JPEG XS Compression for Video Telemetry

Abstract

The Joint Photographic Experts Group (JPEG) XS video compression standard provides visually lossless imagery at typical compression ratios of 10:1, allowing transmission of multiple channels of high-definition video over GbE networks. As video bandwidths continue to increase with higher definition and higher frame rate sources, visually lossless encoding with sub-millisecond encode/decode latencies, a networked infrastructure, and a reliably constant bitrate transport stream, becomes a compelling alternative to uncompressed video. An overview of JPEG XS and transport methods is presented along with targeted test platform applications, including ground network video distribution and AI-enabled object detection-based Region of Interest (ROI) use cases for bandwidth-limited data links. Comparisons and tradeoffs with other light and higher compression technologies are discussed.