Skip to main content

Resultados relacionados

Resultados totales (Incluyendo duplicados): 13
Encontrada(s) 1 página(s)

Investigo. Repositorio Institucional de la Universidade de Vigo
Publication Artículo científico (JournalArticle)

Multimodal deep learning for point cloud panoptic segmentation of railway environments

  • Grandio Gonzalez, Javier
  • Riveiro Rodríguez, Belén
  • Lamas Novoa, Daniel
  • Arias Sánchez, Pedro
The demand for transportation asset digitalisation has significantly increased over the years. For this purpose, mobile mapping systems (MMSs) are among the most popular technologies that allow capturing high precision three-dimensional point clouds of the infrastructure. In this paper, a multimodal deep learning methodology is presented for panoptic segmentation of the railway infrastructure. The methodology takes advantage of image rasterisation of the point clouds to perform a rough segmentation and discard more than 80% of points that are not relevant to the infrastructure. With this approach, the computational requirements for processing the remaining point cloud are highly reduced, allowing the process of dense point clouds in short periods of time. A 90 km-long railway scenario was used for training and testing. The proposed methodology is two times faster than the current state-of-the-art for the same point cloud density, and pole-like object segmentation metrics are improved., Fundación BBVA, Agencia Estatal de Investigación | Ref. PID2019-108816RB-I00, Ministerio de Universidades | Ref. FPU20/01024, Universidade de Vigo/CISUG
Investigo. Repositorio Institucional de la Universidade de Vigo
Publication Artículo científico (JournalArticle)

Assessment of vegetation indices for mapping burned areas using a deep learning method and a comprehensive forest fire dataset from Landsat collection

  • Suárez Fernández, Gabriel Eduardo
  • Martínez Sánchez, Joaquín
  • Arias Sánchez, Pedro
Forested areas, crucial for their multifunctional roles, face significant risks from wildfires. Recent methods have enabled accurate delineation of burned areas (BA) using satellite images, which is essential for assessing impacts and recovery. Nevertheless, many such techniques rely on short time series for training data and utilize many predictors without thorough analysis of efficient computational frameworks. Therefore, this paper presents a novel method that builds a forest fire dataset while minimizing technical and technological costs, while also assessing the benefits of incorporating vegetation spectral indices (VIs) versus specific spectral bands in a Convolutional Neural Network (CNN) detector. The methodology involves creating a dataset of real forest fires from 1985 to 2021 using two consecutive Landsat images and applying VIs and unsupervised clustering techniques. The dataset is then used to explore the feasibility of integrating VIs from single-temporal Landsat images into a U-Net-based CNN, measuring performance across multiple settings. A real dataset was successfully acquired, achieving a 73.68 % level of agreement with the limited years available in the Spanish Ministry’s official records, while the analysis of the introduction of VIs into a CNN revealed that using single-temporal images with the Landsat blue band and VIs achieved a 73 % Intersection over Union (IoU) metric, leading to a 46 % reduction in storage requirements. In contrast, using the green band with the blue band and VIs resulted in a 75 % IoU and a 31 % reduction in storage, comparable to configurations using visible and infrared bands without VIs. Consequently, this methodology successfully collects a dataset of BA from wildfires, extending the digitized time scale while demonstrating the advantages of applying VIs, rather than specific spectral bands, in a CNN for delineating burned areas, achieving data optimization and normalization., Ministerio de Universidades | Ref. FPU21/03038, Agencia Estatal de Investigación | Ref. TED2021-132000B-I00, Universidade de Vigo/CISUG
Investigo. Repositorio Institucional de la Universidade de Vigo
Publication Artículo científico (JournalArticle)

