[Defense] Resource-Efficient Hardware-Software Co-Design for the Indexing and Storage of Voxelized 3D Point-Clouds
Thursday, December 9, 2021
3:00 pm - 4:00 pm
will defend his proposal
Resource-Efficient Hardware-Software Co-Design for the Indexing and Storage of Voxelized 3D Point-Clouds
3D maps with many millions to billions of points are now used in an increasing number of applications, with processing rates in the hundreds of thousands to millions of points per second. In mobile applications, power and energy consumption for managing such data and extracting useful information thereof are critical concerns. We have developed structures and methodologies with the purpose of minimizing memory usage and associated energy consumption for indexing and serialization of voxelized point-clouds. The primary source of points in our case is airborne laser scanning, but our methodology is not restricted to only such setting. Our emulated results show a memory usage reduction factor of roughly up to 200x that of Octree/Octomap, and a file size reduction factor of up to 1.65x compared the predominating compression scheme for airborne Lidar data, LASzip. In addition, our structures enable significantly more efficient processing since they are included in a hierarchical structure that captures geometric aspects. We also designed a memory model suitable for hosting our data structures, as well as an appropriate instruction set architecture for manipulating the data. Modules for both the memory model and the ISA were implemented on an FPGA platform in order to measure performance and energy efficiency. In general, this system promises significant resource efficiencies for applications that rely on data structures with a considerable number of memory address pointers, or packed binary data structures.
3:00 PM - 4:00 PM CT
Online via Zoom
Dr. Lennart Johnsson, dissertation advisor
Faculty, students and the general public are invited.