journal article
LitStream Collection
Home
Shi, Xuan; Lai, Chenggang; Huang, Miaoqing; You, Haihang
doi: 10.1111/tgis.12108pmid: N/A
Emerging computer architectures and systems that combine multi‐core CPUs and accelerator technologies, like many‐core Graphic Processing Units (GPUs) and Intel's Many Integrated Core (MIC) coprocessors, would provide substantial computing power for many time‐consuming spatial‐temporal computation and applications. Although a distributed computing environment is suitable for large‐scale geospatial computation, emerging advanced computing infrastructure remains unexplored in GIScience applications. This article introduces three categories of geospatial applications by effectively exploiting clusters of CPUs, GPUs and MICs for comparative analysis. Within these three benchmark tests, the GPU clusters exemplify advantages in the use case of embarrassingly parallelism. For spatial computation that has light communication between the computing nodes, GPU clusters present a similar performance to that of the MIC clusters when large data is applied. For applications that have intensive data communication between the computing nodes, MIC clusters could display better performance than GPU clusters. This conclusion will be beneficial to the future endeavors of the GIScience community to deploy the emerging heterogeneous computing infrastructure efficiently to achieve high or better performance spatial computation over big data.
Guan, Qingfeng; Zeng, Wen; Gong, Junfang; Yun, Shuo
doi: 10.1111/tgis.12109pmid: N/A
This article presents an improved parallel Raster Processing Library – pRPL version 2.0. Since the release of version 1.0, a series of modifications has been made in pRPL to improve its usability, flexibility, and performance. While retaining some of the key features of pRPL, the new version has gained several new features: (1) a new DataManager class has been added for integrated data management, and to facilitate data decomposition, assignment mapping, data distribution, Transition execution, and load‐balancing; (2) a GDAL‐based raster data I/O mechanism has been added to support various geospatial raster data formats, and provide centralized and pseudo parallel I/O modes; and (3) a static load‐balancing mode and a dynamic load‐balancing mode using the task‐farming technique are provided. A parallel zonal statistics tool and a parallel Cellular Automata model were developed to demonstrate the usability and performance of pRPL 2.0. The experiments using the California datasets showed that the performance altered when different pRPL options (i.e. load‐balancing mode, I/O mode and writer mode) were used for different algorithms, datasets, and varying numbers of processes.
Guo, Wei; She, Bing; Zhu, Xinyan
doi: 10.1111/tgis.12121pmid: N/A
Time is a crucial factor for many remote sensing applications such as emergency response. The traditional approach requires users to spend a lot of time downloading, processing, and viewing satellite images with specialized software. Realizing interactive real‐time processing and visualization of satellite images online is our focus. This article presents an On‐Demand computing schema for remote sensing images. A processing chain model is proposed for satellite images on a private cloud computing platform designed for the China Centre for Resources Satellite Data and Application (CCRSDA). The architecture, processing flow, optimization method, fault tolerance, and user interface are described in detail. To test the efficiency and scalability of the platform, 11 processing chains were created and three load balance experiments were executed. The results from these experiments show the validity of the proposed methods and architecture.
Li, Xiaolong; Yang, Jiansi; Guan, Xuefeng; Wu, Huayi
doi: 10.1111/tgis.12127pmid: N/A
The wide use of various sensors makes real‐time data acquisition possible. A new spatiotemporal data model, the Event‐driven Spatiotemporal Data Model (E‐ST), is proposed to dynamically express and simulate the spatiotemporal processes of geographic phenomena. In E‐ST, a sensor object is introduced into the model as a flexible real‐time data source. An event type that is generating and driving conditions is registered into a geographic object, so an event can not only express spatiotemporal change in a geographic object, but also drive spatiotemporal change in some geographic objects. As a dynamic GIS data model, the E‐ST has five characteristics – Temporality and Spatiality, Real‐time, Extendability, Causality, and Realizability. Described and realized in UML, a test‐case deployment demonstrating the impact of urban waterlogging on traffic confirms that a spatiotemporal change process in a geographic phenomena is expressed and simulated by this model. Summarizing this work, four directions for future research are outlined.
doi: 10.1111/tgis.12094pmid: N/A
With the wide use of laser scanning technology, point cloud data collected from airborne sensors and terrestrial sensors are often integrated to depict a complete scenario from the top and ground views, even though points from different platforms and sensors have quite different densities. These massive point clouds with various structures create many problems for both data management and visualization. In this article, a hybrid spatial index method is proposed and implemented to manage and visualize integrated point cloud data from airborne and terrestrial scanners. This hybrid spatial index structure combines an extended quad‐tree model at the global level to manage large area airborne sensor data, with a 3‐D R‐tree to organize high density local area terrestrial point clouds. These massive point clouds from different platforms have diverse densities, but this hybrid spatial index system has the capability to organize the data adaptively and query efficiently, satisfying the requirements for fast visualization. Experiments using point cloud data collected from the Dunhuang area were conducted to evaluate the efficiency of our proposed method.
Guan, Xuefeng; Cheng, Bo; Song, Aihong; Wu, Huayi
doi: 10.1111/tgis.12123pmid: N/A
Web Map Tile Services (WMTS) are widely used in many fields to quickly and efficiently visualize geospatial data for public use. To ensure that a WMTS can successfully fulfill users' expectations and requirements, the performance of a service must be measured to track latencies and bottlenecks that may downgrade the overall quality of service (QoS). Traditional synthetic workloads used to evaluate WMTS applications are usually generated by repeated static URLs, through randomized requests, or by an access log replay. These three methods do not take request characteristics and users' behaviors into consideration, while access logs are not available for systems still under development. Thus, the evaluation outcomes obtained by these methods cannot represent the real performance of online WMTS applications.
Pei, Tao; Sobolevsky, Stanislav; Ratti, Carlo; Amini, Alexander; Zhou, Chenghu
doi: 10.1111/tgis.12128pmid: N/A
The aggregated mobile phone network (AMPN) (i.e. the calling time or numbers are aggregated at every vertex), which records the call volume between different places over time, has been studied extensively to reveal the mobility patterns of residents, etc. Nevertheless, most previous works were implemented based on the non‐directionality of the network model. This simplification may overlook some important characteristics of AMPN. To explore the AMPN as a directional network model, we introduce the concept of directional heterogeneity in the study of AMPN data. The heterogeneity is twofold: (1) the imbalance of vertex (difference between outgoing and incoming calls of the vertex); and (2) the reciprocity of each edge (difference between the directed weights of the same edge). Taking the data of Singapore as an example, we systematically analyze the directional heterogeneity of AMPN. Our findings include three aspects. First, the AMPN shows as more unbalanced in the night‐time than in the daytime, and its imbalance decreases as vertex granularity increases. Second, the directional heterogeneity varied with locations. Specifically, the residential area is dominated by deficits and others by surpluses. Third, the trajectories of incoming and outgoing calls follow a similar geographical pattern (i.e. southeast‐north‐south‐north‐southeast), indicating the calling behavior and routine mobility of users over time and space.
Showing 1 to 8 of 8 Articles