Automatic detection to inventory road slopes using open LiDAR point clouds

  • Rúa Fernández, Erik
  • Núñez Seoane, Antón
  • Arias Sánchez, Pedro
  • Martínez Sánchez, Joaquín
The transport infrastructure of a country facilitates the development and growth of its economy and improves the quality of life of its inhabitants. Increasing its resilience to different types of risks to improve performance is becoming more important. In the current context of climate change, natural hazards are more severe and frequent. In this article, we focus on rockfall as a natural hazard for roads that occurs in small areas in the vicinity of natural or cut slopes, causing road safety problems by invading part of the road. This article aims to inventory the slopes along the road, identifying the area of the road which would be invaded in case of a rockfall. A methodology divided into two blocks is proposed. First, for slope detection and inventory, an algorithm is developed based on open LiDAR point clouds analysis. The second block consists of estimating the invaded road area if a rockfall occurs on each of the inventoried slopes, using a combination of RockGIS software and the Monte Carlo method. The methodology was applied in five case studies: three sections on motorways and two sections on national roads. The results obtained for slope detection show higher rates in the case studies analyzing motorways, with a precision of 100%, a recovery rate of greater than 93.4%, and an F1 score of greater than 0.96. The results in the invaded area of the road show that 11 slopes would cause a total cut of the motorway in one of the directions if a rockfall occurs. These results are useful for infrastructure managers to remotely obtain an inventory of road slopes and know which of them would affect road safety. Also, the results can serve as input for the Intelligent Transportation System and allow the exchange of information under the Building Information Model approach., Agencia Estatal de Investigación | Ref. PID2019-108816RB-I00, Agencia Estatal de Investigación | Ref. PRE2020-096222, European Commission | Ref. H2020, n. 955337
Investigo. Repositorio Institucional de la Universidade de Vigo
Publication Artículo científico (JournalArticle)

Transport infrastructure management based on LiDAR synthetic data: a deep learning approach with a ROADSENSE simulator

  • Comesaña Cebral, Lino Jose
  • Martínez Sánchez, Joaquín
  • Núñez Seoane, Antón
  • Arias Sánchez, Pedro
In the realm of transportation system management, various remote sensing techniques have proven instrumental in enhancing safety, mobility, and overall resilience. Among these techniques, Light Detection and Ranging (LiDAR) has emerged as a prevalent method for object detection, facilitating the comprehensive monitoring of environmental and infrastructure assets in transportation environments. Currently, the application of Artificial Intelligence (AI)-based methods, particularly in the domain of semantic segmentation of 3D LiDAR point clouds by Deep Learning (DL) models, is a powerful method for supporting the management of both infrastructure and vegetation in road environments. In this context, there is a lack of open labeled datasets that are suitable for training Deep Neural Networks (DNNs) in transportation scenarios, so, to fill this gap, we introduce ROADSENSE (Road and Scenic Environment Simulation), an open-access 3D scene simulator that generates synthetic datasets with labeled point clouds. We assess its functionality by adapting and training a state-of-the-art DL-based semantic classifier, PointNet++, with synthetic data generated by both ROADSENSE and the well-known HELIOS++ (HEildelberg LiDAR Operations Simulator). To evaluate the resulting trained models, we apply both DNNs on real point clouds and demonstrate their effectiveness in both roadway and forest environments. While the differences are minor, the best mean intersection over union (MIoU) values for highway and national roads are over 77%, which are obtained with the DNN trained on HELIOS++ point clouds, and the best classification performance in forested areas is over 92%, which is obtained with the model trained on ROADSENSE point clouds. This work contributes information on a valuable tool for advancing DL applications in transportation scenarios, offering insights and solutions for improved road and roadside management., Agencia Estatal de Investigación | Ref. PID2022-140662OB-I00, Agencia Estatal de Investigación | Ref. PCI2020-120705-2-I00
Investigo. Repositorio Institucional de la Universidade de Vigo
Publication Artículo científico (JournalArticle)

A discordance analysis in manual labelling of urban mobile laser scanning data used for deep learning based semantic segmentation

  • Gonzalez Collazo, Silvia María
  • Balado Frías, Jesús
  • González Rodríguez, Maria Elena
  • Nurunnabi, Abdul
Labelled point clouds are crucial to train supervised Deep Learning (DL) methods used for semantic segmentation. The objective of this research is to quantify discordances between the labels made by different people in order to assess whether such discordances can influence the success rates of a DL based semantic segmentation algorithm. An urban point cloud of 30 m road length in Santiago de Compostela (Spain) was labelled two times by ten persons. Discordances and its significance in manual labelling between individuals and rounds were calculated. In addition, a ratio test to signify discordance and concordance was proposed. Results show that most of the points were labelled accordingly with the same class by all the people. However, there were many points that were labelled with two or more classes. Class curb presented 5.9% of discordant points and 3.2 discordances for each point with concordance by all people. In addition, the percentage of significative labelling differences of the class curb was 86.7% comparing all the people in the same round and 100% comparing rounds of each person. Analysing the semantic segmentation results with a DL based algorithm, PointNet++, the percentage of concordance points are related with F-score value in R2 = 0.765, posing that manual labelling has significant impact on results of DL-based semantic segmentation methods., Xunta de Galicia | Ref. ED481B-2019-061, Ministerio de Ciencia e Innovación | Ref. PID2019-105221RB-C43, Ministère de l’Economie of the G. D. of Luxembourg | Ref. SOLSTICE 2019-05-030-24, Universidade de Vigo/CISUG
Investigo. Repositorio Institucional de la Universidade de Vigo
Publication Artículo científico (JournalArticle)

