Systems And Methods for Hybrid Compression Of Spectral Image Data

• US 20140307975A1
• Filed: 04/10/2013
• Published: 10/16/2014
• Est. Priority Date: 04/10/2013
• Status: Active Grant

• Current Assignee: Southwest Research Institute
• Original Assignee: Southwest Research Institute
• Inventors: DeForest, Craig
• Granted Patent: US 9,031,336 B2
• Time in Patent Office: Days
• US Class Current: 382/233
• CPC Class Codes: G06T 9/00 Image coding bandwidth or redundancy reduction for static pictures H04N1/41; coding or decoding of static colour picture signals H04N1/64; methods or arrangements for coding, decoding, compressing or decompressing digital video signals H04N19/00

Abstract

A method and system for generating a compressed spectral image is provided. Spectral image data including a plurality of spectral intensity values is received. The spectral intensity values are associated with a first spatial dimension (x-dimension), a second spatial dimension (y-dimension) and a wavelength dimension (λ-dimension). A window is applied to the spectral image data along the λ-dimension, to select a subset of the spectral image data corresponding to a range of wavelengths. A Fourier transform is performed on the windowed spectral image data along the λ-dimension, at locations along the x-dimension and y-dimension, to generate Fourier coefficients associated with each of the locations. The Fourier transformed data is filtered by retaining a subset of the Fourier coefficients at each of the locations. Wavelet compression is performed on the filtered data along the x-dimension and the y-dimension to generate the compressed spectral image.

Claims

1. A method for generating a compressed spectral image, said method comprising:

receiving spectral image data, said spectral image data comprising a plurality of spectral intensity values, each of said spectral intensity values associated with a first spatial dimension (x-dimension), a second spatial dimension (y-dimension) and a wavelength dimension (λ

-dimension);

applying a window to said spectral image data along said λ

-dimension, said window configured to select a subset of said spectral image data corresponding to a range of wavelengths;

performing a Fourier transform on said windowed spectral image data along said λ

-dimension, said Fourier transform performed at a plurality of locations along said x-dimension and said y-dimension, said Fourier transform configured to generate a plurality of Fourier coefficients associated with each of said locations;

filtering said Fourier transformed data, said filtering comprising retaining a subset of said Fourier coefficients at each of said locations; and

performing wavelet compression of said filtered data along said x-dimension and along said y-dimension to generate said compressed spectral image.

2. The method of claim 1, wherein a compression ratio is selected for said wavelet compression, said compression ratio associated with said Fourier coefficient.

3. The method of claim 1, wherein said Fourier coefficients are associated with a frequency range and said retained subset of Fourier coefficients comprise Fourier coefficients associated with lower frequencies of said frequency range.

4. The method of claim 3, wherein said lower frequencies comprises a lowest 15% of said frequency range of said Fourier coefficients.

5. The method of claim 3, wherein said lower frequencies comprises a lowest 33% of said frequency range of said Fourier coefficients.

6. The method of claim 1, further comprising quantizing said Fourier coefficients to a selected number of bits, wherein said selected number of bits is based on a frequency associated with said Fourier coefficient.

7. The method of claim 1, wherein said x-dimension comprises a range of 400 to 600 samples and said y-dimension comprises a range of 600 to 1000 samples.

8. The method of claim 1, wherein the length of said window comprises a range of 16-128 samples.

9. The method of claim 1, further comprising transmitting said compressed spectral image over a bandwidth limited communications link.

10. The method of claim 1, further comprising:

performing wavelet decompression on said compressed spectral image along said x-dimension and along said y-dimension; and

performing an inverse Fourier Transform on said wavelet decompressed data to generate a decompressed spectral image.

