Lossy compression of hyperspectral data using vector quantization

摘要: Abstract Efficient compression techniques are required for the coding of hyperspectral data. Lossless is in transmission and storage data within distribution .system. Lossy have a role initial analysis where large quantities evaluated to select smaller areas more detailed evaluation. Central lossy developmgent suitable distortion measure, this work discusses applicability extant measures video imagery. Criteria remote sensing measure developed discussed. One [the percentage maximum absolute (PMAD) measure] considered be candidate application remotely sensed images. The effect then investigated on likelihood classification images, both directly original reconstructed features extracted by decision boundary feature extraction (DBFE) technique. PMAD determined an image with varying degrees distortion. Despite some anomalies caused challenging discrimination tasks, accuracy total its constituent classes remains predictable as level increases. Although reduced from 96.8% 82.8% compressed 4% PMAD, loss not significant (less that 8%)for most other than those present problem. Yet 1/17 size original.