scholarly journals A DEEP LEARNING FRAMEWORK FOR ROADS NETWORK DAMAGE ASSESSMENT USING POST-EARTHQUAKE LIDAR DATA

Author(s):  
S. T. Seydi ◽  
H. Rastiveis

Abstract. Roads network are the most important parts of urban infrastructures, which can cause difficulty to the city whenever they undergo a problem. This paper aims to provide and implement a deep learning-based method to determine the status of the streets network after an earthquake using LiDAR point cloud. The proposed framework composes of three main phases: (1) Deep features of LiDAR data are extracted using a Convolutional Neural Network (CNN). (2) The extracted features are used in a multilayer perceptron (MLP) neural network in which debris areas inside the road network are detected. (3) The amount of debris in each road is applied to damage index for classifying the road segments into blocked or un-blocked. To evaluate the efficiency of the proposed framework, LiDAR point cloud of the Port-au-Prince, Haiti after the 2010 Haiti earthquake was used. The overall accuracy of more than 97% proved the high performance of this framework for debris detection. Moreover, analyzing damage assessment of 37 road segments based on the detected debris and comparing to a visually generated damaged map, 31 of the road segments were correctly labelled as either blocked or un-blocked.

Nutrients ◽  
2018 ◽  
Vol 10 (12) ◽  
pp. 2005 ◽  
Author(s):  
Frank Lo ◽  
Yingnan Sun ◽  
Jianing Qiu ◽  
Benny Lo

An objective dietary assessment system can help users to understand their dietary behavior and enable targeted interventions to address underlying health problems. To accurately quantify dietary intake, measurement of the portion size or food volume is required. For volume estimation, previous research studies mostly focused on using model-based or stereo-based approaches which rely on manual intervention or require users to capture multiple frames from different viewing angles which can be tedious. In this paper, a view synthesis approach based on deep learning is proposed to reconstruct 3D point clouds of food items and estimate the volume from a single depth image. A distinct neural network is designed to use a depth image from one viewing angle to predict another depth image captured from the corresponding opposite viewing angle. The whole 3D point cloud map is then reconstructed by fusing the initial data points with the synthesized points of the object items through the proposed point cloud completion and Iterative Closest Point (ICP) algorithms. Furthermore, a database with depth images of food object items captured from different viewing angles is constructed with image rendering and used to validate the proposed neural network. The methodology is then evaluated by comparing the volume estimated by the synthesized 3D point cloud with the ground truth volume of the object items.


2020 ◽  
Vol 11 (28) ◽  
pp. 7335-7348 ◽  
Author(s):  
Timothy E. H. Allen ◽  
Andrew J. Wedlake ◽  
Elena Gelžinytė ◽  
Charles Gong ◽  
Jonathan M. Goodman ◽  
...  

Deep learning neural networks, constructed for the prediction of chemical binding at 79 pharmacologically important human biological targets, show extremely high performance on test data (accuracy 92.2 ± 4.2%, MCC 0.814 ± 0.093, ROC-AUC 0.96 ± 0.04).


2018 ◽  
Vol 246 ◽  
pp. 03044 ◽  
Author(s):  
Guozhao Zeng ◽  
Xiao Hu ◽  
Yueyue Chen

Convolutional Neural Networks (CNNs) have become the most advanced algorithms for deep learning. They are widely used in image processing, object detection and automatic translation. As the demand for CNNs continues to increase, the platforms on which they are deployed continue to expand. As an excellent low-power, high-performance, embedded solution, Digital Signal Processor (DSP) is used frequently in many key areas. This paper attempts to deploy the CNN to Texas Instruments (TI)’s TMS320C6678 multi-core DSP and optimize the main operations (convolution) to accommodate the DSP structure. The efficiency of the improved convolution operation has increased by tens of times.


Author(s):  
Yoichiro Maeda ◽  
Kotaro Sano ◽  
Eric W. Cooper ◽  
Katsuari Kamei ◽  
◽  
...  

