Video Data Compression

In conventional video transmission systems, data is scanned at a constant rate with a fixed optical resolution and transmitted with a given magnitude accuracy. In adaptive video transmission systems, more efficient utilization of bandwidth and/or power is accomplished by varying the optical resolution and amplitude accuracy as a function of picture detail. In picture areas where there is high detail, video information is scanned with high optical resolution and its magnitude is transmitted with low accuracy. This is possible because the eye is relatively insensitive to magnitude errors in high detail regions. Correspondingly, areas of low detail are scanned with low optical resolution but transmitted with high amplitude accuracy. This paper describes several methods for implementing these concepts. It is shown how redundancy reductionalgorithms, employing ??a posteriori?? curve fitting techniques, are used to adaptively change the effective resolution as a function of picture detail. Adaptive aperture is also demonstrated as a technique for changing video data accuracy in accordance with picture detail. Several examples of compressed and reconstructed video are shown. These include TIROS video, RANGER photos, and high resolution aerial reconnaissance photographs. The reduction in bandwidth obtained is demonstrated to be as high as 6 to 1. A complete system for compressing video data, encoding the compressed data, transmitting it, and reconstructing the data into the original form is also described.