scholarly journals An Automated Training of Deep Learning Networks by 3D Virtual Models for Object Recognition

Symmetry ◽  
2019 ◽  
Vol 11 (4) ◽  
pp. 496 ◽  
Author(s):  
Kamil Židek ◽  
Peter Lazorík ◽  
Ján Piteľ ◽  
Alexander Hošovský
Keyword(s):  
Deep Learning ◽  
Object Recognition ◽  
Augmented Reality ◽  
3D Models ◽  
Processing Unit ◽  
Assembly Process ◽  
Learning Networks ◽  
Embedded Devices ◽  
New Approach ◽  
Virtual Models

Small series production with a high level of variability is not suitable for full automation. So, a manual assembly process must be used, which can be improved by cooperative robots and assisted by augmented reality devices. The assisted assembly process needs reliable object recognition implementation. Currently used technologies with markers do not work reliably with objects without distinctive texture, for example, screws, nuts, and washers (single colored parts). The methodology presented in the paper introduces a new approach to object detection using deep learning networks trained remotely by 3D virtual models. Remote web application generates training input datasets from virtual 3D models. This new approach was evaluated by two different neural network models (Faster RCNN Inception v2 with SSD, MobileNet V2 with SSD). The main advantage of this approach is the very fast preparation of the 2D sample training dataset from virtual 3D models. The whole process can run in Cloud. The experiments were conducted with standard parts (nuts, screws, washers) and the recognition precision achieved was comparable with training by real samples. The learned models were tested by two different embedded devices with an Android operating system: Virtual Reality (VR) glasses, Cardboard (Samsung S7), and Augmented Reality (AR) smart glasses (Epson Moverio M350). The recognition processing delays of the learned models running in embedded devices based on an ARM processor and standard x86 processing unit were also tested for performance comparison.

scholarly journals Augmented Reality Maintenance Assistant Using YOLOv5

2021 ◽  
Vol 11 (11) ◽  
pp. 4758
Author(s):  
Ana Malta ◽  
Mateus Mendes ◽  
Torres Farinha
Keyword(s):  
Neural Network ◽  
Deep Learning ◽  
Object Recognition ◽  
Augmented Reality ◽  
Real Time ◽  
Recognition System ◽  
High Accuracy ◽  
Video Streams ◽  
The Neural Network ◽  
Deep Learning Neural Network

Maintenance professionals and other technical staff regularly need to learn to identify new parts in car engines and other equipment. The present work proposes a model of a task assistant based on a deep learning neural network. A YOLOv5 network is used for recognizing some of the constituent parts of an automobile. A dataset of car engine images was created and eight car parts were marked in the images. Then, the neural network was trained to detect each part. The results show that YOLOv5s is able to successfully detect the parts in real time video streams, with high accuracy, thus being useful as an aid to train professionals learning to deal with new equipment using augmented reality. The architecture of an object recognition system using augmented reality glasses is also designed.

Geologic Modelling Using Augmented Reality

2021 ◽  
Author(s):  
Phathompat Boonyasaknanon ◽  
Raymond Pols ◽  
Katja Schulze ◽  
Robert Rundle
Keyword(s):  
Augmented Reality ◽  
Real Time ◽  
3D Models ◽  
New Approach ◽  
Domain Experts ◽  
3D Objects ◽  
Track Record ◽  
Initial Work ◽  
Geologic Model ◽  
Time Critical

Abstract An augmented reality (AR) system is presented which enhances the real-time collaboration of domain experts involved in the geologic modeling of complex reservoirs. An evaluation of traditional techniques is compared with this new approach. The objective of geologic modeling is to describe the subsurface as accurately and in as much detail as possible given the available data. This is necessarily an iterative process since as new wells are drilled more data becomes available which either validates current assumptions or forces a re-evaluation of the model. As the speed of reservoir development increases there is a need for expeditious updates of the subsurface model as working with an outdated model can lead to costly mistakes. Common practice is for a geologist to maintain the geologic model while working closely with other domain experts who are frequently not co-located with the geologist. Time-critical analysis can be hampered by the fact that reservoirs, which are inherently 3D objects, are traditionally viewed with 2D screens. The system presented here allows the geologic model to be rendered as a hologram in multiple locations to allow domain experts to collaborate and analyze the reservoir in real-time. Collaboration on 3D models has not changed significantly in a generation. For co-located personnel the approach is to gather around a 2D screen. For remote personnel the approach has been sharing a model through a 2D screen along with video chat. These approaches are not optimal for many reasons. Over the years various attempts have been tried to enhance the collaboration experience and have all fallen short. In particular virtual reality (VR) has been seen as a solution to this problem. However, we have found that augmented reality (AR) is a much better solution for many subtle reasons which are explored in the paper. AR has already acquired an impressive track record in various industries. AR will have applications in nearly all industries. For various historical reasons, the uptake for AR is much faster in some industries than others. It is too early to tell whether the use of augmented reality in geological applications will be transformative, however the results of this initial work are promising.