In recent years, much research on the unmanned control of a moving vehicle has been conducted, and various robots and motor vehicles moving automatically are being used. However, the more complicated the environment is, the more difficult it is for the autonomous vehicles to move automatically. Even in such a challenging environment, however, an expert with the necessary operation skill can sometimes perform the appropriate control of the moving vehicle. In this research, a method for learning a human’s operation skill using a convolutional neural network (CNN) and setting visual information for input is proposed for learning more complicated environmental information. A CNN is a kind of deep-learning network, and it exhibits high performance in the field of image recognition. In this experiment, the operation knowledge was also visualized using a fuzzy neural network with obtained input-output maps to create fuzzy rules. To verify the effectiveness of this method, an experiment involving operation skill acquisition by some subjects using a drone control simulator was conducted.


Author(s):  
M. Yadav ◽  
B. Lohani ◽  
A. K. Singh

<p><strong>Abstract.</strong> The accurate three-dimensional road surface information is highly useful for health assessment and maintenance of roads. It is basic information for further analysis in several applications including road surface settlement, pavement condition assessment and slope collapse. Mobile LiDAR system (MLS) is frequently used now a days to collect detail road surface and its surrounding information in terms three-dimensional (3D) point cloud. Extraction of road surface from volumetric point cloud data is still in infancy stage because of heavy data processing requirement and the complexity in the road environment. The extraction of roads especially rural road, where road-curb is not present is very tedious job especially in Indian roadway settings. Only a few studies are available, and none for Indian roads, in the literature for rural road detection. The limitations of existing studies are in terms of their lower accuracy, very slow speed of data processing and detection of other objects having similar characteristics as the road surface. A fast and accurate method is proposed for LiDAR data points of road surface detection, keeping in mind the essence of road surface extraction especially for Indian rural roads. The Mobile LiDAR data in <i>XYZI</i> format is used as input in the proposed method. First square gridding is performed and ground points are roughly extracted. Then planar surface detection using mathematical framework of principal component analysis (PCA) is performed and further road surface points are detected using similarity in intensity and height difference of road surface pointe in their neighbourhood.</p><p>A case study was performed on the MLS data points captured along wide-street (two-lane road without curb) of 156<span class="thinspace"></span>m length along rural roadway site in the outskirt of Bengaluru city (South-West of India). The proposed algorithm was implemented on the MLS data of test site and its performance was evaluated it terms of recall, precision and overall accuracy that were 95.27%, 98.85% and 94.23%, respectively. The algorithm was found computationally time efficient. A 7.6 million MLS data points of size 27.1<span class="thinspace"></span>MB from test site were processed in 24 minutes using the available computational resources. The proposed method is found to work even for worst case scenarios, i.e., complex road environments and rural roads, where road boundary is not clear and generally merged with road-side features.</p>