11. A system for generating a compressed spectral image, said system comprising:

an image compression module configured to receive spectral image data, said spectral image data comprising a plurality of spectral intensity values, each of said spectral intensity values associated with a first spatial dimension (x-dimension), a second spatial dimension (y-dimension) and a wavelength dimension (λ

-dimension);

a windowing module configured to apply a window to said spectral image data along said λ

-dimension, said window configured to select a subset of said spectral image data corresponding to a range of wavelengths;

a Fourier transform module configured to perform a Fourier transform on said windowed spectral image data along said λ

-dimension, said Fourier transform performed at a plurality of locations along said x-dimension and said y-dimension to generate a plurality of Fourier coefficients associated with each of said locations;

a filter module configured to filter said Fourier transformed data, said filtering comprising retaining a subset of said Fourier coefficients at each of said locations; and

a wavelet compression module configured to perform wavelet compression of said filtered data along said x-dimension and along said y-dimension to generate said compressed spectral image.

12. The system of claim 11, wherein a compression ratio is selected for said wavelet compression, said compression ratio associated with said Fourier coefficient.

13. The system of claim 11, wherein said Fourier coefficients are associated with a frequency range and said retained subset of Fourier coefficients comprise Fourier coefficients associated with lower frequencies of said frequency range.

14. The system of claim 13, wherein said lower frequencies comprises a lowest 15% of said frequency range of said Fourier coefficients.

15. The system of claim 13, wherein said lower frequencies comprises a lowest 33% of said frequency range of said Fourier coefficients.

16. The system of claim 11, wherein said filter module is further configured to quantize said Fourier coefficients to a selected number of bits, wherein said selected number of bits is based on a frequency associated with said Fourier coefficient.

17. The system of claim 11, wherein said x-dimension comprises a range of 400 to 600 samples and said y-dimension comprises a range of 600 to 1000 samples.

18. The system of claim 11, wherein the length of said window comprises a range of 16-128 samples.

19. The system of claim 11, further comprising a transmission module configured to transmit said compressed spectral image over a bandwidth limited communications link.

20. The system of claim 11, further comprising an image decompression module configured to perform wavelet decompression on said compressed spectral image along said x-dimension and along said y-dimension and further configured to perform an inverse Fourier Transform on said wavelet decompressed data to generate a decompressed spectral image.

21. An article comprising a non-transitory storage medium having stored thereon instructions that when executed by a machine result in the following operations:

receiving spectral image data, said spectral image data comprising a plurality of spectral intensity values, each of said spectral intensity values associated with a first spatial dimension (x-dimension), a second spatial dimension (y-dimension) and a wavelength dimension (λ

-dimension);

applying a window to said spectral image data along said λ

-dimension, said window configured to select a subset of said spectral image data corresponding to a range of wavelengths;

performing a Fourier transform on said windowed spectral image data along said λ

-dimension, said Fourier transform performed at a plurality of locations along said x-dimension and said y-dimension, said Fourier transform configured to generate a plurality of Fourier coefficients associated with each of said locations;

filtering said Fourier transformed data, said filtering comprising retaining a subset of said Fourier coefficients at each of said locations; and

performing wavelet compression of said filtered data along said x-dimension and along said y-dimension to generate said compressed spectral image.

22. The article of claim 21, wherein a compression ratio is selected for said wavelet compression, said compression ratio associated with said Fourier coefficient.

23. The article of claim 21, wherein said Fourier coefficients are associated with a frequency range and said retained subset of Fourier coefficients comprise Fourier coefficients associated with lower frequencies of said frequency range.

24. The article of claim 23, wherein said lower frequencies comprises a lowest 15% of said frequency range of said Fourier coefficients.

25. The article of claim 23, wherein said lower frequencies comprises a lowest 33% of said frequency range of said Fourier coefficients.

26. The article of claim 21, further comprising the operation of quantizing said Fourier coefficients to a selected number of bits, wherein said selected number of bits is based on a frequency associated with said Fourier coefficient.

27. The article of claim 21, wherein said x-dimension comprises a range of 400 to 600 samples and said y-dimension comprises a range of 600 to 1000 samples.

28. The article of claim 21, wherein the length of said window comprises a range of 16-128 samples.

29. The article of claim 21, further comprising the operation of transmitting said compressed spectral image over a bandwidth limited communications link.

30. The article of claim 21, further comprising the operations of:

performing wavelet decompression on said compressed spectral image along said x-dimension and along said y-dimension; and

performing an inverse Fourier Transform on said wavelet decompressed data to generate a decompressed spectral image.