scholarly journals 3D Fine-scale Terrain Variables from Underwater Photogrammetry: A New Approach to Benthic Microhabitat Modeling in a Circalittoral Rocky Shelf

2020 ◽  
Vol 12 (15) ◽  
pp. 2466
Author(s):  
Elena Prado ◽  
Augusto Rodríguez-Basalo ◽  
Adolfo Cobo ◽  
Pilar Ríos ◽  
Francisco Sánchez
Keyword(s):  
Deep Learning ◽  
Learning Algorithm ◽  
High Reliability ◽  
3D Models ◽  
Additive Models ◽  
Fine Scale ◽  
New Approach ◽  
Deep Learning Algorithm ◽  
The Relationship ◽  
Very High

The relationship between 3D terrain complexity and fine-scale localization and distribution of species is poorly understood. Here we present a very fine-scale 3D reconstruction model of three zones of circalittoral rocky shelf in the Bay of Biscay. Detailed terrain variables are extracted from 3D models using a structure-from-motion (SfM) approach applied to ROTV images. Significant terrain variables that explain species location were selected using general additive models (GAMs) and micro-distribution of the species were predicted. Two models combining BPI, curvature and rugosity can explain 55% and 77% of the Ophiuroidea and Crinoidea distribution, respectively. The third model contributes to explaining the terrain variables that induce the localization of Dendrophyllia cornigera. GAM univariate models detect the terrain variables for each structural species in this third zone (Artemisina transiens, D. cornigera and Phakellia ventilabrum). To avoid the time-consuming task of manual annotation of presence, a deep-learning algorithm (YOLO v4) is proposed. This approach achieves very high reliability and low uncertainty in automatic object detection, identification and location. These new advances applied to underwater imagery (SfM and deep-learning) can resolve the very-high resolution information needed for predictive microhabitat modeling in a very complex zone.

scholarly journals Augmented Reality Application for Laptop Assembly with Assembly Complexity Study

2021 ◽  
Author(s):  
Chiew Jin Hong ◽  
Aun Naa Aun Sung
Keyword(s):  
User Interface ◽  
Speech Recognition ◽  
Augmented Reality ◽  
3D Models ◽  
Assembly Process ◽  
Assembly Sequence ◽  
Augmented Reality Application ◽  
Assembly Instructions ◽  
Guided Assembly ◽  
High Level

Abstract Augmented Reality (AR) in the assembly process will improve the user's experience by providing interactive instructions in real time. However, no previous application of AR guided assembly for laptops with a high level of assembly complexity has been developed. The research aims to develop an AR guided assembly application to provide instruction on the assembly of a laptop. The assembly complexity of the laptop was also investigated. The development of the AR application involves the creation of model target, 3D models and animations, and the development of user interface. The laptop assembly consists of ten steps. Each step comprises animated 3D models and text detailing the assembly instructions. Speech recognition has been used to navigate the assembly sequence. The AR application has successfully been developed for laptop assembly with an assembly complexity of 6.63. With the developed application, the performance of the laptop assembly can be accelerated.

scholarly journals UAVs in rail damage image diagnostics supported by deep-learning networks

2021 ◽  
Vol 11 (1) ◽  
pp. 339-348
Author(s):  
Piotr Bojarczak ◽  
Piotr Lesiak
Keyword(s):  
Deep Learning ◽  
Head Width ◽  
Learning Networks ◽  
Image Brightness ◽  
Learning Network ◽  
Rail Head ◽  
Image Recording ◽  
Python Language ◽  
Deep Learning Network ◽  
Efficiency Rate

Abstract The article uses images from Unmanned Aerial Vehicles (UAVs) for rail diagnostics. The main advantage of such a solution compared to traditional surveys performed with measuring vehicles is the elimination of decreased train traffic. The authors, in the study, limited themselves to the diagnosis of hazardous split defects in rails. An algorithm has been proposed to detect them with an efficiency rate of about 81% for defects not less than 6.9% of the rail head width. It uses the FCN-8 deep-learning network, implemented in the Tensorflow environment, to extract the rail head by image segmentation. Using this type of network for segmentation increases the resistance of the algorithm to changes in the recorded rail image brightness. This is of fundamental importance in the case of variable conditions for image recording by UAVs. The detection of these defects in the rail head is performed using an algorithm in the Python language and the OpenCV library. To locate the defect, it uses the contour of a separate rail head together with a rectangle circumscribed around it. The use of UAVs together with artificial intelligence to detect split defects is an important element of novelty presented in this work.

Sign in / Sign up

Export Citation Format

Share Document