2019 ◽  
Vol 1 ◽  
pp. 1-1
Author(s):  
Tee-Ann Teo

<p><strong>Abstract.</strong> Deep Learning is a kind of Machine Learning technology which utilizing the deep neural network to learn a promising model from a large training data set. Convolutional Neural Network (CNN) has been successfully applied in image segmentation and classification with high accuracy results. The CNN applies multiple kernels (also called filters) to extract image features via image convolution. It is able to determine multiscale features through the multiple layers of convolution and pooling processes. The variety of training data plays an important role to determine a reliable CNN model. The benchmarking training data for road mark extraction is mainly focused on close-range imagery because it is easier to obtain a close-range image rather than an airborne image. For example, KITTI Vision Benchmark Suite. This study aims to transfer the road mark training data from mobile lidar system to aerial orthoimage in Fully Convolutional Networks (FCN). The transformation of the training data from ground-based system to airborne system may reduce the effort of producing a large training data set.</p><p>This study uses FCN technology and aerial orthoimage to localize road marks on the road regions. The road regions are first extracted from 2-D large-scale vector map. The input aerial orthoimage is 10&amp;thinsp;cm spatial resolution and the non-road regions are masked out before the road mark localization. The training data are road mark’s polygons, which are originally digitized from ground-based mobile lidar and prepared for the road mark extraction using mobile mapping system. This study reuses these training data and applies them for the road mark extraction using aerial orthoimage. The digitized training road marks are then transformed to road polygon based on mapping coordinates. As the detail of ground-based lidar is much better than the airborne system, the partially occulted parking lot in aerial orthoimage can also be obtained from the ground-based system. The labels (also called annotations) for FCN include road region, non-regions and road mark. The size of a training batch is 500&amp;thinsp;pixel by 500&amp;thinsp;pixel (50&amp;thinsp;m by 50&amp;thinsp;m on the ground), and the total number of training batches for training is 75 batches. After the FCN training stage, an independent aerial orthoimage (Figure 1a) is applied to predict the road marks. The results of FCN provide initial regions for road marks (Figure 1b). Usually, road marks show higher reflectance than road asphalts. Therefore, this study uses this characteristic to refine the road marks (Figure 1c) by a binary classification inside the initial road mark’s region.</p><p>To compare the automatically extracted road marks (Figure 1c) and manually digitized road marks (Figure 1d), most road marks can be extracted using the training set from ground-based system. This study also selects an area of 600&amp;thinsp;m&amp;thinsp;&amp;times;&amp;thinsp;200&amp;thinsp;m in quantitative analysis. Among the 371 reference road marks, 332 can be extracted from proposed scheme, and the completeness reached 89%. The preliminary experiment demonstrated that most road marks can be successfully extracted by the proposed scheme. Therefore, the training data from the ground-based mapping system can be utilized in airborne orthoimage in similar spatial resolution.</p>


A rapid dissemination of Android operating system in smart phone market has resulted in an exponential growth of threats to mobile applications. Various studies have been carried out in academia and industry for the identification and classification of malicious applications using machine learning and deep learning algorithms. Convolution Neural Network is a deep learning technique which has gained popularity in speech and image recognition. The conventional solution for identifying Android malware needs learning based on pre-extracted features to preserve high performance for detecting Android malware. In order to reduce the efforts and domain expertise involved in hand-feature engineering, we have generated the grayscale images of AndroidManifest.xml and classes.dex files which are extracted from the Android package and applied Convolution Neural Network for classifying the images. The experiments are conducted on a recent dataset of 1747 malicious Android applications. The results indicate that classes.dex file gives better results as compared to the AndroidManifest.xml and also demonstrate that model performs better as the image become larger.


Author(s):  
Hatem Keshk ◽  
Xu-Cheng Yin

Background: Deep Learning (DL) neural network methods have become a hotspot subject of research in the remote sensing field. Classification of aerial satellite images depends on spectral content, which is a challenging topic in remote sensing. Objective: With the aim to accomplish a high performance and accuracy of Egyptsat-1 satellite image classification, the use of the Convolutional Neural Network (CNN) is raised in this paper because CNN is considered a leading deep learning method. CNN is developed to classify aerial photographs into land cover classes such as urban, vegetation, desert, water bodies, soil, roads, etc. In our work, a comparison between MAXIMUM Likelihood (ML) which represents the traditional supervised classification methods and CNN method is conducted. Conclusion: This research finds that CNN outperforms ML by 9%. The convolutional neural network has better classification result, which reached 92.25% as its average accuracy. Also, the experiments showed that the convolutional neural network is the most satisfactory and effective classification method applied to classify Egyptsat-1 satellite images.


