Abstract
Fiducial markers play a fundamental role in various fields in which precise localization and tracking are paramount. In Augmented Reality, they provide a known reference point in the physical world so that AR systems can accurately identify, track, and overlay virtual objects. This accuracy is essential for creating a seamless and immersive AR experience, particularly when prompted to cope with the sub-millimeter requirements of medical and industrial applications. This research article presents a comparative analysis of four fiducial marker tracking algorithms, aiming to assess and benchmark their accuracy and precision. The proposed methodology compares the pose estimated by four algorithms running on Hololens 2 with those provided by a highly accurate ground truth system. Each fiducial marker was positioned in 25 sampling points with different distances and orientations. The proposed evaluation method is not influenced by human error, relying only on a high-frequency and accurate motion tracking system as ground truth. This research shows that it is possible to track the fiducial markers with translation and rotation errors as low as 1.36 mm and 0.015 degrees using ArUco and Vuforia, respectively.

Abstract
Welding is a challenging, risky, and time-consuming profession. Recently, there has been a documented shortage of trained welders, and as a result, the market is pushing for an increase in the rate at which new professionals are trained. To address this growing demand, training institutions are exploring alternative methods to train future professionals. The emergence of virtual reality technologies has led to initiatives to explore their potential for welding training. Multiple studies have suggested that virtual reality training delivers comparable, or even superior, results when compared to more conventional approaches, with shorter training times and reduced costs in consumables. This paper conducts a comprehensive review of the current state of the field of welding simulators. This involves exploring the different types of welding simulators available and evaluating their effectiveness and efficiency in meeting the learning objectives of welding training. The aim is to identify gaps in the literature, suggest future research directions, and promote the development of more effective and efficient welding simulators in the future. The research also seeks to develop a categorical system for evaluating and comparing welding simulators. This system will enable a more systematic and objective analysis of the features and characteristics of each simulator, identifying the essential characteristics that should be included in each level of classification.

Abstract
Machine vision systems are widely used in assembly lines for providing sensing abilities to robots to allow them to handle dynamic environments. This paper presents a comparison of 3D sensors for evaluating which one is best suited for usage in a machine vision system for robotic fastening operations within an automotive assembly line. The perception system is necessary for taking into account the position uncertainty that arises from the vehicles being transported in an aerial conveyor. Three sensors with different working principles were compared, namely laser triangulation (SICK TriSpector1030), structured light with sequential stripe patterns (Photoneo PhoXi S) and structured light with infrared speckle pattern (Asus Xtion Pro Live). The accuracy of the sensors was measured by computing the root mean square error (RMSE) of the point cloud registrations between their scans and two types of reference point clouds, namely, CAD files and 3D sensor scans. Overall, the RMSE was lower when using sensor scans, with the SICK TriSpector1030 achieving the best results (0.25 mm +/- 0.03 mm), the Photoneo PhoXi S having the intermediate performance (0.49 mm +/- 0.14 mm) and the Asus Xtion Pro Live obtaining the higher RMSE (1.01 mm +/- 0.11 mm). Considering the use case requirements, the final machine vision system relied on the SICK TriSpector1030 sensor and was integrated with a collaborative robot, which was successfully deployed in an vehicle assembly line, achieving 94% success in 53,400 screwing operations.

Abstract
The pose estimation of a mobile robotic system is essential in many autonomous applications. Inertial sensors provide high-frequency measurements that can be used to estimate the displacement, however, for estimating the orientation, an additional filter is required. Some of the newest Attitude and Heading Reference Systems can provide a referenced estimation of the orientation of the device, allowing it to retrieve the orientation of a robotic system. However, magnetic field perturbations caused by ferromagnetic objects or induced magnetic fields might influence these systems and, consequently, lead to the accumulation of errors over time. In this paper, the performance of the Xsens fusion filter is compared with a stateof-the-art algorithm to estimate the orientation of the system under dynamic movements and in the presence of magnetic perturbations, with the goal of finding the most suitable for an Unmanned Aerial Vehicle. The results show that both filters are robust and perform well in the target scenario, with a root mean squared error between 2 and 5 degrees; however, the Xsens fusion filter does not require an extra computer to process the data.

Abstract
The Multi-View Stereo (MVS) is a key process in the photogrammetry workflow. It is responsible for taking the camera's views and finding the maximum number of matches between the images yielding a dense point cloud of the observed scene. Since this process is based on the matching between images it greatly depends on the abil-ity of features matching throughout different images. To improve the matching performance several researchers have proposed the use of Convolutional Neural Networks (CNNs) to solve the MVS problem. Despite the progress in the MVS problem with the usage of CNNs, the Video RAM (VRAM) consumption within these approaches is usually far greater than classical methods, that rely more on RAM, which is cheaper to expand than VRAM. This work then follows the progress made in CasMVSNet in the reduction of GPU memory usage, and further study the changes in the feature extraction process. The Average Group-wise Correlation is used in the cost vol-ume generation, to reduce the number of channels in the cost volume, yielding a reduction in GPU memory usage without noticeable penalties in the result. The deformable convolutions are applied in the feature extraction net -work to augment the spatial sampling locations with learning offsets, without additional supervision, to further improve the network's ability to model transformations. The impact of these changes is measured using quanti-tative and qualitative tests using the DTU and the Tanks and Temples datasets. The modifications reduced the GPU memory usage by 32% and improved the completeness by 9% with a penalty of 6.6% in accuracy on the DTU dataset.(c) 2021 Published by Elsevier B.V.