Share Email Print

Proceedings Paper

Evaluation of H.264 and H.265 full motion video encoding for small UAS platforms
Author(s): Christopher D. McGuinness; David Walker; Clark Taylor; Kerry Hill; Marc Hoffman
Format Member Price Non-Member Price
PDF $17.00 $21.00

Paper Abstract

Of all the steps in the image acquisition and formation pipeline, compression is the only process that degrades image quality. A selected compression algorithm succeeds or fails to provide sufficient quality at the requested compression rate depending on how well the algorithm is suited to the input data. Applying an algorithm designed for one type of data to a different type often results in poor compression performance. This is mostly the case when comparing the performance of H.264, designed for standard definition data, to HEVC (High Efficiency Video Coding), which the Joint Collaborative Team on Video Coding (JCT-VC) designed for high-definition data. This study focuses on evaluating how HEVC compares to H.264 when compressing data from small UAS platforms. To compare the standards directly, we assess two open-source traditional software solutions: x264 and x265. These software-only comparisons allow us to establish a baseline of how much improvement can generally be expected of HEVC over H.264. Then, specific solutions leveraging different types of hardware are selected to understand the limitations of commercial-off-the-shelf (COTS) options. Algorithmically, regardless of the implementation, HEVC is found to provide similar quality video as H.264 at 40% lower data rates for video resolutions greater than 1280x720, roughly 1 Megapixel (MPx). For resolutions less than 1MPx, H.264 is an adequate solution though a small (roughly 20%) compression boost is earned by employing HEVC. New low cost, size, weight, and power (CSWAP) HEVC implementations are being developed and will be ideal for small UAS systems.

Paper Details

Date Published: 13 May 2016
PDF: 12 pages
Proc. SPIE 9841, Geospatial Informatics, Fusion, and Motion Video Analytics VI, 98410M (13 May 2016); doi: 10.1117/12.2238133
Show Author Affiliations
Christopher D. McGuinness, Univ. of Dayton Research Institute (United States)
David Walker, Univ. of Dayton Research Institute (United States)
Clark Taylor, Air Force Research Lab. (United States)
Kerry Hill, Air Force Research Lab. (United States)
Marc Hoffman, Air Force Research Lab. (United States)

Published in SPIE Proceedings Vol. 9841:
Geospatial Informatics, Fusion, and Motion Video Analytics VI
Matthew F. Pellechia; Kannappan Palaniappan; Peter J. Doucette; Shiloh L. Dockstader; Gunasekaran Seetharaman; Paul B. Deignan, Editor(s)

© SPIE. Terms of Use
Back to Top