2018 ◽  
Vol 10 (9) ◽  
pp. 1461 ◽  
Author(s):  
Yongyang Xu ◽  
Zhong Xie ◽  
Yaxing Feng ◽  
Zhanlong Chen

The road network plays an important role in the modern traffic system; as development occurs, the road structure changes frequently. Owing to the advancements in the field of high-resolution remote sensing, and the success of semantic segmentation success using deep learning in computer version, extracting the road network from high-resolution remote sensing imagery is becoming increasingly popular, and has become a new tool to update the geospatial database. Considering that the training dataset of the deep convolutional neural network will be clipped to a fixed size, which lead to the roads run through each sample, and that different kinds of road types have different widths, this work provides a segmentation model that was designed based on densely connected convolutional networks (DenseNet) and introduces the local and global attention units. The aim of this work is to propose a novel road extraction method that can efficiently extract the road network from remote sensing imagery with local and global information. A dataset from Google Earth was used to validate the method, and experiments showed that the proposed deep convolutional neural network can extract the road network accurately and effectively. This method also achieves a harmonic mean of precision and recall higher than other machine learning and deep learning methods.


2020 ◽  
Vol 128 ◽  
pp. 19-28
Author(s):  
Teresa Pamuła

The estimation of energy consumption has become an important prerequisite for planning the implementation of electric buses and the required infrastructure for charging them in public urban transport. The article proposes a model for estimating electric bus energy consumption for the bus line of public urban transport. The developed model uses a deep learning network to estimate bus energy consumption, stop by stop, accounting for the road characteristics. The research aimed to develop a neural model for estimating electric energy consumption so that it can be easily applied in large bus networks using real data sources that are widely available to bus operators. The deep learning networks allow for the effective use of a large number of sample data (big data). The energy needed to power a bus which travels a distance from a bus stop to a bus stop is a function of selected parameters, such as distance between stops, driving time between stops, time at the bus stop, average number of passengers, the slope of the road, average speed between stops, extra energy – fixed value for the section. The given relationships were mapped using a neural network. A neural model for estimating the energy consumption of an electric bus can be used in works for determining the necessary battery capacity, for the design of optimized charging strategies and to determine charging infrastructure requirements for electric buses in a public transport network. Ocena zapotrzebowania na energię stała się ważnym warunkiem wstępnym planowania wdrażania autobusów elektrycznych oraz wymaganej infrastruktury do ich ładowania w publicznym transporcie miejskim. W artykule zaproponowano model szacowania zużycia energii przez autobus elektryczny dla linii autobusowej przedsiębiorstwa komunikacji miejskiej. W opracowanym modelu do wyznaczenia zapotrzebowania na energię autobusu na odcinku drogi od przystanku do przystanku z uwzględnieniem charakterystyki drogi lokalnej użyto sieci neuronowej typu deep learning. Celem badań było opracowanie neuronowego modelu szacowania zużycia energii elektrycznej tak, aby można go było łatwo zastosować w dużych sieciach autobusowych przy użyciu rzeczywistych źródeł danych, które są powszechnie dostępne dla operatorów transportu autobusowego. Użycie sieci typu deep learning pozwala na efektywne wykorzystanie dużej liczby danych wzorcowych (tzw. big data). Przyjęto, że wartość energii potrzebna do pokonania odległości od przystanku do przystanku autobusowego jest funkcją wybranych parametrów, takich jak: odległość między przystankami, czas trwania jazdy na odcinku między przystankami, czas przebywania autobusu na przystanku, średnia liczba pasażerów, kąt nachylenia drogi, średnia prędkość na odcinku, energia dodatkowa – stała wartość dla odcinka. Podane zależności zostały odwzorowane za pomocą sieci neuronowej. Neuronowy model oszacowania zużycia energii przez autobus elektryczny może zostać użyty w pracach mających na celu określenie niezbędnej pojemności akumulatorów, zaprojektowanie zoptymalizowanych strategii ładowania oraz określenie wymogów w zakresie infrastruktury ładowania dla autobusów elektrycznych w sieci transportu publicznego.


Sign in / Sign up

Export Citation Format

Share Document