Qualitative fault tree for the analysis of slope stability loss in road infrastructure

  • Rúa Fernández, Erik
  • Martínez Sánchez, Joaquín
  • Arias Sánchez, Pedro
TRA Lisbon 2022 Conference Proceedings Transport Research Arena (TRA Lisbon 2022),14th-17th November 2022, Lisboa, Portugal, Landslides are natural hazards all over the world, being the loss of slope stability in transport infrastructures considered small-scale landslides causing economic and human losses. Although there are many examples in the literature for the general analysis of landslides susceptibility, little information is available to analyze slopes at a local scale. The objective of this work is to identify all the slopes in road infrastructure, select those located in landslide susceptibility areas and apply a qualitative fault tree, also defined in this work, to determine the factors which can cause the slope stability loss. The results obtained are useful for each critical slope., Agencia Estatal de Investigación | Ref. PID2019-108816RB-I00, Agencia Estatal de Investigación | Ref. PRE2020-096222
Investigo. Repositorio Institucional de la Universidade de Vigo
Publication Artículo científico (JournalArticle)

Analysis of the mechanical behaviour of an innovative removable clamps-based connection system for steel square and rectangular profiles

  • Cabaleiro Núñez, Manuel
  • González Gaya, Cristina
  • Conde Carnero, Borja
  • Caamaño Martinez, Jose Carlos
This work studies a new 90º connection system based on clamps for a square or rectangular steel profile, which can be used on existing structures without any prior or invasive operations. The main field where this connection system could be used would be in industrial plants, since it allows the reuse of industrial structures as many times as necessary, depending on each layout change, thus obtaining significant economic savings. This work aims to analyse the behaviour of this new type of connection using a 3D FEM-based numerical analysis procedure, previously validated against experimental tests. In this research, it has been possible to identify four behavioural modes, the first of which, linear behaviour, is the most suitable for the operational work of this type of connection. The distance between the bolts, one of the most important variables affecting the mechanical response, is parameterised and analysed in a numerical model. According to the results, the strength of the joint improves as the distance between the bolts increases, but this also means larger clamps and the need for more space., Agencia Estatal de Investigación | Ref. TED2021-130497A-I00
Investigo. Repositorio Institucional de la Universidade de Vigo
Publication Artículo científico (JournalArticle)

Analysis of sun glare on roundabouts with aerial laser scanning data

  • Gonzalez Collazo, Silvia María
  • Balado Frías, Jesús
  • González, E.
  • Soilán Rodríguez, Mario
Road geometry and sun glares play an important role concerning road safety. In this research, the direct sunlight in a roundabout sited in Ávila (Spain) is analysed using Aerial Laser Scanning (ALS) point clouds. First, the roundabout is divided in 8 sections, obtaining the driver bearing vectors of the roundabout. Entrances and exits driver bearing vectors of the roundabout are also considered. Then, sun rays are generated for a specific location of the roundabout and in a specific day and time. The incidence of the sun rays with the driver’s vision angle is analysed based on human vision model. Finally, intersections of sun rays with obstacles are calculated utilizing ALS point clouds. ALS data is processed (removing outliers, reducing point density, and computing a Delaunay Triangulation) in order to obtain accurate intersection results with obstacles and optimise the computational time. The method was tested in a roundabout, considering different driver bearings, the slope of the road and the elevation of the terrain. The results show that sun glares are detected at any day and time of the year, therefore areas with risk of direct sun glare within the roundabout are identified. The sun ray’s incidence in the vision angle of the driver is higher during winter solstice, and intersections with obstacles occur mainly during sunrise and sunset. In roundabout vector 7, during winter solstice there is direct sun glare for 7 hours 30 minutes, at the equinoxes for 6 hours 15 minutes and during summer solstice there is no direct sun glare., Xunta de Galicia | Ref. ED481B-2019-061, Xunta de Galicia | Ref. ED431C 2020/01, Ministerio de Ciencia e Innovación | Ref. PID2019-105221RB-C43, Ministerio de Ciencia e Innovación | Ref. TIN2016-77158 -C4-2-R, Ministerio de Ciencia e Innovación | Ref. FJC2018-035550-I
Investigo. Repositorio Institucional de la Universidade de Vigo
Publication Artículo científico (JournalArticle)

