can be implemented easily as an extension to existing appearance based methods. An extensive experimental evaluation was performed to provide insights into how much the recognition accuracy can be improved. The results are presented and discussed considering different types of subjects and populations of different height distributions.

Guo, G., 2019. Video image acquisition technology of underwater vehicle based on 3D sensor. In: Hoang, A.T. and Aqeel Ashraf, M. (eds.), Research, Monitoring, and Engineering of Coastal, P...

In this work, a novel salient point descriptor for 3D point clouds, called Covariance Matrix Pyramids (CMPs), is presented. With CMPs it is possible to compare unstructured and unequal numbers of points which is an important characteristic when working with point clouds. Corresponding points from different scans are matched in a pyramidal approach combined with Particle Swarm Optimization. The flexibility of CMPs is demonstrated on the basis of several databases with objects, such as 3D faces, 3D apples, 3D kitchen scenes, 3D human–machine interaction gesture sequences, and 3D buildings all recorded with different 3D sensors. Quantitative results are given and compared with other state-of-the-art descriptors, whereby CMPs show promising performance.

ndustrial robot can extend its capability adopting sensors such as camera and laser range finder. In this paper, we deal with object manipulation including grasping objects on the table and inserting them into the hole on the vertical plane. Kinect is used as the 3D sensor. First, object location and target location is computed after processing 3D data. Robot motion obtained by solving inverse kinematics can have errors so that we adopt visual feedback style approach. Chessboard type marker is attached on the end-effector of the robot and its location with respect to the target is used to correct the robot motion.

For 3D sensor networks to monitor the ocean, forest and aerosphere, etc., sensors can forward their data to the base station by greedy routing. It is critical to quantitatively greedy routing’s deliverability for evaluating network’s functionality and performance. The probability that all sensors can successfully send their data to the base station by greedy routing is usually modeled as the probability of guaranteed delivery. For a typical spherical cap 3D sensor network deployment scenario where nodes follow a homogeneous Poisson point process, the relationship between the sensor transmission radius and the probability of guaranteed delivery is studied, and a tight analytical upper bound on the sensor transmission radius to ensure the designed deliverability probability is derived in this paper. The correctness and tightness of the derived upper bound are verified by extensive simulations.

Recently, cheap 3D sensor such as Kinect is widely used in diverse applications. In this paper, we use Kinect as the 3D sensor for the object manipulation using robot arm. Our target is insert objects on the horizontal table into the corresponding location on the hole in the vertical plate. First, we find two dominant planes corresponding to the horizontal and vertical table by processing 3D points by Kinect. Then, we find each object on the horizontal table using contour processing. Also, object on the vertical plane is processed using the same algorithm. Finally, object matching is done using Hu invariant moments.

We report on experimental results and TCAD simulations addressing the impact of layout on the electrical characteristics of double-sided 3D diodes fabricated at Fondazione Bruno Kessler (FBK), Trento, Italy. Simulations are found to accurately reproduce the device characteristics, thus explaining the basic mechanisms governing the breakdown behavior and capacitance of different devices and providing useful hints for layout optimization.

Roller skis are used by cross-country skiers, biathletes and ski-orienteers for their snow-free training and in roller ski competitions. Additionally, much of the current sports research into the physiology and biomechanics of cross- country skiing is conducted indoors on treadmills using roller skis. For elite athletes, the differences in performance are quite small, thus emphasising the importance of knowing the roller skis’ rolling resistance coefficient, especially in connection to research and roller ski competitions. The purpose of this study was to develop a roller ski rolling resistance measurement system (P-RRMS) that is portable and therefore useful in different contexts and locations. The P-RRMS was designed as a small treadmill, equipped with roller ski stabilizing lateral supports and a screwed bar for applying different vertical loads on the roller ski. The design uses only one force sensor, with possible measurements of three directions of force and torque around three axes. The weight of the P-RRMS is 100kg and it is equipped with wheels to facilitate transportation.