LOSSLESS COMPRESSION ALGORITHM FOR HYPERSPECTRAL DATA
First Claim
1. A method of compressing a stream of input data elements to produce an output data stream, the method comprising the steps of:
- (a) selecting at least one model, from a set of predetermined mathematical models, each mathematical model parameterizing a space with a set of model specific coordinates, the coordinates represented by a tuple of at least one coordinate values, and outputting for each selected model parameters comprising a map, from a tuple of coordinate values in the model space to an approximating set of data values for at least one subportion of the data elements,(b) determining, for at least one subportion of the data elements, a label indicating a selected model, and the values of a tuple of model coordinates which map to a tuple of data values which, according to some measure of merit, best approximate the subportion of the data elements, and(c) outputting at least the model coordinates as at least a portion of an output data stream.
1 Assignment
0 Petitions
Accused Products
Abstract
Lossless compression techniques provide efficient compression of hyperspectral satellite data. The present invention combines the advantages of a clustering with linear modeling. A number of visualizations are presented, which help clarify why the approach of the present invention is particularly effective on this dataset. At each stage, the algorithm achieves an efficient grouping of the data points around a relatively small number of lines in a very large dimensional data space. The parametrization of these lines is very efficient, which leads to efficient descriptions of data points. The method of the present invention yields compression ratios that compare favorably with what is currently achievable by other approaches.
61 Citations
22 Claims
-
1. A method of compressing a stream of input data elements to produce an output data stream, the method comprising the steps of:
-
(a) selecting at least one model, from a set of predetermined mathematical models, each mathematical model parameterizing a space with a set of model specific coordinates, the coordinates represented by a tuple of at least one coordinate values, and outputting for each selected model parameters comprising a map, from a tuple of coordinate values in the model space to an approximating set of data values for at least one subportion of the data elements, (b) determining, for at least one subportion of the data elements, a label indicating a selected model, and the values of a tuple of model coordinates which map to a tuple of data values which, according to some measure of merit, best approximate the subportion of the data elements, and (c) outputting at least the model coordinates as at least a portion of an output data stream. - View Dependent Claims (2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22)
-
Specification