Comparing Mobile and Aerial Laser Scanner point cloud data sets for automating the detection and delimitation procedure of safety-critical near-road slopes

  • Núñez Seoane, Antón
  • Martínez Sánchez, Joaquín
  • Rúa Fernández, Erik
  • Arias Sánchez, Pedro
An inappropriately maintained road cut-slope is likely to fail, resulting in landslides or falling rocks that compromise road safety. Thus, road managers need to know the location of dangerous slopes along the road in order to prevent these events from happening. In this article, we compare two different approaches for conducting the digitization of the road environment and the automatic detection and delimitation of road slopes: Mobile Laser Scanners (MLS) and Aerial Laser Scanners (ALS). The point clouds obtained using the first kind of devices are dense, rich in detail and generated from a ground perspective; the second type of scanners produce less dense clouds from a zenithal perspective. We explore what is the effect of the point cloud density and scanner point of view over the slope detection procedure. Two road segments from the Spanish A55 and A52 highways were used as study zones, and a total of 28.61 km were analyzed. Better detection and delimitation results were achieved when using the ALS data and its corresponding algorithm. It was observed that the higher point density and detail of the MLS clouds were not an advantage for the slope detection task, and that measuring the road from a terrestrial perspective affected in a negative way during the detection process: the crest of the slopes often remained unmeasured, hidden behind vegetation or man-made elements, thus resulting in the slopes not being complete in the MLS clouds. Meanwhile, the whole slope structure is scanned when the scene is measured from an aerial perspective, henceforth obtaining better detection rates despite the relatively low resolution. The findings of this study provide valuable information in the field of road asset management, and help road managers make decisions when choosing what technology to use for the data gathering process., Agencia Estatal de Investigación | Ref. PID2022-140662OB-I00, Universidade de Vigo/CISUG
Investigo. Repositorio Institucional de la Universidade de Vigo
Publication Artículo científico (JournalArticle)

Santiago urban dataset SUD: Combination of Handheld and Mobile Laser Scanning point clouds

  • Gonzalez Collazo, Silvia María
  • Balado Frías, Jesús
  • Garrido González, Iván
  • Grandio Gonzalez, Javier
  • Rashdi , Rabia
  • Tsiranidou , Elisavet
  • Del Rio Barral, Pablo
  • Rúa Fernández, Erik
  • Puente Luna, Iván
  • Lorenzo Cimadevila, Henrique Remixio
Santiago Urban Dataset SUD is a real dataset that combines Mobile Laser Scanning (MLS) and Handheld Mobile Laser Scanning (HMLS) point clouds. The data is composed by 2 km of streets, sited in Santiago de Compostela (Spain). Point clouds undergo a manual labelling process supported by both heuristic and Deep Learning methods, resulting in the classification of eight specific classes: road, sidewalk, curb, buildings, vehicles, vegetation, poles, and others. Three PointNet++ models were trained; the first one using MLS point clouds, the second one with HMLS point clouds and the third one with both H&MLS point clouds. In order to ascertain the quality and efficacy of each Deep Learning model, various metrics were employed, including confusion matrices, precision, recall, F1-score, and IoU. The results are consistent with other state-of-the-art works and indicate that SUD is valid for comparing point cloud semantic segmentation works. Furthermore, the survey's extensive coverage and the limited occlusions indicate the potential utility of SUD in urban mobility research., Agencia Estatal de Investigación | Ref. PID2019-105221RB-C43, Xunta de Galicia | Ref. ED481B-2019-061, Xunta de Galicia | Ref. ED431C 2020/01, Universidade de Vigo/CISUG
Investigo. Repositorio Institucional de la Universidade de Vigo
Publication Artículo científico (JournalArticle)

