Ramanathan Sugumaran

Distributed LiDAR data processing in a high-memory cloud-computing environment

This article explores the use of an advanced, high-memory cloud-computing environment to process large-scale topographic light detection and ranging (LiDAR) data. The new processing techniques presented herein for LiDAR point clouds intelligently filter and triangulate a data set to produce an accurate digital elevation model. Ample amounts of random-access memory (RAM) allow the employment of efficient hashing techniques for spatial data management; such techniques were utilized to reduce data distribution overhead and local search time during data reduction. Triangulation of the reduced, distributed data set was performed using a local streaming approach to optimize processor utilization. Computational experiments used Amazon Web Services Elastic Compute Cloud resources. Analysis was performed to determine (1) the accuracy of the binning/array-based reduction, as measured by root mean square error and (2) the scalability of the approach on varying-size clusters of high-memory instances (nodes having 244 GB of RAM). For experimental data sets, topographic LiDAR data generated by the Iowa LiDAR Mapping Project was used. This article concludes that the data-reduction strategy is computationally efficient and outperforms a comparable randomized filter control when moderate reduction is undertaken – e.g., when the data set is being reduced by between 30% and 70%. Performance speed-up ratios of up to 3.4, comparing between a single machine and a 9-node cluster, are exhibited. A task-specific stratification of the results of this work demonstrates Amdahl’s law and suggests the evaluation of distributed databases for geospatial data.

Parallel Landscape Driven Data Reduction & Spatial Interpolation Algorithm for Big LiDAR Data

Airborne Light Detection and Ranging (LiDAR) topographic data provide highly accurate digital terrain information, which is used widely in applications like creating flood insurance rate maps, forest and tree studies, coastal change mapping, soil and landscape classification, 3D urban modeling, river bank management, agricultural crop studies, etc. In this paper, we focus mainly on the use of LiDAR data in terrain modeling/Digital Elevation Model (DEM) generation. Technological advancements in building LiDAR sensors have enabled highly accurate and highly dense LiDAR point clouds, which have made possible high resolution modeling of terrain surfaces. However, high density data result in massive data volumes, which pose computing issues. Computational time required for dissemination, processing and storage of these data is directly proportional to the volume of the data. We describe a novel technique based on the slope map of the terrain, which addresses the challenging problem in the area of spatial data analysis, of reducing this dense LiDAR data without sacrificing its accuracy. To the best of our knowledge, this is the first ever landscape-driven data reduction algorithm. We also perform an empirical study, which shows that there is no significant loss in accuracy for the DEM generated from a 52% reduced LiDAR dataset generated by our algorithm, compared to the DEM generated from an original, complete LiDAR dataset. For the accuracy of our statistical analysis, we perform Root Mean Square Error (RMSE) comparing all of the grid points of the original DEM to the DEM generated by reduced data, instead of comparing a few random control points. Besides, our multi-core data reduction algorithm is highly scalable. We also describe a modified parallel Inverse Distance Weighted (IDW) spatial interpolation method and show that the DEMs it generates are time-efficient and have better accuracy than the one’s generated by the traditional IDW method.