Quality Assessments of Standard Video Compression Techniques Applied to Hyperspectral Data Cubes

With a satellite-borne low-cost Hyper Spectral Imager (HSI) a large - target area can be imaged. HSIs can detect oceanic phenomena e.g. algal distribution and environmental spills, enabling quicker reactions by authorities. The HSI provides spatially resolved spectral information. The resultant datasets are large, and the capacity to transmit data to the ground is severely limited. To reduce the size of the dataset, compression is required. Various freely available compression algorithms exist. In this paper, algorithms are assessed for their suitability for this application. Uncompressed reference datasets from the HSI are compressed with the H.263, H.264, and H.265 algorithms, varying the Quantization Parameter (QP). The compressed datasets are compared to the original data using several tests. H.263 and H.264 perform the spectral tests poorly, but H.265 (QP=30) passes the spectral tests. Moreover, H.265 achieves the best balance between quality and data reduction and is recommended for the satellite-borne HSI.