Detection of direct sun glare on drivers from point clouds

  • Gonzalez Collazo, Silvia María
  • Del Rio Barral, Pablo
  • Balado Frías, Jesús
  • González Rodríguez, Maria Elena
Sunlight conditions can reduce drivers’ visibility, which is a safety concern on road networks. This research introduces a method to study sun glare incidence in road environments. Sun glare areas during daylight hours are automatically detected from mobile laser scanning (MLS) and aerial laser scanning (ALS) point clouds. The method comprises the following steps. First, the Sun’s position (solar altitude and azimuth) referring to a location is calculated. Second, the incidence of sun glare with the user’s angle of vision is analyzed based on human vision. Third, sun ray intersections with near obstacles (vegetation, building, etc.) are calculated utilizing MLS point clouds. Finally, intersections with distant obstacles (mountains) are calculated utilizing ALS point clouds. MLS and ALS data are processed in order to combine both data types, remove outliers, and optimize computational time for intersection searches (point density reduction and Delaunay triangulation). The method was tested on two real case studies, covering roads with different bearings, slopes, and surroundings. The combination of MLS and ALS data, together with the solar geometry, identify areas of risk for the visibility of drivers. Consequently, the proposed method can be utilized to reduce sun glare, implementing warnings in navigation systems., Xunta de Galicia | Ref. ED481B-2019-061, Xunta de Galicia | Ref. ED431C 2020/01, Agencia Estatal de Investigación | Ref. PID2019-105221RB-C43, Ministerio de Economía, Industria y Competitividad | Ref. TIN2016-77158- C4-2-R
Investigo. Repositorio Institucional de la Universidade de Vigo
Publication Artículo científico (JournalArticle)

Gamification for road asset inspection from Mobile Mapping System data

  • Barros Sobrín, Álvaro
  • Balado Frías, Jesús
  • Soilán Rodríguez, Mario
  • Mingueza Bauzá, Enrique
Gamification techniques have been proven effective in various fields such as education and industry. In this paper, we introduce a novel approach that applies gamification techniques to the identification of road assets in Mobile Laser Scanning (MLS) data. Our method utilises three gamification techniques: avatar (vehicle), point cloud segmentation into levels, and scoring. We implemented these techniques in Unreal Engine and evaluated their performance using three real-world case studies. We also compared two ways of point cloud visualisation: mesh-based and point-based. Our results demonstrate that our gamification approach improves the handling and visualisation of point clouds when compared to other free software such as Cloud Compare. Specifically, the point-based visualisation method provides a more accurate representation of the road environment and the input point cloud and is easier to import into Unreal Engine. However, this method requires more computational resources for visualisation. On the other hand, level segmentation ensures a constant frame rate of 60 frames per second. Furthermore, our gamification approach enhances the experience of road asset identification, making it more enjoyable for the user. However, we acknowledge that the quality of the point cloud remains the primary factor affecting the accuracy of asset identification, regardless of the software used. Overall, our proposed gamification approach offers a promising solution for improving the identification of road assets in MLS data and has the potential to be applied to other fields beyond road asset identification., Xunta de Galicia | Ref. ED481B-2019-061, Agencia Estatal de Investigación | Ref. PID2019-105221RB-C43, Agencia Estatal de Investigación | Ref. RYC2021‐033560‐I
Investigo. Repositorio Institucional de la Universidade de Vigo
Publication Artículo científico (JournalArticle)

Multi feature-rich synthetic colour to improve human visual perception of point clouds

  • Balado Frías, Jesús
  • González Rodríguez, Maria Elena
  • Rodríguez Somoza, Juan Luis
  • Arias Sánchez, Pedro
Although point features have shown their usefulness in classification with Machine Learning, point cloud visualization enhancement methods focus mainly on lighting. The visualization of point features helps to improve the perception of the 3D environment. This paper proposes Multi Feature-Rich Synthetic Colour (MFRSC) as an alternative non-photorealistic colour approach of natural-coloured point clouds. The method is based on the selection of nine features (reflectance, return number, inclination, depth, height, point density, linearity, planarity, and scattering) associated with five human perception descriptors (edges, texture, shape, size, depth, orientation). The features are reduced to fit the RGB display channels. All feature permutations are analysed according to colour distance with the natural-coloured point cloud and Image Quality Assessment. As a result, the selected feature permutations allow a clear visualization of the scene's rendering objects, highlighting edges, planes, and volumetric objects. MFRSC effectively replaces natural colour, even with less distorted visualization according to BRISQUE, NIQUE and PIQE. In addition, the assignment of features in RGB channels enables the use of MFRSC in software that does not support colorization based on point attributes (most commercially available software). MFRSC can be combined with other non-photorealistic techniques such as Eye-Dome Lighting or Ambient Occlusion., Xunta de Galicia | Ref. ED481B-2019-061, Xunta de Galicia | Ref. ED431F 2022/08, Agencia Estatal de Investigación | Ref. PID2019-105221RB-C43, Universidade de Vigo/CISUG